Podcasts about series c

The latest episode of the Cambridge Tech Podcast features a fascinating panel discussion on international trade recorded live at this year's Cambridge Wide Open Week.Steven Lynch, Director of International Trade at the British Chambers of Commerce (BCC), opens with a sobering statistic: whilst only 10% of British businesses export, the BCC's network achieves a remarkable 45%. This disparity represents a significant untapped opportunity. As Lynch points out, "If we increased UK exports by 2%, we could improve UK GDP by 0.6%."The problem isn't a lack of innovation or ambition; it's a delivery gap. UK founders and scientists often lack the business expertise and market clarity needed to confidently enter international markets. This is where the new trade accelerator comes in.Rather than offering generic webinars about exporting, the BCC has created something fundamentally different: an end-to-end, process-driven accelerator specifically targeting life sciences companies interested in Singapore.Singapore might not be the obvious first choice for UK startups - the US typically dominates - but it's positioned as a strategic gateway to Southeast Asia. With a population of 6 million, Singapore punches well above its weight as a hub for healthcare innovation and life sciences.Jay Sadanandan, a Singapore life sciences expert, highlights the market's unique advantages:The Health Sciences Authority recently launched a close collaboration with the UK's MHRA, enabling regulatory expediency for innovative technologiesAccess to world-class clinical research infrastructure and patient populationsSingapore's ambitious government vision to solve prevention through treatment across healthcareThe 10 selected companies span pre-seed to Series C stages and focus on areas where UK strength aligns with Singapore opportunity: genomics, molecular diagnostics, advanced therapeutics, and enabling technologies like biologic stabilisation. Importantly, companies aren't required to relocate or abandon UK operations. Instead, they use Singapore as a commercial testbed, a lower-risk environment to validate their business model before scaling across Asia Pacific, the US, or beyond.What's Next?The pilot launches in September 2026, but the ambitions extend far beyond Singapore. Lynch reveals plans to scale to Hong Kong, the Gulf, and other strategic markets. The ultimate goal? Moving from 15 businesses entering one market annually to 15 businesses entering different markets each month.For UK startup founders and investors watching the export opportunity, this episode offers genuine insight into how the next generation of British life sciences companies might achieve global scale, and why Singapore might be the launchpad you haven't considered yet.Listen to the full episode on the Cambridge Tech Podcast to hear the complete discussion, including audience Q&A and practical advice for companies considering international expansion.Headline sponsor Holden Polestar Produced by Cambridge TV #CamTechPod Hosted on Acast. See acast.com/privacy for more information.

Yasir Arafat, Co-Founder, CTO, & President, stops by the Energy News Beat podcastOn July 4th, 2024—America's 250th birthday—something extraordinary happened that most people missed. Allo Atomics turned on a brand new advanced nuclear reactor for the first time in 50 years. But here's what makes this truly revolutionary: they designed, built, and achieved criticality in less than 12 months. In this episode of the Energy Newsbeat Podcast, host Stu Turley sits down with Yasir Arafat, CTO and President of Allo Atomics, to explore how a company that was just two people 2.5 years ago is now rebuilding American manufacturing muscle and solving one of humanity's greatest challenges—powering the AI revolution while ending global energy poverty.From factory-based mass production to autonomous reactors powered by AI, from 36-day construction timelines to modular power plants that rival grid reliability, this conversation reveals how American innovation is about to transform energy forever.This is the story of how nuclear went from a decade-long project to a factory product—and why it matters for every country on Earth.Buckle up, as this was a huge discussion, and I hope that it is the first of many with Aalo Atomics. They are going places, and we need them to be as successful as they seem likely to be. They have the nuclear market and are designing the entire supply chain, from standardized fuel to go-to-market strategies, controlling it along the way.Secretary Chris Wright was at their offices and personally signed their approvals for critical mass testing. This is huge, and they were one of four reactors that kicked off our 250th anniversary in style.This is just about as cool as it gets - Well done Aalo Atomics and Secretary Chris Wright!1. Advanced Nuclear Reactor Development & AchievementYasir Arafat, CTO and President of Allo Atomics, discusses their groundbreaking accomplishment of building and turning on a brand new advanced nuclear reactor on July 4th—the first time this has been done in 50 years. This was achieved in less than 12 months as part of an executive order challenge from President Trump.2. Rapid Company Growth & Manufacturing Scale-UpAllo Atomics grew from a 2-person company 2.5 years ago to nearly 200 employees. The company is expanding its manufacturing facility from 40,000 square feet to 200,000 square feet this year, with plans to reach a million square feet within three years—positioning it as the largest nuclear reactor manufacturing facility in the United States.3. Factory-Based Mass Production ModelRather than traditional project-based nuclear construction, Allo is pioneering a factory mass-production approach similar to the automotive and aerospace industries. This includes automated welding systems and modular component design that can be transported and assembled like “Lego blocks.”4. AI Data Center Power DemandsThe podcast highlights the massive energy needs of AI data centers, which require 60 gigawatts of power over the next five years. Nuclear provides the ideal solution as a 24/7, clean baseload power source that doesn't depend on weather.5. Fuel Supply Chain & EnrichmentDiscussion of uranium enrichment levels, with Allo's reactors operating at 5% enrichment using commercially available uranium dioxide fuel—avoiding the need for exotic fuel forms or military-grade materials that would complicate supply chains.6. Modular Power Plant Design (AuloPOD)Allo designed a 50-megawatt power plant with multiple reactors and turbines providing N+1 redundancy, achieving 99.9% availability. This allows data centers to operate independently from the grid while maintaining reliability.7. Accelerated Construction & DeploymentThe company built its first nuclear reactor building in just 36 days—reducing construction costs by one-fifth. Their goal is to compress the entire deployment process from order to operation in less than 12 months, compared to the traditional 6-10+ years.8. Vertical Integration & Supply Chain SolutionsAllo manufactures the majority of hardware in-house while partnering with 100+ suppliers. The philosophy is “no unobtanium”—every component must be obtainable from existing suppliers or manufactured internally.9. AI Integration in Nuclear Design & OperationsAllo uses AI extensively for engineering design, licensing acceleration, and autonomous reactor operations. They have 62+ major AI projects running, with human validation at every step to ensure quality and safety.10. Global Energy Impact & Future ApplicationsThe conversation extends beyond data centers to broader applications: ending energy poverty worldwide, powering remote locations, military bases, marine vessels, and even lunar bases. The vision is to transform the U.S. into an energy exporter and uplift nations through reliable, low-cost nuclear power.11. Series C Funding & IPO PlansAllo is currently raising Series C funding and plans to go public only after demonstrating its ability to execute at scale with licensed products and generate revenue.This is a compelling discussion about how American innovation and manufacturing expertise can solve the global energy crisis through advanced nuclear technology.Check out Aalo Atomics here: https://www.aalo.com/Connect with Yasir on his LinkedIn here: https://www.linkedin.com/in/yasiraalo/A shout-out to Steve Reese and the Reese Energy Consulting group for sponsoring the Podcast https://reeseenergyconsulting.com/.Data2 if you have any business systems, can you trust A? Well, they have the patent on validation. . https://data2.zoholandingpage.com/energyAnd we have WellDatabase rolling in as a new sponsor. https://welldatabase.com/

Today I'm delighted to welcome Andy Parker, CEO of Step Pharma.With over 25 years of experience across AstraZeneca, Shire, Zealand Pharma, andventure capital, Andy has led Step Pharma since 2019. The company is pioneeringa targeted approach to cancer and blood disorders by inhibiting the enzymeCTPS1. Their lead candidate, dencatistat, blocks this pathway that certaincancer cells and activated immune cells rely on, while sparing healthy cellsthat use the related CTPS2 enzyme.In this episode, we'll dive into the science behind this mechanism, explore StepPharma's expanding pipeline from lymphomas and solid tumours to essentialthrombocythaemia, and discuss their recent €38 million Series C financing.We'll also look ahead to the future of precision oncology.01:17 Meet Andy Parker06:12 The biotech ecosystem around Geneva07:51 The CTPS1 enzyme and why cancer cells depend on it14:08 Pipeline-in-a-product strategy across three indications20:14 The series C: €38 million raise25:41 Partnering with big pharma: possibilities and limits28:11 The future of precision oncology and metabolic targetingInterested in being a sponsor of an episode of our podcast? Discover how you can get involved here! Stay updated by subscribing to our newsletterTo dive deeper into the topic: Step Pharma and Concr to partner on cancer treatmentStep Pharma moves into oncology clinical trialsStep Pharma announces promising pre-clinical cancer data

We've been running a bit of an Agent Cloud series surveying all the top inference/compute/cloud providers, from Databricks to Daytona to Railway and, even further back, E2B, but we're excited to conclude this series returning to Modal, which has just raised a monster $355M Series C.The cloud was built for developers. But agents are now changing that.The old infra stack was designed for a human who could read docs, reason through YAML, and understand dashboards to figure out what they need when something broke. While this was painful for developers, it worked since they could fill in missing context in their heads.However, agents don't have that luxury. Now in this new era of agents, everything has to be tighter.They need a place to write code, run it, inspect the output, change the environment, debug failures, and try again. Fast iteration and feedback loops with all the necessary context are crucial for agents to operate properly. Furthermore, sandboxes are a clear representation of this shift as agents can easily spin up isolated environments. This programmatic infra even extends to research:Two years ago, we were one of the first to cover Modal with CEO Erik Bernhardsson and Alessio designed our favorite LS thumbnail of all time:At the time, Modal was just a teeny little company with a $17M Series A.Today, fresh off their $355M Series C, Modal is one of the clearest examples of the agent cloud future being built in real time: a cloud platform moving past traditional web app assumptions toward the workloads AI actually creates such as elastic inference, sandboxes, GPU burst, post-training, background agents, and infrastructure that agents themselves can operate.In this episode, Modal CTO Akshat Bubna joins swyx and Vibhu to unpack why AI applications don't fit traditional cloud assumptions, why Kubernetes was never designed for bursty compute-heavy workloads, and why Modal is now shifting from developer experience to agent experience.We go deep on Modal's AI infra stack: serverless functions, decorator-based infrastructure, elastic inference for custom models, GPU snapshotting, DeFlash, speculative decoding, Auto Endpoints, sandboxes, persistent storage, networked containers, private IPv6, RDMA, multi-node training, and Modal's capacity pool across 17 cloud providers. Akshat also explains why RL rollouts can require 100,000 sandboxes, why production agents need hard guardrails, why observability may matter more than reading code, and why AI has made infrastructure exciting again.We discuss:* Why Kubernetes wasn't built for bursty AI workloads* How Modal started as a better runtime before becoming an AI cloud* Why Modal added GPUs before ChatGPT* The shift from developer experience to agent experience* Why observability matters when agents are writing the code* Elastic inference for custom models across audio, video, robotics, and comp bio* GPU snapshotting, cold starts, and why inference workloads are so bursty* Why RL rollouts can require 100,000 sandboxes* DeFlash, speculative decoding, and frontier-level inference performance* Auto Endpoints and making optimized inference easier to deploy* What Modal adds beyond vLLM, SGLang, and raw GPU rental* Modal's 17-cloud capacity pool and supercloud strategy* Networked sandboxes, sidecars, private IPv6, and RDMA* Serverless multi-node training for post-training and research workloads* Auto-research, model-guided sweeps, and agents launching GPU experiments* Compute strategy, capacity planning, and batch tiers* Why production agents need specialized sandboxes and hard guardrails* Modal's take on managed agents, CI, Gitpod/Ona, Python, TypeScript, and Modal BenchAkshat Bubna* LinkedIn: https://www.linkedin.com/in/akshat-bubna-188885103* X: https://x.com/akshat_bModal* Website: https://modal.comTimestamps00:00:00 Introduction00:00:39 Modal's origin and why Kubernetes wasn't enough00:04:32 Developer Experience → Agent Experience00:06:21 Modal's AI cloud primitives00:09:14 Sandboxes, agent loops, and proto-Cognition00:12:12 Elastic inference, GPU snapshotting, and 100,000 sandboxes00:15:24 DeFlash, speculative decoding, and Auto Endpoints00:19:59 Production-grade inference beyond raw GPUs00:22:00 Background agents, Ramp Inspect, and the agent lifecycle00:24:08 Modal's 17-cloud supercloud strategy00:26:40 Networked sandboxes, private IPv6, and RDMA00:32:48 Multi-node training, post-training, and auto research00:37:36 Compute strategy, capacity planning, and batch tiers00:40:55 Open models, real-time AI, and production agent infra00:43:06 Hard guardrails, managed agents, and specialized sandboxes00:46:06 Why AI made infrastructure exciting again00:48:30 Model APIs, differentiated products, and agentic video00:51:50 CI, coding-agent infra, SDKs, and Modal Bench00:57:28 Closing ThoughtsTranscriptIntroduction: Modal, Series C, and the Art PartySwyx [00:00:00]: We're here with Akshat, CTO of Modal, together with Vibhu. Congrats on your Series C.Akshat [00:00:10]: Thank you.Swyx [00:00:11]: Your party yesterday was amazing.Akshat [00:00:15]: Yeah.Swyx [00:00:15]: From all the photos and all the swag.Akshat [00:00:17]: We had a bunch of art installations, which was fun, seeing, like, our products on pedestals next to, like, Rodin.Swyx [00:00:25]: Very nice. Very nice. When you started, it was not the GPU inference company. Maybe it was in your mind. Take us back to the origin story.Modal's Origin: A New Runtime Beyond KubernetesAkshat [00:00:39]: I first met Eric, who's the CEO, through an investor. Back then Eric was already thinking about building, a new runtime, and he got there thinking through why are workflow orchestration products so hard to use. It's because you have to run them on Kubernetes. Kubernetes is hard to manage. It's not built for burstiness and, custom images,Swyx [00:01:03]: YeahAkshat [00:01:03]: It has a terrible developer experience.Swyx [00:01:05]: And I'll, I'll interjectAkshat [00:01:06]: YeahSwyx [00:01:07]: For listeners, who are new, we interviewed Eric two years ago, and there's a bit more of the story there from Spotify and all those things.Swyx [00:01:14]: And I came across Eric through Data Council because he did that talk on the serverless container stack that you guys did, which was like, that was my first like, “Okay, I need to take Modal very seriously” moment.Akshat [00:01:26]: Yeah.Swyx [00:01:26]: But it was still very unclear, like, do I need all this for just my data pipelines?Akshat [00:01:33]: Yeah. initially what we were thinking about was if we build a better runtime, it's a very useful primitive in itself. It's There's a lot of things that, get solved by serverless functions, like you can do, ETL stuff, you can do job queues, you can do all this, like, bursty processing, which it turns out every company had needs for. but then we also were thinking about this as like, this is a primitive that we can build a whole collection of products on, which are very verticalized. So perhaps data engineering would've been the first one, but we were thinking about inference. Back then it was more classical inference, like computer vision stuff and running XGBoosts and whatnot. But we added GPUs to the product a year before ChatGPT came out.From Serverless Containers to GPU WorkloadsSwyx [00:02:19]: Nice.Akshat [00:02:19]: We just didn't think it would be that big of a deal.Swyx [00:02:22]: Yeah, just like add A100.Vibhu [00:02:23]: Was there any, like, early key problem that really sparked off why you built it?Akshat [00:02:28]: Yeah. Primarily it's just, none of the tooling that was out there was built for, one, a really great developer experience, and also there's a general trend of, a lot of the workloads that we were seeing were very. I wish there was a better word for it, but compute-heavy. Like, they need, one, like, need a lot more resources, so you need to burst up and down a lot, versus like Kubernetes designed for, like, slow scaling and, more for, like, web server use cases. And also there's just a lot more specialization in, like, what kinds of environments these workloads run in. Like, we had sometimes they need accelerators, sometimes they need different kinds of images, and this is just like a consistent thing that we saw across a lot of companies. That would be the next step.Software-Defined Infrastructure and Decorator-Based DXSwyx [00:03:13]: Yeah. Yeah. Be nice. I don't know how much this factored into the early story, but I wrote a post when I was at Temporal about infrastructure, software-defined infrastructure or something like that.Akshat [00:03:22]: Yeah, the self-provisioningSwyx [00:03:23]: Self-provisioning.Akshat [00:03:24]: Yeah.Swyx [00:03:24]: Yeah. I can't even remember my own post.Swyx [00:03:26]: And then you put me on the landing page.Akshat [00:03:28]: Yeah. We really like, the term and so we stole it.Swyx [00:03:32]: Because you had the insight that everything can just be in decorators co-located with the code, right?Akshat [00:03:37]: Yeah.Swyx [00:03:37]: Was that a big part of the originalAkshat [00:03:39]: YesSwyx [00:03:39]: Story or it was just like a DX layer?Akshat [00:03:41]: That was, really important because we really didn't want people to spend, so much time, writing YAML, and it seemed like you could really condense the surface area of what you're doing, put it in code so you can operate on it just like you operate on other code, and like build stuff that's more expressive and dynamic. and so yeah, that was always a very important part.Swyx [00:04:04]: Then the pushback is this is a DSL.Akshat [00:04:07]: Yeah.Swyx [00:04:07]: It's you're closed source. I am locked into Modal.Akshat [00:04:11]: Yeah. We never really got pushback for that because the nice thing about Modal is you can bring whatever code you have, and sure, the DSL is at the configuration layer for, what hardware you're using, how you're scaling things up, but you still own the code.Akshat [00:04:27]: And that's, that's been an important, part of our story, even as we do inference now.Swyx [00:04:32]: Yeah.Vibhu [00:04:32]: How much of do you think still stays the same today? Like if you were to build something today, DevX very important, but I feel like, a lot of this has been changed with just hook it up to an agent, have Claude Code, have Codex implement a tool. there's very agent native primitives that are different than if I'm doing this myself, right?Developer Experience → Agent ExperienceAkshat [00:04:54]: We've changed our SDK team to think about agent experience instead of, developer experience and we think that the same benefits that apply for DX also apply for AX, which is why would you have an agent read through hundreds of Kubernetes files and like write YAML that's not even typed when it can make a couple of changes in a decorator and it gets this self-provisioning runtime of, being able to see its changes live in action? yeah, it just seems from the customers we talk to, they find Modal is much faster for agents to use versus operating on a different substrate.Swyx [00:05:34]: Yeah, because like you, again, you co-locate the infrastructure requirements to the code that runs it.Akshat [00:05:38]: Yeah.Swyx [00:05:38]: Well, the negative thesis now is that nobody's looking at their code anymore, so there's no point.Akshat [00:05:44]: Yeah, people aren't looking at code. one thing we still see is really important is observability.Swyx [00:05:51]: Yeah.Akshat [00:05:51]: Like how good is your dashboard? And of course, like we have, we push a lot of it to the CLI so the agents can do their own investigation, but you still need humans to go interpret what's going on and, make judgment calls and whatnot. and that's I feel like, Maybe more important now than looking at the code itself.Swyx [00:06:11]: Yes, because like, you can try to treat the code as a black box and then use, see the observable action that comes out of it, and then just prompt a change.What Modal Is For: AI Cloud PrimitivesAkshat [00:06:21]: Yeah.Swyx [00:06:22]: So I think it takes a bit of restraint to not specialize, to say, “I want to ship a new primitive,” and then just be general purpose.Swyx [00:06:31]: People ask you, “What are you for?” You're like, “ I don't know. We can do this, we can do that.”Vibhu [00:06:36]: Well, I'd be curious to see, like, okay, if we were to ask you, like, what is Modal for even at a high level? There's a lot you guys do, sandboxes, GPUs, everything. How do you answer?Akshat [00:06:46]: Modal is a cloud platform that's built for, where we've built the primitives from scratch for AI applications. and right now it covers, inference, training, batch processing, and sandbox workloads.Akshat [00:07:00]: But we're building a lot moreSwyx [00:07:02]: I noticed you didn't say web server, so there is still a role for, like, the always-on large-scale Kubernetes type things.Akshat [00:07:09]: Yeah, absolutely. We're, we're not trying to compete with the renders of the world, because yeah, we think the differentiator for us is the, are the workloads that need specialized compute, need to scale up and down a lot. yeah, they're, they're, they're just shaped differently.Working Alongside Frontier StartupsVibhu [00:07:26]: I think you're building a lot of it alongside the startups, right? They're innovating quite a bit, even in your, like, latest blog post. Like, even in the series C, the customers that you mention here, the cognitions, technical ones, ramps and whatnot, they're, they're innovating with you, right? And that's not something AWS is doing directly with.Akshat [00:07:45]: Yeah, absolutely. I think, this is again classic. We're a small team. We can move really fast. our engineers are working with our customers and figuring it out. Yeah.Swyx [00:07:54]: So my first week at Cognition, I walked in, there was someone wearing a Modal shirt. I was like, “What are you doing here?” They're like, “Yeah, I just. I am embedded inside of Cog.”Akshat [00:08:05]: Yeah, I think that was Peyton. We sent him overSwyx [00:08:07]: Yeah.Akshat [00:08:07]: Because, the latency of communication was too high otherwise.Swyx [00:08:12]: Yeah, distributed node, you have to - you have to place one and collocate.Vibhu [00:08:16]: Yeah.Swyx [00:08:16]: So I had a, I had direct personal experience, right? So I worked on smol developer three years ago. it was inspired by Claude 1. I think you onboarded me at some point, like, just before, and I was like, “Oh, like, I need some bursty compute. Like, I was just gonna try using Modal.” And it was a, it was a pretty pleasant experience. apparently, I showed up in the board meeting, like the analytics.smol developer, Sandboxes, and Proto-CognitionAkshat [00:08:39]: Yeah, you blew up on Hacker News and,Swyx [00:08:41]: YeahAkshat [00:08:41]: We got a big traffic spike. I. I think the way you used smol developer was Modal functions for running stuff, which was. Like, the, that was a good use case. but then, yeah.Swyx [00:08:53]: Yeah. That - So to me, that was proto-cognition.Akshat [00:08:55]: Right.Swyx [00:08:56]: If only I had, like, stuck to it.Swyx [00:08:58]: Like, that was like, if - did you say draw the tech treeAkshat [00:09:00]: AbsolutelySwyx [00:09:00]: You're just like, “Yeah, like, probably this will happen.”Akshat [00:09:02]: Yeah. Like, he was so close. You were just rebuilding upon usSwyx [00:09:04]: I just didn't realize.Akshat [00:09:05]: But the funny story there is at the same time, we were talking to a bunch of customers who needed something like sandboxing.Swyx [00:09:14]: Yeah.Akshat [00:09:14]: This is like twenty-three.Swyx [00:09:15]: Yeah.Akshat [00:09:16]: So we builtSwyx [00:09:17]: You introduced a new API right after that.Akshat [00:09:18]: Yeah.Swyx [00:09:19]: Yes.Akshat [00:09:19]: Like, we built sandboxes in May of twenty-three before anyone was even knew this was gonna be a thing. And the first example we published was, we took smol developerSwyx [00:09:28]: Smol developerAkshat [00:09:28]: And put it in a loop, so the agent can iterate on itself.Swyx [00:09:33]: Loops are hot these days.Vibhu [00:09:34]: It's the looper.Akshat [00:09:34]: Yeah.Vibhu [00:09:35]: Loops in. When was this, twenty-three?Akshat [00:09:38]: Yeah.Vibhu [00:09:39]: A small check.Akshat [00:09:39]: Yeah.Swyx [00:09:39]: It's like twenty-three. so the. the, those for listeners, like, the problem was the models are not built for any of this, right?Swyx [00:09:46]: Like, you're just trying to like. They're not post-training to understand, like, looping and, like, self-correction and tool calling was there, but, like, also not that great.Akshat [00:09:55]: Yeah.Akshat [00:09:55]: I don't remember if you used tool calling in this one, but yeah, the models would just diverge after like ten iterations and not produce anything meaningful.Swyx [00:10:03]: Yeah. But like, then. So okay, like now talking to myself three years ago, the answerVibhu [00:10:08]: Of course they will get betterSwyx [00:10:09]: Collect all the failures, build benchmark, and then collect all the, examples, build the RL environmentAkshat [00:10:15]: RightSwyx [00:10:15]: Sell it for like ten billion dollars to Meta.Swyx [00:10:17]: And then also train a model and then sell that for sixty billion dollars to Elon. And this isAkshat [00:10:23]: Yeah, of courseSwyx [00:10:23]: The funny machine. Like, it's like, it's about the hardware.Akshat [00:10:28]: It's hard to have that inherent conviction that the stuff will get that much better.Swyx [00:10:33]: In retrospect, it's so f*****g obvious.Akshat [00:10:36]: Fair enough.Swyx [00:10:37]: Like, what else were we doing back then? I don't know. anyway. Yeah. So this. That was the start of your sandboxing journey, right? I feel like it didn't blow up until, like, last year.Akshat [00:10:49]: Yeah.Swyx [00:10:50]: So there was like a couple years of quietness.Akshat [00:10:52]: Exactly, yeah. We wereVibhu [00:10:53]: I think very underrated product value. Like, my experience with Modal, Charles, before he had joined Modal, met this guy at a hackathon, and he really insisted we wanted to run some small model, not hosted anywhere, and he's like, “ there's this cool company, Modal. They'll like spin up a GPU sandbox, we can throw it on there. They'll take a Hugging Face link.” And like there's so much value just right there, right? Like instant hosting, spin it up, spin it down. It'll stay cold, but we run the demo a few days later, it'll come back up and like all this stuff in retrospect, like it's still what we needed like today.Akshat [00:11:27]: Yeah, it's still needed today. workload shapes have changed a lot as, we run stuff for people with really massive production scale and, there it's it's not about scaling from zero to one, but it's how do we scale really elastically, from like thousand to fifteen hundred GPUs very quickly in a given region. It's the same shape problem.Elastic Inference, GPU Autoscaling, and Custom ModelsVibhu [00:11:50]: Okay. So you look at, say, Cursor Composer, right?Akshat [00:11:53]: Yeah.Vibhu [00:11:53]: They had a. “We'll do RL on a model every couple hours.” you guys have a whole version of RL inference gym and whatnot.Vibhu [00:12:01]: When you look at workloads like that, you're doing train runs where you need to scale up, scale down every hour thousands of GPUs, right? That's the example for we do need it, right?Akshat [00:12:12]: Yeah. Well, so I'll, I'll take a step back and, maybe talk about like how people use Modal today. because our biggest use case is, elastic inference. And the thing we first found product market fit, with was inference for custom models. So we stayed away from the LLM space, and we were serving companies like Suno for audio, Runway for video, robotics, comp bio companies that train their own model elsewhere. But Modal is the best black box that for deployment, scaling to however many GPUs you need as your traffic pattern changes. And we saw all of them like have a very unpredict- predict- predictable, traffic pattern. it's like diurnal. It's Some days, like the company will do a launch and, they'll need like, way more. And it's not just one model that they deploy. They-- all these companies deploy, lots of different models in different regions, and so the autoscaling problem becomes even harder because then you have to scale within a certain region, and those cycles are offset. So different times you scale up in different regions.Akshat [00:13:20]: So that's like our sortVibhu [00:13:22]: And thatAkshat [00:13:22]: YeahVibhu [00:13:22]: That in and of itself is a huge category. There's a bunch of inference providers which, provide this fireworks, does this as a service together, whatnot, Base10. that's carved into its own niche for language models, at least right now.Akshat [00:13:36]: Yeah. the thing that we have specialized in is the autoscaling aspect.Vibhu [00:13:41]: Yeah.Akshat [00:13:41]: Because we found that it's not universally true that everyone else can autoscale, and we've gone deeper into it on the tech side by, we've incorporated GPU snapshotting into the product so we can take the GPU state, like your torch.compile model, snapshot it, and the next cold start is way faster. And so going back to your question, it's That's why you need a lot of burstiness for inference. But then people also do a lot of demand training, like for RL stuff, your rollouts are bursty, as you said. People also do a lot of batch jobs. So we'll see, a lot of companies, before they have a training run, they'll need thousands of GPUs to run encoding or something like that. And I think those things are much more bursty than. I agree that agents are not that bursty. sandboxes are, except when you're doing RL. RL is justRL, Batch Jobs, and 100,000 SandboxesVibhu [00:14:28]: Or commerceAkshat [00:14:28]: Insanely bursty.Vibhu [00:14:29]: Yeah.Akshat [00:14:30]: Yeah. Like when you're doing, rollouts, you sometimes need a hundred thousand sandboxes in your sandboxes.Vibhu [00:14:37]: Yeah. I'm curious if you've seen early sparks of continual learning. There are some people, like our friends, ngram, recently announced thisAkshat [00:14:45]: YeahVibhu [00:14:45]: They're, they're trying to do training. That also seems like a different workload, right? If you're doing training twenty-four/seven per se, there's a very weird dynamic of how you're using GPUs between people and whatnot, but seems like something you guys would work for.Akshat [00:15:00]: As you said, we're, we're fortunate to work with a number of, customers at the frontier and grab some of our customers. and they are taking the primitives we have, and trying to use them in very interesting ways, like continual learning. It's possible as the stuff gets better, some of that will be part of, our offering as well if, more people need it. but we're, we're just waiting to seeVibhu [00:15:23]: YeahAkshat [00:15:23]: How it shakes out.Vibhu [00:15:24]: Is there a primitive that you added after sandboxing that was the next step in the story?LLM Inference, DeFlash, and Speculative DecodingAkshat [00:15:32]: I guess we've been going much deeper into LLM inferenceVibhu [00:15:35]: YeahAkshat [00:15:35]: Because we realized that some of the advantages we have with like autoscaling, again, especially in different regions and whatnot, are, not present elsewhere. and the place where we had a gap was we weren't, working on the model layer itself. Like we were a black box. And, we realized that, we can get to frontier-level model performance, with, by having great people who work on this. And, we've been open sourcing a lot of our work, in terms of, Recently, we, shared our work on DeFlash, which is a block-based, speculator, and we've open sourced, all of it. So, you can - By using open source DeFlash, you can get the same performance as you would with one of the proprietary providers. And the next thing we're thinking about hereVibhu [00:16:23]: I thought this wasAkshat [00:16:24]: YeahVibhu [00:16:24]: An interesting blog post as well, right? Like, I think in here you make a claim that. Not a claim, just that how effective speculative deco-decoding really just get to.Akshat [00:16:33]: Yeah.Vibhu [00:16:33]: Anything you wanna point out from this around, what people should know?Akshat [00:16:39]: Yeah, absolutely. the high-level summary is, it would help to describe what speculative decoding is.Vibhu [00:16:44]: Yes.Akshat [00:16:44]: I will, yes.Vibhu [00:16:45]: I think, likeAkshat [00:16:46]: YeahVibhu [00:16:46]: So we've covered like Eagle and all thisAkshat [00:16:47]: YeahVibhu [00:16:47]: Like Hydra and all those things, but it was like two years ago.Akshat [00:16:51]: Yeah.Vibhu [00:16:51]: I think it doesn't hurt, right?Akshat [00:16:52]: Yeah. Speculative decoding is you have a smaller model, called a draft model, predict tokens ahead of the bigger model, and then you have the bigger model, verify all of this, all the tokens are predicted. And the reason it's faster is if you're predicting, one token at once, you're bound by memory bandwidth. But if you can batch the verification of, the draft model, then you're much more efficient using compute, and it's faster, and as long as your draft model is producing a lot of tokens that can get accepted, which is called the accept length, you can get a speed up that's, multiple times of, the original model speed. and well, that's what we highlight here. It's Like people talk a lot about we made these kernels faster and whatnot, but improving kernel will only give you like few percentage points of improvement, and, increasing accept length, literally is a multiplicative decreaseVibhu [00:17:47]: Like two to four X.Akshat [00:17:48]: Yeah, exactly.Vibhu [00:17:48]: Without much head-on performance.Akshat [00:17:50]: Yeah. I think it may - you are running a second model, right? So it may be something more expensive in the compute,Vibhu [00:17:57]: I meant quality performanceAkshat [00:17:58]: Probably not by muchVibhu [00:17:58]: But yeah. I thinkAkshat [00:17:59]: So there's no drop in quality performanceVibhu [00:18:01]: YeahAkshat [00:18:01]: Because you're always. You're never accepting a token that the big modelVibhu [00:18:04]: It's strictly betterAkshat [00:18:05]: YeahVibhu [00:18:05]: Or it's same.Akshat [00:18:06]: Exactly.Vibhu [00:18:07]: Right. Yeah.Akshat [00:18:08]: And so we've been working a bunch on DeFlash, which is a block-based speculator. so it's instead of predicting, one token at a time, it's predicting a block. And we've been open sourcing our work with it. The next thing for us here is for helping people train speculators and custom models. it's it's something that traditionally is very forward-deployed engineering driven, support deployed, engineer driven, like you work with customers and help them do that. And our vision for. This is why we launched Auto Endpoints, is we want to make frontier-level performance available to everyone. And so, we mentioned this in the announcement, we teased it. The next thing we're, we're launching is, as you run an auto endpoint, we shadow trafficAuto Endpoints and Frontier-Level PerformanceVibhu [00:18:54]: Do you want to explain what auto endpoints are?Akshat [00:18:57]: Yeah.Vibhu [00:18:57]: I lovely, yeah.Akshat [00:18:58]: Yeah. So, this is, I guess, going back to your Modal is you touch the code, but, sometimes people don't wanna touch the code, and they wanna get started with an endpoint that works and has all the great performance and, scalability that Modal has. So we've made that easier with, a way to create an endpoint from our UI, from the CLI, that has all of our optimizations that we talked about, like the DeFlash stuff already baked in, and there's full transparency. So we give you the code, you can go run it yourself, and if you want, you can eject out into the full Modal experience, which we see as people get sophisticated, they do wanna tweak the models, they wanna, fine-tune stuff. You can still do all of that. It's it's not a black box. And yeah, the next thing, as we teased later in the post, is how do we give you value even beyond this in terms of having your draft models evolve as your data distribution evolves, again, without having to talk to a person and, yeah.Vibhu [00:19:59]: I guess just to understand it directly, you have the GPUs, you have an endpoint that's compatible, you serve open model. If someone was to do this themselves, what's the delta that you guys provide? So you do a lot of open source great work on effective inference. how does it compare to, say, I take the same model, 5.2 FP8, take shelf inference engine, vLLM, SGLang, get compute of similar capacity, similar cost. What's the delta that plugging into something this, like this offers outside of the benefit of, scaling?Production Inference Beyond Raw GPUsAkshat [00:20:34]: It's interesting because we've taken the approach of open sourcing our contributions and upstreaming them. we work closely with the SGLang team. We want the improvements that our team, comes up with to be, there in open source for others to use, even outside of Modal. The benefit to us is we have a team that has significant expertise in terms of if you do have something that is not there, our team can help you get that performance, first. the other thing is with these endpoints, we are way more elastic, as you said, than, anyone else, and you have true scaling to zero. you have true, burstiness, and in practice, that matters a lot more to people than just finding, the GPU and, running Modal code on something.Vibhu [00:21:20]: Yeah. And I will say it's not that straightforward to just. like what I said is easier said than done, right?Akshat [00:21:26]: Yeah.Vibhu [00:21:27]: It's I think still for the average person, still hard to just gut check using different. There's, there's quite a bit of combinations you can make there. the trade-offs aren't really known at face value.Akshat [00:21:40]: Yeah. it's it's not just that. I think it's it's that running production-grade inference is a hard infer problem.Vibhu [00:21:49]: YeahAkshat [00:21:49]: Even if you subtract out the autoscalingVibhu [00:21:50]: YeahAkshat [00:21:51]: Is controlling things like tail latency and, making sure every, request is delivered at least once and whatnot.The Model and Agent LifecycleVibhu [00:22:00]: There's a lot of innovation that you can do here. I think, it's very interesting that you're starting to encroach on, like as you become a full cloud, you're starting to encroach on other people's turf.Vibhu [00:22:09]: What will you not do?Akshat [00:22:13]: Well, we wanna follow our users and, make sure they get like a platform that has everything that works well together. so right now we're focused on the model lifecycle and the agent, lifecycle. so both like going from data prep to training to inference, and then also if I want to deploy a background agent, let's say, sandbox, do persistent storage, a whole bunch of other stuff.Vibhu [00:22:38]: We talked to Cole, who did, OpenInspect. Yeah.Akshat [00:22:42]: Yeah.Vibhu [00:22:42]: And RealInspect also is on Modal.Akshat [00:22:44]: Yeah. So Ramp Inspect was a great example of a background agent that was really successful because they, were able to use some of the primitives like snapshotting and fast scaling to just have something that feels really reactive and works well.Ramp Inspect and Background AgentsVibhu [00:23:02]: Yeah. That's the new CTO of, Ramp right there.Akshat [00:23:05]: Yeah, Rahul.Vibhu [00:23:08]: It was really fun. yeah, okay, I think, all very bullish. Like, one of my reflections was also I did not originally. So when I met you guysThe Inference Inflection: CPU, GPU, and Co-LocationVibhu [00:23:19]: You weren't that much in the GPU game, and now you're all about, inference. And one of the points that I hinged on for Jensen's keynote at GTC this year was, what we're calling like the inference inflection, right? That let's say in AI workloads or machine learning workloads, it used to be like, let's call it eight to one GPU to CPU, and now it's more like one to one, which is like a interesting. Like, - because of how much agents are blocked or call out to this, to CPU heavy stuff the actual, like, limiting factor, like, swings back and forth from GPU to CPU a lot more than it used to be all GPU and then occasional CPU.Akshat [00:24:01]: Yeah.Vibhu [00:24:02]: GPU, CPU. And now it's like just constantly, and you just have to locate everything.Seventeen Clouds and the Supercloud StrategyAkshat [00:24:08]: Yeah. And that's one of the things that, again, we see as, something appealing about Modal, which is we've built this capacity pool that spans, 17 cloud providers, so we're, we're very good at Running on various kinds of cloud capacity across the worldSwyx [00:24:24]: You don't have your own data centers?Akshat [00:24:25]: We don't have our own data centers. We just run across a lot of neo cloudsSwyx [00:24:29]: Yeah. AreAkshat [00:24:30]: Metal providers.Swyx [00:24:30]: Yeah. Question mark.Swyx [00:24:31]: Yeah. You're, you're running the math, and you're like, “What's the cutover point where you're like.”Akshat [00:24:36]: Yeah, it's a good question. part of it is we see our differentiator in the software layer, and, being capital light and focusing on the software helps us move really fast. so far it's worked out well because there are so many other people building data centers that we're able to work effectively with them, and again, focus on what makes us, special.Swyx [00:24:55]: Yeah.Swyx [00:24:56]: 17 gets you into, like, the local providers sometimes. LikeAkshat [00:25:00]: The,Swyx [00:25:01]: Which was the most interesting one?Akshat [00:25:02]: There are a lot more neo clouds than you expect, and they all have various degrees of, various levels of reliability. And, that's why it's something we've invested a lot of time in, is building our own reliability layer on top. so if the GPU falls off the bus or something happens, we user workloads are not affected, and that lets us use a lot more capacity than,Swyx [00:25:30]: YeahAkshat [00:25:30]: You as a user would be able to.Swyx [00:25:32]: It's a useful thing to have because like now everyone knows, like, what layer you are and, like, you optimize for being the super cloud of all clouds.Akshat [00:25:41]: Yeah. That's, that's, that's the idea. and so I guess when you mentioned colocation, that's, that's another interesting thing where, one thing we've seen is people come to us when they want, very specifically located, CPUs or GPUs, like they wantSwyx [00:25:57]: Oh, they pin it in likeAkshat [00:25:58]: YeahSwyx [00:25:58]: EU?Akshat [00:25:59]: Exactly. Or EU, US.Swyx [00:26:01]: Right. Data resiliencyAkshat [00:26:02]: AustraliaSwyx [00:26:02]: Locality thing or performance or what?Akshat [00:26:04]: It's either data locality or latency, yeah.Swyx [00:26:07]: Yeah.Akshat [00:26:07]: Like, you want your. They're running sandboxes and model. They want them to be right next to aSwyx [00:26:10]: Yeah, it's easy thenAkshat [00:26:11]: YeahSwyx [00:26:12]: To. That is important in all those things. and so, like, you've accidentally, I don't know if it's accident, but, like, you've built the perfect primitive for agents to express themselves. And then, like, it's almost very funny how every extra development just involves more file system, just involves more CPU.Akshat [00:26:30]: Yeah.Swyx [00:26:31]: Just like the things that you already have. I don't know much about, if there's any, like, networking usages that are interesting, but you've also done some good work on networking.Networking, Sidecars, Private IPv6, and SandboxesAkshat [00:26:40]: Yeah, that's exactly right. Like, we're just taking compute storage and networking and building stuff on that layer, for, again, the stuff people need.Swyx [00:26:49]: YeahAkshat [00:26:50]: We see a few interesting networking things coming up. one is people want networked sandboxes. so we haveSwyx [00:26:57]: For like a Docker cluster type thing.Akshat [00:26:59]: Yeah.Swyx [00:26:59]: Sorry, Docker Swarm. Oh, f**k. What is it called?Akshat [00:27:02]: Compose.Swyx [00:27:03]: Compose type thing.Akshat [00:27:04]: Yeah. So if you want Docker Compose, our sandboxes now support, this thing called sidecars. So you can. A sandbox is a pod of containers, and you can run multiple containers in, a sandbox. also useful because, going back to networking, people want a lot of control over, outbound networking from a sandbox.Swyx [00:27:23]: Yeah.Akshat [00:27:23]: Like, they might wanna run a middle proxy for, like, maybe logging stuff for RL or, controlling how egress can happen to a domain, injecting credentials. and yeah. So we've, we've had to build a lot of that stuff ourselves.Swyx [00:27:38]: Yeah.Akshat [00:27:39]: But then also sometimes people want, sandboxes spanning multiple nodes to talk to each other, which is an emerging thing we're seeing. We have support for that for a different reason, and yeah, we'll see if that becomes stable.Swyx [00:27:52]: Like, just an open socket. It's a. This is directly like mTLS.Akshat [00:27:56]: We do support that, which is you can, expose a tunnel inside a sandbox.Swyx [00:28:01]: Yeah.Akshat [00:28:01]: And then you can either expose it to public internet or it can be, you can add like a HTTP, auth layer above it. But we have this thing called I6PN, which we haven't talked about, which is this, like, overlay network using IPv6 addresses. so if Modal containers, within the same workspace, when this is enabled, can address each other using this private IPv6 address, and no one else can.Akshat [00:28:28]: So it's like private networking, for containers. We built it because we needed it as a primitive for our distributed training product. so we have this other feature, which is you can add a decorator to a function, and you get a cluster of GPUs. and they have RDMA networking. so you can run a distributed training job, that's truly serverless. and we did the overlay network for that. But then we've seen that people are using it for other reasons, and, I'm intrigued to yeah, what would people do with it.Swyx [00:28:59]: Build primitives and let people figure it out, right?Akshat [00:29:01]: Yeah, exactly.Swyx [00:29:02]: You put out a pretty interestingAkshat [00:29:03]: They're like, they read the docs webpage. Let me use thatSwyx [00:29:06]: YeahAkshat [00:29:06]: Something they never intended to work. This is literally not even in our docs page. People somehow found it, and they're using it.RDMA, Memory Movement, and Distributed TrainingSwyx [00:29:12]: Huh.Swyx [00:29:14]: The way you portrayed it with, like, RDMA versus TCP, like, very well laid out, but just the transfer speed change at scale for RL, like yeah, you have it, you have it built in. I'm sure someone found it. It's found it to be a lot more efficient before you made a thing out of it, right?Akshat [00:29:32]: Yeah. And not to split hairs, I guess the overlay network is the TCP overlay network.Akshat [00:29:39]: The reason we have that is you need that to do the key exchange for RDMA before you set up the RDMA network on top of that. but then people found the TCP part.Swyx [00:29:48]: Can I tell you, this is like a big aha moment for me becauseAkshat [00:29:51]: YeahSwyx [00:29:51]: So I review 2,200 submissions for the World's Fair.Akshat [00:29:56]: Yeah.Swyx [00:29:57]: And then I got this from John OsterhoutAkshat [00:29:58]: HuhSwyx [00:29:59]: Who I don't know if. Do John Osterhout by name?Akshat [00:30:01]: The name sounds familiar.Swyx [00:30:02]: He published a. He's a well-known professor, published a lot of interesting software design books, and this is the talk he chose to submit, is on RDMA at Inference. And I'm like, you wouldn't think that this guy, who is like operating systems guy, would care about RDMA.Akshat [00:30:20]: I, it makes sense to me because I,Swyx [00:30:24]: This is the cloud, right? YeahAkshat [00:30:25]: Like, the way you move around your KV cache and how efficiently you can do it, how efficiently you move, your weights from your training GPUs to your inference GPUs in RL is there's a lot of degrees of freedom, and it is a systems problemSwyx [00:30:41]: YeahAkshat [00:30:41]: Moving memory aroundSwyx [00:30:42]: YeahAkshat [00:30:43]: Scheduling.Swyx [00:30:44]: This shows you how primitive my understanding of networking stuff is.Swyx [00:30:46]: Is this like the domain of WireGuard as well?Akshat [00:30:50]: Not quite.Swyx [00:30:51]: It's adjacent?Swyx [00:30:53]: Explain everything.Akshat [00:30:54]: Sure.Swyx [00:30:56]: How do we move memory around GPUs?Akshat [00:30:58]: Well, so sorry. Yeah, that is memory. Sorry, I was talking more, and maybe I was talking like five minutes back, about the private IPv6, addressing that you've set up.Swyx [00:31:09]: Yeah.Akshat [00:31:09]: Is it like it's a VPN?Swyx [00:31:10]: Yeah, it is like a VPN, and yeah, WireGuard is, yeah, you're right. It is,Akshat [00:31:16]: Right. Yeah, you already moved on to new topicsSwyx [00:31:17]: A similarAkshat [00:31:18]: OkaySwyx [00:31:19]: In the same space, WireGuard is, encrypted and this is,Akshat [00:31:23]: And you don't need encryption.Swyx [00:31:23]: Yeah.Akshat [00:31:24]: Yeah.Swyx [00:31:24]: This is not encrypted. that's the main difference. This is TCP and we have eBPF programs that will reject or allow the TCP connection based on whether you're allowed to do it.Akshat [00:31:35]: Used to involve a full sidecar, but now you have eBPF in the Linux kernel.Swyx [00:31:39]: Yeah.Akshat [00:31:40]: Yeah. I don't know if this is a natural follow-on to the topic of like my skepticism on distributed training is that while, like, people spend a lot of money on, like, cables to hook up GPUs, and even that is not, like, fast enough, and that's the bottleneck, is your networking fast enough?Swyx [00:31:59]: Yeah. So I guess you're talking about fully distributed training like, Dialog or something which is like cross data centerAkshat [00:32:06]: That would be, yes.Swyx [00:32:07]: That's the extreme.Akshat [00:32:08]: Yeah.Swyx [00:32:08]: You're in the middle, and then other people would have like the Mellanox cables up in, like, their actual data center.Akshat [00:32:14]: When you run multi-node training on Modal, RDMA, I think Mellanox, is, or InfiniBand is like a, is all seen as RDMA. but it's a way to bypass the TCP networking stack and, transfer, stuff much faster, between one node, to the other. And we have I think like 3 terabit per second, internal networkingSwyx [00:32:40]: OkayAkshat [00:32:40]: Which is the standard that's needed.Swyx [00:32:42]: Okay. So I misunderstood whatAkshat [00:32:43]: 50Swyx [00:32:43]: What part of the stack you wereAkshat [00:32:44]: 50 gigs overSwyx [00:32:45]: YeahAkshat [00:32:45]: If you wentSwyx [00:32:45]: YeahAkshat [00:32:46]: RDMA.Swyx [00:32:46]: Okay.Swyx [00:32:48]: Yeah. I, very impressive work.Multi-Node Training, Post-Training, and Auto ResearchSwyx [00:32:52]: So effectively you're extending like the model philosophy to the training cluster, like, yeah.Akshat [00:32:59]: Yeah. And we're, we're not going for like large scale training runs. the thing that we've built multi-node training for is, we see a lot of, smaller scale post-training. like, people are post-training like medium sized fund models, so they can, get higher quality on inference. this is a perfect fit, for something like that.Swyx [00:33:21]: Yeah. That is my impression of how a lot of these labs explore branches in post-training and then eventually merge whatever they find in.Akshat [00:33:31]: Yeah. The other use case we've seen for multi-node training is even if you have a big cluster, your researchers are still doing small runsSwyx [00:33:38]: YesAkshat [00:33:39]: Having elasticity thereSwyx [00:33:40]: Right, sureAkshat [00:33:40]: Matters a lot more.Swyx [00:33:41]: Yeah. the, like, this is like the current limiting factor for auto research, which is like you need to give your model some GPUs in order for it to completely run.Akshat [00:33:51]: We have a blog post on auto resource and model is,Swyx [00:33:55]: YeahAkshat [00:33:56]: Yeah, like, turns out to be pretty good substrate for that.Swyx [00:33:59]: So my impression is auto research means many things, likeAkshat [00:34:01]: YeahSwyx [00:34:01]: Anything that Andrej coins. Right now it's still science fair, right? Like not like, I don't know how many people are doing this.Akshat [00:34:08]: We're having a golf.Swyx [00:34:08]: Yeah.Akshat [00:34:09]: I thought the same thing.Swyx [00:34:11]: Yeah, you would know.Akshat [00:34:12]: We, like, our internal both training and inference teams use this the general shape of this quite a bit. like we have this one internal repo called auto inference, which essentially we've automated our own forward-deployed engineering efforts using, this harness, which is, the agent will just spin up a sweep of different things. It'll even run like, NVIDIA inside profiler and it'll like tweak configs and it'll arrive the right thing. it'll change your GPUs both from H200 to B200, and works really well.Swyx [00:34:47]: Nice.Akshat [00:34:47]: So yeah.Swyx [00:34:48]: By the way, I enjoy that your forward-deployed engineering is so technical that you have to do these things.Swyx [00:34:52]: It's very different from forward-deployed engineering from other people.Akshat [00:34:54]: Yeah. For our forward-deployed engineering team is, essentially they're like applied inference researchers or applied training researchers.Swyx [00:35:02]: Someone told me like they have to be able to build, but they also have to be able to sell. do they have to sell or are they like they're good, they're just like post-sale type of thing?Akshat [00:35:09]: It does, being able to talk to a customer and engage effectively with themSwyx [00:35:13]: YeahAkshat [00:35:13]: Matters a lot.Swyx [00:35:14]: They want the same thing.Akshat [00:35:15]: Yeah.Swyx [00:35:15]: ?Akshat [00:35:15]: But it's it's not really a sales, thing. We pair them with-- We have solution architects as well that are more on the sales side.Swyx [00:35:23]: Okay. Let's spend a bit more time on auto research. This is a big focus for for this year. Where does this go? like, have people explored enough? Like, there's all these beautiful charts of like improve and then level off a bit and then you find the next thing. Is this one abstraction up from normal training? Is that how we think about it, or do you think about it differently? Like model level training versus high, like driven hyperparameter search.Auto Inference and Modal BenchAkshat [00:35:51]: Yeah, like,Swyx [00:35:51]: Someone, some people call it like neural architecture search or whatever, right? Like.Akshat [00:35:54]: Yeah, - So the stuff I've seen people do with it is nowhere on the architecture level. It's pretty much tweaking parameters, but it's it's a hyperparameter sweep that's guided by some model intuition, so it's like much more efficient than, whatever other, sweep you would have.Swyx [00:36:12]: Yeah, it's just, it's just a question of where you want to spend your compute?Akshat [00:36:16]: Right.Swyx [00:36:16]: ‘Cause yeah, you can just throw infinite amounts of money on this and somehow you'll bang out Shakespeare?Akshat [00:36:22]: Yeah, infinite monkey.Swyx [00:36:24]: Yeah, so like the very good for model. and I think it's also very important that agents can spin up other agents, can spin up their infrastructure. Like very good for you. how good is our LLMs at generating model code? Like the benefit of existing LLMs is that you are in the data.Akshat [00:36:42]: Yeah. They're, they're surprisingly good. I think like pre Cloud 4 they were not, and then now they're able to shot, stuff out of the box. But we're playing around with releasing like a Modal Bench for like the harderSwyx [00:36:55]: YeahAkshat [00:36:55]: Things, that the LLMs cannot do yet and maybeSwyx [00:36:59]: What's an example of that?Akshat [00:37:01]: I think the things that- Sometimes agents struggle with, without right guidance and a skill is, how to, use the rest of our observability. Like how to. Something is failing, like how do you look at the logs and then update the right thing? It's reasoning about that. But they're able to shot, likeSwyx [00:37:23]: Yeah. You can just add a skill to it?Compute Strategy and Capacity PlanningAkshat [00:37:26]: Yeah. So we have a Modal skill now that. Which is why we built this Modal Bench. It's to find things like that, so we can address them in our tool.Swyx [00:37:35]: Tune a skill. Yeah.Akshat [00:37:36]: Yeah.Swyx [00:37:36]: No. it's it's good. are you facing any shortages? like we talk a lot about GPU shortages, but also CPU, also memory.Swyx [00:37:44]: Yeah.Akshat [00:37:45]: We have had a lot of growth, which means that, there's - we've had to be much better aboutSwyx [00:37:53]: PlanningAkshat [00:37:54]: Proactive capacity planning.Swyx [00:37:55]: Yeah.Akshat [00:37:55]: So we have,Swyx [00:37:57]: Which by the way, like it's like a MBA's like dreamAkshat [00:38:00]: YesSwyx [00:38:00]: Is like just planning this stuff. I think last time you and I talked about something maybe about this.Akshat [00:38:03]: Yeah. we have a really competent team of people that we call, The role is called compute strategy. so yeah, if anyone listening here or wants to work on thatSwyx [00:38:13]: Compute strategy?Akshat [00:38:13]: Yeah.Swyx [00:38:14]: I think,Akshat [00:38:14]: I feel like,Swyx [00:38:15]: I think the normies call it FP&A or something.Akshat [00:38:18]: Well, it's more It's it's not FP&A. It's it's There's a lot of interesting financial questions of like what is the blend between one year and three-year reservations? how do we forecast our own capacity? how do we. especially since our capacity is very fungible across different GPU types and different regions, like you have to model a lot of it. and you also have to have an opinion on how the supply chain is gonna evolve, and then you have to like, take bets,Swyx [00:38:49]: YeahAkshat [00:38:49]: Based on that.Swyx [00:38:50]: Tokenomics.Akshat [00:38:50]: Yeah.Swyx [00:38:51]: This is like probably a not a real point, but, I was trying to think about like what other industries. I was trying to think about like, we cannot be first to like these kinds of problems.Akshat [00:38:59]: Yeah.Swyx [00:39:00]: And what other industries have had this? And I was like, airlines with fuel and like they have to hedge their fuel and like, I think for a long time Southwest because they made like a hero fuel bet, they like were like super low cost becauseAkshat [00:39:12]: OhSwyx [00:39:12]: Compared to everyone else.Akshat [00:39:14]: Yeah. I hadn't thought about that.Vibhu [00:39:16]: We're at a fun time too?Akshat [00:39:18]: Yeah. It's. A lot of the compute business in general, for us is also about being very good about capacity management. That is how you have great unit, economics. but also over time it's how you can unlock more value for customers. Like, one of the things we're building now is like a way for customers to get, If they don't care about latency, like get much cheaper pricing and they'll get results back in like next 24 hours or something, like a batch tier essentially.Batch Tiers and Latency-Insensitive WorkloadsSwyx [00:39:47]: Yeah.Akshat [00:39:47]: And those are levers we have because we control the whole stack and scheduling and whatnot to give people a sufficientSwyx [00:39:53]: Yeah. I feel like they're not as popular. Like those, like the Frontier Labs have all those APIs. They're not as popular as they should be.Akshat [00:40:00]: The demand that we see for something like that is not for LLMs. although sometimes people wanna run evals andSwyx [00:40:08]: OkayAkshat [00:40:08]: Synthetic data prep and there it makes sense.Swyx [00:40:10]: Okay.Akshat [00:40:11]: But it's from a lot of LLM companies, like people who are doing computational bio, like they have to run really big batch jobs and they don't care about when they get it back.Swyx [00:40:22]: Yeah. And like they have a reasonable. It's it's also like a cousin to the stopping problem of like, will this finish in time?Akshat [00:40:30]: Yeah. You can bound it.Swyx [00:40:33]: Yeah.Akshat [00:40:33]: Like you can give peopleSwyx [00:40:34]: YeahAkshat [00:40:34]: SLAs on it.Swyx [00:40:35]: Yeah. I think what's, what's interesting is like the next phase of model.Swyx [00:40:38]: Like what, do people expect from you, now that you're established and you're like well-known compute player among all these leading companies. You had an inference launch week, and we talked a little bit about the launches. like what else? Like what else should people know?What Modal Builds NextAkshat [00:40:55]: We are building primitives that make our users' lives much easier. So, I think for example, with LLM inference, thousands more companies are gonna post-train their own models and, deploy open source models for inference. so we're thinking a lot about what is the best product shape for that. And, that involves everything from our training gym to, then, endpoints that get frontier-level performance. again, but I haven't talked to anyone. It looks somewhat different on other verticals. Like, we're also seeing a lot of real-time, audio-video stuff in there, which is why like, we're working on things like regional routing, with fallbacks. So you can get GPUs that are as close to users as possible. so you get like low latency for video streaming and whatnot. And then on the agent side, it's,Akshat [00:41:52]: We're still working very closely with our customers because stuff is changing so fast in terms of what they need. And, I think beyond sandboxes and persistent file systems, there's a lot of other things people will need from this agent stack as they build production agents. So yeah, we're thinking about those other things that fit in there.Swyx [00:42:13]: I want to ask what the other things are.Akshat [00:42:15]: Yeah. I probably should share right now.Swyx [00:42:17]: I think-- I think, okay, so, I do think a lot about the principal components of cloud, and you do talk about compute storage networking.Akshat [00:42:25]: Yeah.Swyx [00:42:25]: Because so far for me, it's fine. so far for the. the first couple generations of cloud, it's fine. What's different, qualitatively different about agents that you need some new permission level? Like a lot of people, okay, and I'll just kinda spew tokens at you until it like hopefully sparks something.Akshat [00:42:43]: Yeah.Swyx [00:42:44]: Like the new level now is whatever Claude Code does, which is dangerously scope permissions or like allow list by command or like whatever, right? And sometimes they're like, “Well, okay, we have like this adaptive thinking mode where like, just trust me, bro. I will make the calls for you.” Is that it? like mediated permissions.Hard Guardrails vs. LLM-Mediated PermissionsVibhu [00:43:03]: Now you're looping it with a goal and letting it roll.Akshat [00:43:06]: Yeah, I'm, I'm skeptical of LLM media permission for stuff that is at the sandbox level because you do want hard boundaries.Swyx [00:43:16]: Yeah.Akshat [00:43:16]: Otherwise, someone can exfiltrate stuff.Swyx [00:43:20]: But likeAkshat [00:43:20]: YeahSwyx [00:43:20]: Maybe that's old school thinking. Maybe we're the dinosaurs.Swyx [00:43:23]: Maybe the AI OS or the LLM OS is really the kernel is a goddamn LLM.Swyx [00:43:30]: Like it makes you feel uncomfortable.Akshat [00:43:31]: Yeah, I'm, I'm toldSwyx [00:43:32]: But that's what trusting the LLM is. Like imagine a spherical cow perfect LLM.Akshat [00:43:36]: Right.Swyx [00:43:37]: That it.Akshat [00:43:39]: Maybe.Swyx [00:43:41]: I wanna test the boundaries, right?Akshat [00:43:42]: Yeah.Swyx [00:43:42]: Like, and I don't believe that, but I wanna see where I'm wrong ‘cause that's, that's the consensus.Akshat [00:43:49]: Yeah. I think you always need hard guardrails when you want, And you can pair those with softer guardrails, right? And that's gonna be a lot of mediated.Managed Agents and Specialized SandboxesSwyx [00:44:00]: There. I'll also get you a end with a couple of your commentary on like the ecosystem outside of Modal. Manage agents. Everyone has one. Gemini, OpenAI, Claude, very useful for you, but also like it is their way of starting to edge into your space.Akshat [00:44:17]: Yeah.Swyx [00:44:17]: What's going on?Akshat [00:44:19]: Yeah, we're, very excited to partner with Anthropic and some of the other foundation labs, will not name who we're also working with. the way we see it is the manage agent thing is a great place to start if you're starting out building an agent and, But then when you get to, building something more production grade, like you're a company that's like Ramp that's building their own, Ramp also runs their accounting agent on us, so their external-facing agent. You need a lot more control over, your compute primitive on things like, what sort - how do you persist different files that the agent has access to, and how do you snapshot and restore? How do you control the networking? maybe you want GPUs. When you get to that point, you kinda want, a specialized sandbox provider, that gives you those things, and that's the role that we are trying to play.Swyx [00:45:15]: YeahAkshat [00:45:16]: We don't really have an opinion on the harness, whether it runs - it's a cloud-managed agent, and you hook it up to Model Sandbox, or you run the harness in Model Sandbox. We'll see where people converge with that.Swyx [00:45:26]: Yeah. Do you any opinions on like the meta harnesses, or just another layer on top of these things?Akshat [00:45:31]: You mean like the OpenPipeSwyx [00:45:33]: OpenPipe is one. I think Vercel had one, which I can't remember the name of right now. Fredshot had one. and then, to me, most recently was Data Databricks that had Omnigen. All these are meta harness. Like it's kinda pseudo agent cloud type things.Akshat [00:45:50]: I personally have not played around with them.Swyx [00:45:53]: Yeah.Akshat [00:45:53]: Build agents with them.Swyx [00:45:54]: Everything's bullish Modal, as long as it consumes more infra.Akshat [00:45:57]: That's why we're focusing on the infra layer. It's somewhere where our, relative competence is and, also it's a hard problem to solve.Swyx [00:46:06]: Yeah. I will say like just generally reflecting on that, I don't know if - if there's other topics on Modal, but like just generally reflecting as an infra person, not as intense as you, but in that field, this has like been the most exciting time in infra. Like it was boring for a while, and you couldn't really get people excited about data infrastructure. Like Eric would get on Data Console, everyone just watched the video and like say, “Look at how many sandboxes I can spin up,” and no one gave a crap.Why Infrastructure Became Exciting AgainAkshat [00:46:39]: Yeah.Swyx [00:46:40]: And like now everyone gives a crap.Akshat [00:46:42]: That's true. It is a very exciting time, and I think a lot of that's driven by just the amount of scale all of this stuff needs.Swyx [00:46:50]: I think the, like a lot of your initiatives or a lot of your like product directions make sense in retrospect, which is like the best kind, but I wouldn't necessarily have thought about it myself, which.Akshat [00:47:00]: We need the predictions.Swyx [00:47:02]: I think there's a lot that you just don't even see, right? Like you have the batch, you have the voice, you have the multimodal, but what else?Akshat [00:47:10]: What else is coming up for usSwyx [00:47:11]: Yeah. Where do you see things going?Akshat [00:47:13]: Yeah. I, in generalBiotech, Robotics, and Non-LLM AI WorkloadsAkshat [00:47:15]: It's it's clear that there's there's a huge shift happening. I think one thing that's not as obvious to people because LLM inference gets talked about so much and is also we work a lot of companies that are, doing things like drug discovery and computational bio, like the Chai Discoveries of the world. Big things are probably gonna happen there. we work a lot of robotics companies that are putting robots in like active deployments and getting good results out of them.Swyx [00:47:45]: Is there Air Gap Modal? Is there a version that is like prem air gapped whatever?Akshat [00:47:50]: No. We,Swyx [00:47:51]: You should cloud only.Akshat [00:47:51]: Yeah.Swyx [00:47:52]: Yeah. Okay. But yeah, so what you're saying is like because you're focused on primitives and they're good primitives, you find use cases in all these kinds of things.Akshat [00:48:01]: Yeah.Swyx [00:48:01]: Probably diversifies you a little bit away from LMS all the time.Akshat [00:48:05]: Yeah, absolutely. We're, we'- our goal isn't to only serve the LLM inference market.Swyx [00:48:10]: There are a lot just on the website, the audio,Akshat [00:48:12]: Yeah. We said both onSwyx [00:48:14]: Computational bio images. Yeah, there's a lot here. There's QTA TTS, customizing. Oh, Chatterbox. there was customizing Whisper.Akshat [00:48:24]: Okay. Yeah.Swyx [00:48:25]: This screen reminds me of a fallen competitor, which Replicate.Model APIs vs. Differentiated AI ProductsSwyx [00:48:31]: What's your postmortem on what happened?Akshat [00:48:34]: This is one thing we've stayed away from is providing an API for models because I think providing model APIs is some of it ends up serving like a really hobbyist market, which is much less sticky.Swyx [00:48:50]: Yeah.Akshat [00:48:50]: And we've always wanted to build for companies that are building products and need more flexibility that's not just an API.Swyx [00:48:57]: Which you can build an API for a model and this is clearly what it is. But you - but what you're saying, you can wrap it into a more fully functioning back end that you run.Akshat [00:49:06]: Yeah. So all of our examples, it's not that spin up this model, here's an API token, use it. They're all code.Swyx [00:49:13]: Okay.Akshat [00:49:13]: And so the point is that this is just an example.Swyx [00:49:16]: Starter code.Akshat [00:49:17]: Yeah. But you can tweak it however you want.Swyx [00:49:20]: Yeah.Akshat [00:49:21]: And if you're like a company building a product, like, computational bio whatnot, yeah.Swyx [00:49:26]: I guess I'm trying to tease out for listenersAkshat [00:49:28]: YeahSwyx [00:49:28]: When does it stop becoming, oh, you're just an API call and you're just a wrapper on API to becoming what you call a product, right?Swyx [00:49:36]: Like, what is that layer? Like what-- Like, more lines of code, but like beyond that, what is the substance that people add that qualifies it to be something more?Akshat [00:49:46]: I think there's a little bit of like a selection effect of like a lot of the companies who do wanna get deeper into that level are probably building something that's more differentiated. And, I think, an example is like - with LLM inference, originally we, worked with companies that were building their own post-training frameworks or they were, - Ramp early in the day was training their own tokenizer and like swapping out the tokenizer in Llama and whatnot. I'm not saying that's, that successful, in that case. But a better example is like, let's say Suno. because Suno, does not use Modal for training.Swyx [00:50:26]: Mikey on the pod. Yeah.Akshat [00:50:27]: But they use Modal for all their inference and that's because they have like a custom-- They have completely custom model architecture and that means that they have to be at the code level and tweak things that are not, just an API.Swyx [00:50:41]: It's interesting as well, like we had, Ethan, most recently on the xAI Groq team make a prediction that like the next tier in video gen is not a better video model, it's a better model or agent that orchestrates video models.Video Agents and Production WorkflowsAkshat [00:50:56]: Oh, interesting.Vibhu [00:50:56]: Language model backbone that can use toolsAkshat [00:50:58]: RightVibhu [00:50:59]: And write code.Akshat [00:51:00]: Like, yes, I can make my second video or my second video from Groq, but I want my minute video.Akshat [00:51:06]: And I'm not going there through normal video gen.Swyx [00:51:10]: Yeah, that's interesting. I - So we have GPU sandboxes and recently have seen a few companies doing agents that do video manipulation or,Akshat [00:51:22]: Yeah. Give it FFmpeg and just do it.Swyx [00:51:23]: Run FFmpeg. But likeAkshat [00:51:25]: That's not enough.Swyx [00:51:25]: Yeah.Akshat [00:51:26]: You need to give it Adobe.Swyx [00:51:27]: Yeah, I hadn't put it together with like it would be a video production thing. in my mind these things were going more towards editingAkshat [00:51:36]: Yeah.Vibhu [00:51:36]: Well, shout out Mantis.Akshat [00:51:37]: I think about this a lot.Swyx [00:51:38]: .Akshat [00:51:41]: Yeah. Sorry.Vibhu [00:51:41]: Luma. Luma Agent is a version of this for video production, but it's a off.Swyx [00:51:46]: I was gonna get your quick takes, on some other stuff that happensGitpod/Ona, CI, and Runtime SandboxesSwyx [00:51:50]: In recent news and just-just see if you have anything interesting. Gitpod, very li

.entry-img img{ display:none !important; } .single .hentry .entry-img{ display:none !important; } https://open.spotify.com/episode/6raW3lf3gJuwTrYNbdkf0F Too many organisations are pouring time and money into AI only to find that the promised efficiency gains and cost savings never materialise, leaving CFOs struggling to justify the investment. Understanding why most AI projects fail to deliver ROI, and what finance leaders can do differently, is now a critical skill for anyone responsible for steering strategy, systems, and spend. In this GrowCFO Show episode, host Kevin Appleby sits down with Sinohe Terrero, CFO and COO of Envoy, to explore why so many AI initiatives fall short and how finance leaders can change the outcome. Drawing on his experience as a serial startup CFO and operator in high-growth tech companies, Sinohe reframes AI as a practical toolkit for augmentation, task automation, and application development, and explains how confusion between these use cases leads to poor deployment and weak returns. Throughout the conversation, Sinohe shares real examples from Envoy's finance function, from AI-powered reconciliations and automated interview workflows to custom dashboards that bring data together in one place. He also dives into AI governance, describing the AI council he leads and the data policies that allow innovation while protecting sensitive information, ultimately positioning the CFO as a hands-on AI leader focused on both value creation and risk management. Key topics covered: Companies misunderstand what AI can do, deploy it inappropriately (e.g., trying to “fully automate everything”), and often lack in-house application developers who can tailor solutions to their actual workflows. Sinohe breaks AI use into augmentation, task automation, and application development, arguing that most ROI today comes from targeted task automation and small, purpose-built tools, not sweeping end-to-end automation projects. Envoy's finance team used AI to automate health insurance and other reconciliations, identifying about $40,000 in recoveries and turning tedious, quarterly work into a largely automated process. Sinohe personally builds AI-powered applications to reconcile accounts, summarize emails and Slack, prep and debrief interviews, and create a “morning coffee” dashboard that consolidates operational and financial insights into a single pane of glass. As head of Envoy's AI council, Sinohe has helped design a data governance matrix that clarifies what data can be used in which tools, allowing experimentation and creativity while strictly protecting company and customer data. Sinohe is bullish on increased data accessibility (e.g., via banks and platforms like Salesforce) and predicts a shift toward custom, CFO‑designed dashboards and tools, with legacy point solutions being displaced by in‑house applications that do exactly what the business needs. Links Sinohe Terrero on LinkedIn Kevin Appleby on LinkedIn GrowCFO Mentoring Timestamps:  0:01:36 – Sinohe explains Envoy as a workplace technology platform focused on managing physical spaces (visitor check-in, security, emergency notifications, desk allocation) with 6,000+ global customers and around 250 employees. 0:03:35 – He shares how timing, a tight investor story, and demonstrating strong cash flow and operational discipline were critical to a successful Series C raise during a turbulent market. 0:04:47 – Sinohe lays out the core reasons AI fails in many organizations and introduces his three-part framework: augmentation, task automation, and application development. 0:07:11 – He describes teaching himself to build AI-powered applications, including an asset-account reconciliation tool that cut a two-hour monthly process down to about two minutes. 0:12:21 – Using tools like Scribe to document workflows, Envoy's finance team identifies automation candidates; a payroll-led AI skill for health insurance reconciliations surfaced roughly $40,000 owed to the company. 0:17:53 – Sinohe explains Envoy's AI council, clear AI policies, and a data governance matrix that defines what data can be used where, enabling safe experimentation at scale. 0:21:17 – He details his personal AI setup: automated interview briefing/debriefing via Granola + Claude, daily digests of emails/Slack/meetings, and automated summaries of operational metrics and customer activity. 0:24:58 – Sinohe predicts job disruption in large teams (e.g., 100 accountants potentially shrinking to 60) but sees smaller teams using AI to focus on higher-value, advisory work rather than basic reconciliations. 0:26:30 – He describes replacing tools like Flowcast, Asana/Monday, and other SaaS products with custom AI-enabled applications that do 75% of what generic tools do—but 100% of what Envoy actually needs. 0:33:36 – Sinohe forecasts greater bank and platform data accessibility, more automated reconciliations, and a shift that frees CFOs from operational drudgery so they can focus on higher‑value strategic work. Find out more about GrowCFO If you enjoyed this podcast, you can subscribe to the GrowCFO Show with your favorite podcast app. The GrowCFO show is listed in the Apple podcast directory, Spotify and many others. Why not subscribe there today? That way, you never miss an episode. GrowCFO is a great place to extend your professional network. Join GrowCFO as a free member today and participate in our regular networking events and webinars. Premium members can also access our extensive training center and CFO Digital Toolkit. You can enroll in our flagship Future CFO or Finance Leader programs here. You can find out more and join today at growcfo.net

Good morning from Pharma Daily: the podcast that brings you the most important developments in the pharmaceutical and biotech world. Today, we're diving into a wave of exciting advancements and strategic movements shaping the landscape of this ever-evolving industry. Starting with a significant breakthrough in cell therapy, the FDA has granted approval to Orca Bio's Tregzi, a novel treatment aimed at preventing graft-versus-host disease (GVHD) in blood cancer patients undergoing allogeneic transplants. This approval emphasizes the potential of regulatory T cells in mitigating the severe complications that often accompany such transplants. The success of Tregzi not only marks a pivotal moment for cell therapy but also heralds a new era in transplant medicine where cellular therapies can significantly improve patient outcomes by reducing GVHD incidence. In other strategic moves, Ipsen's acquisition of Memo Therapeutics for up to $796 million highlights the industry's focus on addressing unmet needs in transplant medicine. Memo's work on monoclonal antibodies targeting the BK virus—a significant threat to kidney transplant patients—reflects a growing trend towards personalized medicine and targeted biologics. This acquisition is poised to bring much-needed therapeutic interventions to market, underscoring the sector's commitment to innovative solutions. Meanwhile, Anthropic is making waves with the introduction of Claude Science AI Workbench, accompanied by an internal drug discovery program aimed at neglected diseases. This initiative underscores the transformative role of artificial intelligence in drug discovery, particularly in areas previously overlooked due to limited commercial incentives. By accelerating the identification and development of novel therapeutics, AI-driven platforms promise breakthroughs in treating rare and neglected diseases. In clinical trial news, Abivax has reported promising Phase 3 safety data for obefazimod, a miR-124 enhancer targeting ulcerative colitis. By alleviating previous cancer-related concerns, obefazimod stands out as a potential small molecule oral therapy for autoimmune conditions. This advancement demonstrates ongoing innovation in small molecule therapeutics designed to modulate immune responses, expanding treatment options for ulcerative colitis patients. On the financial front, several companies are bolstering their pipelines through significant funding efforts. Alvotech has secured a $75 million term loan to enhance its biosimilar pipeline, reflecting the increasing importance of cost-effective biologic alternatives. AB Science has raised approximately $2.62 million focusing on oncology and rare diseases like acute myeloid leukemia (AML) and amyotrophic lateral sclerosis (ALS). Additionally, Flare Therapeutics' $85 million Series C round emphasizes precision medicine approaches targeting transcription factors, illustrating the industry's dedication to precision medicine. Regulatory landscapes are also shifting with Sarepta Therapeutics' FDA acceptance for full approval bids on Amondys 45 and Vyondys 53 despite confirmatory trial challenges. These therapies leverage exon skipping technology for Duchenne muscular dystrophy treatment, highlighting the intricate balance between accelerated approvals and robust clinical evidence requirements. In mergers and acquisitions news, Kimball Electronics' acquisition of Helvoet Polymer Technologies expands their drug delivery systems capabilities across Europe and India. This strategic alignment reflects broader industry trends towards integrated service offerings and global expansion efforts. Regulatory challenges continue with leadership changes at the FDA as Vijay Kumar steps down amid ongoing turbulence within gene and cell therapy sectors. Additionally, Angitia Biopharmaceuticals' termination of its Phase 3 trial for its BMP-6 candidate illustrates inherent risks and strategic pivots often necessary in drug development. Turning our focus globally, China's approval of its first CAR-T therapy for solid tumors signals a pivotal moment in oncology. This milestone could spur similar advancements worldwide, particularly drawing interest on when such treatments might receive approval in other regions like the United States. Moderna's expansion into in vivo CAR-T therapies for autoimmune diseases further marks a strategic divergence from traditional approaches, aiming to create off-the-shelf solutions that redefine treatment paradigms. Gene therapy continues to gain traction as Uniqure navigates regulatory reversals from the FDA regarding its Huntington's disease gene therapy filing for accelerated approval. This development hints at a broader shift within regulatory bodies towards fostering innovation in rare disease treatments under new leadership directives. Industry-wide trends also reveal a notable shift towards automation in cell therapy production led by companies like Cellares and Ori. This reflects growing demands for scalable manufacturing processes that enhance production capabilities while reducing costs. Overall, these developments illustrate a dynamic landscape where scientific innovations, strategic partnerships, regulatory adjustments, and technological integrations drive progress across various medical fields. As these sectors evolve, they hold vast potential for transforming patient care through groundbreaking therapeutic options and improved healthcare delivery systems worldwide.Support the show

We talk the foldable iPhone, Michael Jackson's death, and Trump passports. Some other notable news:Micron delivered one of the most stunning quarters in semiconductor history, reporting roughly $41.5 billion in fiscal-Q3 revenue, up about 346% year over year, and guiding next quarter to about $50 billion with an 81% gross margin. The read-through is that AI has turned high-bandwidth memory from a boom-bust commodity into a scarce, contracted input for the next generation of compute.SpaceX's record IPO unwound almost as violently as it launched, with the stock falling 31% in four sessions from its June 16 peak and erasing more than $600 billion of market value. A 4.2% public float, a $20 billion bond sale, newly listed options, August lockup risk, and a $4.9 billion 2025 net loss all collided in one of the clearest market-structure lessons of the AI trade.Oracle disclosed about 21,000 job cuts and directly tied the reductions to AI adoption in its own annual filing, even as it expands AI cloud capacity through major data-center deals linked to OpenAI and Meta. China also reclaimed the world's-fastest-supercomputer crown with LineShine, an all-domestic CPU-only system that hit 2.198 exaflops under U.S. export controls.A fatal Tesla crash in Katy, Texas, reopened the self-driving accountability fight after the driver said he had been using Tesla's partially automated driving system and both NHTSA and NTSB opened investigations. Robotics funding added the other side of the physical-AI story, with about $55.8 billion raised so far in 2026 and Figure banking a $1 billion Series C at a $39 billion valuation.The runner-ups: FedEx beat expectations and completed the FedEx Freight spin-off, onsemi agreed to buy Synaptics for about $7 billion to push deeper into edge and physical AI, and UN Secretary-General Antonio Guterres pressed AI companies to disclose data-center emissions, water use, land use, and energy sources. The 30,000-ft view: Q1 GDP was revised up to 2.1%, PCE inflation ran at 4.6%, markets repriced toward possible Fed hikes, Nvidia's Vera CPUs entered full production, and the mega-IPO pipeline still has Anthropic and OpenAI queued. If you want a prize, send us a DM: instagram.com/rickerandbon tiktok.com/@rickerandbon youtube.com/@rickerandbon

In this episode of The Product Podcast by Product School, Carlos González de Villaumbrosia sits down with Cristina Cordova, Chief Operating Officer at Linear, the product development system built for teams and agents. Linear raised $82 million in a Series C round in June 2025 at a $1.25 billion valuation. The company has been profitable since 2021, and serves over 20,000 paid business customers, from seed-stage startups to Fortune 100 enterprises, with a team of just 140 people. Before Linear, Cristina joined Stripe as one of its first employees, and led Platform and Partnerships at Notion.What you'll learn:Why keeping headcount intentionally lean is a strategic advantageReplacing traditional interviews with paid two to five-day projectsWhy PMs are the fastest-growing power users of agentic toolsKey takeaways:A small team is not a small business. Revenue, customers, and growth rate matter more than headcount.If you fully delegate your AI thinking, you lose your native understanding of how these products actually workAgentic workflows are now the default, not a feature. The companies that treat them that way will pull ahead.Credits:Host: Carlos Gonzalez de VillaumbrosiaGuest: Cristina CordovaSocial Links:Find out more about Product School hereFollow our Podcast on TikTok hereFollow Product School on LinkedIn here

Every founder gets a version of the same advice: don't pick a fight with an entrenched industry. The incumbents have the relationships, the regulatory cover, the deep pockets - you'll bleed out trying.But some of the most interesting companies of the last decade were built ignoring that advice, winning over markets that were nearly impenetrable.In this episode, Yaniv Bernstein is joined by Brandon Weber - co-founder and CEO of Nava Benefits, a Series C-funded AI-powered health benefits brokerage. Before Nava, Brandon co-founded Hightower, a commercial real estate startup that merged with VTS and went on to run over half of all office buildings in the United States. Brandon has now done this twice in two completely different industries, and has developed a repeatable playbook for breaking into entrenched markets and using AI as a structural advantage.Yaniv and Brandon dig into what actually makes a market 'broken', why the entry point needs to be far narrower than most founders think, and how to build the conviction to keep going when a thousand people tell you it won't work. In this episode, you will:Understand the 'burning platform' signal - what makes a market 'broken', but worth spending a years breaking intoLearn why your entry point needs to be far narrower than feels comfortable, and how Brandon went from targeting 'the health insurance market' to 'employers with 50-500 employees who can't afford a dedicated benefits team'Hear why 'disrupting from within' is often smarter than disrupting head-on - and how Nava built a broker-shaped entity that the industry's immune system couldn't rejectDiscover how to design a human-AI system (what Brandon calls a 'cybernetic' service model) where agents handle 80-85% of the work and licensed professionals operate at the top of their licenseTimestamps00:00 Coming Up…00:45 On Today's Show: Brandon Weber on Fixing Broken Industries01:43 How To Spot Broken Markets03:59 Why Most Healthcare Startups Fail (Distribution)05:35 Lessons From Building Hightower and VTS08:41 How Do We Think Smaller? Finding the 'Narrow Wedge'10:57 What It Means To 'Disrupt From Within'16:53 Choosing the ICP18:35 The Innovator's Dilemma and Moving Upmarket22:57 Scaling with AI: A Business in Two Phases26:29 Service as a Software34:02 Attract and Hire Industry Insiders36:44 When to Acquire39:06 Closing AdviceResources mentioned in this episodeNava Benefits (Brandon's company): https://www.navabenefits.comGary Lo's previous TSP episode: https://youtu.be/jtMgd7Nv_HYThe Innovator's Dilemma by Clayton Christensen (framework discussed at length): https://www.amazon.com/Innovators-Dilemma-Revolutionary-Change-Business/dp/0062060244The PactHonor the Startup Podcast Pact! If you have listened to TSP and gotten value from it, please:Follow, rate, and review us in your listening appSecure your official TSP merchandise at https://shop.tsp.show/Follow us on YouTube for full-video episodes: https://www.youtube.com/@startup-podcastGive us a public shout-out on LinkedIn or anywhere you have a social media followingKey linksThis episode of the Startup Podcast is sponsored by .tech domains. Forget weird prefixes and creative misspellings; the availability for .tech domains is simply way better than .com. For a clean name that highlights your tech credentials, get a .tech domain at your favorite registrar.The Startup Podcast website: https://www.tsp.show/episodes/Learn more about Chris and YanivWork 1:1 with Chris: http://chrissaad.com/advisory/Follow Chris on Linkedin: https://www.linkedin.com/in/chrissaad/Follow Yaniv on Linkedin: https://www.linkedin.com/in/ybernstein/Producer: Justin McArthur https://www.linkedin.com/in/justin-mcarthurAssistant Producer: Steph Hefferan https://www.linkedin.com/in/steph-heff/Intro Voice: Jeremiah Owyang https://web-strategist.com/

https://novacut.ai/  https://genaimeetup.com/  Anthropic has officially closed a $65 billion Series H at a $965 billion valuation, nearly 2.5x its valuation from just 100 days ago. Meanwhile, funding is flowing across the ecosystem: Frameworks AI at $15B, Baseten at $11B, OpenRouter's $113M Series B, and Cognition AI's $1B Series D. NVIDIA went on an open-source super week with Nemotron 3 Ultra, Cosmos 3, and Nemotron 3.5 ASR. Microsoft dropped 5 new MAI models. Google released Gemma 4 12B, and Anthropic shipped Opus 4.8. On the benchmarks front, DeepSWE crowns GPT-5.5 as the leader in long-horizon coding tasks, while ITBench shows even frontier models struggle with real-world SRE incidents — Claude Opus 4.7 tops out at just 47%. Plus: Cloudflare acquires VoidZero to build the future of AI-native edge development, and Google is paying SpaceX $920M/month for compute. Topics covered: • Anthropic's $65B Series H and path to $1T • Fireworks AI, Baseten, OpenRouter & Cognition funding rounds • Microsoft's 5 new MAI models • NVIDIA's open-source super week (Nemotron, Cosmos 3) • MiniMax M3, Gemma 4 12B, JetBrains Mellum2, Opus 4.8 • DeepSWE benchmark: GPT-5.5 leads long-horizon coding • ITBench: Frontier models under 50% on real SRE tasks • Cloudflare + VoidZero for AI-native edge dev • Google's $920M/month SpaceX compute deal #AI #Anthropic #NVIDIA #OpenAI #AInews #TechNews #LLM     Funding rounds Anthropic formally confirmed the closure of its $65 billion Series H funding round at a post-money valuation of $965 billion. This represents a 2.5-fold increase over its $380 billion Series G valuation from February 2026, adding $585 billion in value in approximately 100 days https://www.anthropic.com/news/series-h  Frameworks AI raising at 15B valuation representing a near fourfold increase from its $4 billion Series C valuation recorded in October 2025 processing 15 trillion tokens daily for major production clients including Cursor, Notion, and Perplexity https://finance.yahoo.com/sectors/technology/articles/fireworks-ai-eyes-15-billion-174609357.html Baseten is raising 1B at 11B valuation annualized revenue, which skyrocketed from $200 million to $600 million over a single quarter https://techstartups.com/2026/05/26/ai-inference-startup-baseten-in-talks-to-raise-1-billion-at-11-billion-valuation/  OpenRouter has secured a $113 million Series B funding OpenRouter has experienced exponential traffic growth, with weekly production throughput expanding fivefold from 5 trillion to 25 trillion tokens over a six-month horizon https://www.businesswire.com/news/home/20260526953416/en/OpenRouter-Raises-%24113-Million-CapitalG-led-Series-B-as-Weekly-Volume-Explodes-to-25T-Tokens  Further up the stack: Cognition AI secured a $1 billion Series D round led by Lux Capital and 8VC https://cognition.ai/blog/series-d   Model Releases MAI models: MAI-Code-1-Flash: A 5-billion active parameter model optimized for ultra-low latency within GitHub Copilot and VS Code. MAI-Image-2.5: A high-fidelity image generation model ranking third on global image evaluation arenas, outperforming competing architectures like Nano Banana Pro. MAI-Transcribe-1.5: A multi-lingual speech processing engine offering fivefold speed improvements across 43 languages. MAI-Voice-2: Natural audio and voice generation across 15 languages, available at a highly competitive price point. Web IQ: A search-grounding API engineered to directly compete with Perplexity. https://microsoft.ai/models/    https://www.peoplematters.in/news/ai-and-emerging-tech/uber-imposes-dollar1500-monthly-ai-spending-limit-on-employees-amid-rising-costs-50073    Nvidia has executed an "Open-Source Super Week," positioning itself as a dominant software and model publisher: Nemotron 3 Ultra (best US open source open weights model but behind china): A massive 550-billion parameter MoE (55 billion active) designed with a 1-million token context window, optimized specifically for high-throughput, cyclical agent loops. It achieved peak throughput rates of 400 tokens per second on day-zero optimized clusters. Cosmos 3: A physical AI world-modeling framework comprising 16-billion Nano and 64-billion Super variants. Built on a Mixture-of-Transformers (MoT) architecture, Cosmos 3 natively binds textual, visual, auditory, and physical kinetic vectors. Nemotron 3.5 ASR: A highly compact 0.6-billion parameter streaming speech recognition model pushing sub-100 millisecond latencies across 40 language locales.   https://www.minimax.io/models/text/m3  MiniMax M3: A 1-million token context model hitting 59.0% on SWE-Bench Pro and 74.2% on MCP Atlas, though noted for high token consumption due to intensive internal self-validation loops.   https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemma-4-12b/  Gemma 4 12B: Google's Apache 2.0 on-device model, which utilizes an encoder-free architecture that projects vision and audio vectors directly into the text-token space, bypassing separate CLIP-style encoders to minimize local memory footprints. https://www.jetbrains.com/mellum/  JetBrains Mellum2: A compact 12-billion parameter MoE (2.5 billion active) engineered for ultra-low latency routing and retrieval-augmented generation (RAG) sub-agents within developer IDEs. Opus 4.8 https://www.anthropic.com/news/claude-opus-4-8    https://www.cnbc.com/2026/06/05/google-to-pay-spacex-920-million-a-month-for-xai-compute-capacity.html      Benchmarks: https://deepswe.d atacurve.ai/blog https://venturebeat.com/technology/deepswe-blows-up-the-ai-coding-leaderboard-crowns-gpt-5-5-and-finds-claude-opus-exploiting-a-benchmark-loophole (GPT 5.5 the winner in long horizon tasks) a highly complex software engineering benchmark focused on original, long-horizon tasks across five distinct programming languages. Comprising 113 chaotic tasks across 91 live, production-grade repositories, DeepSWE forces agents to generate 5.5 times more code and modify an average of 7 separate files per task compared to standard evaluations. On this challenging leaderboard, GPT-5.5 leads with a score of 70%, establishing a significant 16-percentage-point lead over contemporary alternatives I think older benchmarks where models reach ~90% accuracy can be considered saturated. Few percentage points don't give us any good signal.  https://research.ibm.com/publications/developing-ai-agents-for-it-automation-tasks-with-itbench  ITBench-AA, an evaluation framework focusing on live Kubernetes incident response and Site Reliability Engineering (SRE) operations. Comprising 59 live, containerized SRE incident snapshots, the results are remarkably sobering: every frontier model scored under 50% on successful incident resolution, with Claude Opus 4.7 leading at 47% and GPT-5.5 following closely at 46%.   Edge AI announcements: https://www.cloudflare.com/press/press-releases/2026/cloudflare-acquires-voidzero-to-build-the-future-of-the-ai-native-web/  The consolidation of the AI-native developer stack has reached the runtime virtualization layer. Cloudflare recently completed the acquisition of VoidZero, the development group responsible for Vite, Vitest, Rolldown, and Oxc, backing the transaction with a $1 million open-source ecosystem fund. This acquisition is highly strategic; as autonomous agents write an increasing proportion of production software, local development environments, compilation pipelines, and bundlers must be optimized for execution speeds that match agent speeds. Cloudflare's goal is to construct a localized, full-stack edge playground. In this sandbox, AI agents can generate, test, bundle (utilizing the highly parallelized, Rust-based Oxc and Rolldown engines), and deploy entire web applications end-to-end within milliseconds. This architecture completely bypasses traditional local machine container bottlenecks, enabling high-velocity agent loops to execute in a fully sandboxed, web-scale edge runtime.

Good morning from Pharma Daily: the podcast that brings you the most important developments in the pharmaceutical and biotech world. Today, we're diving into some of the most significant advancements in scientific research, clinical trials, and regulatory landscapes within the industry. These developments are shaping the future of patient care and drug development significantly. Starting with Legend Biotech's LB2501, which achieved an impressive 100% response rate in a Phase 1 study for non-Hodgkin lymphoma using in vivo CAR T-cell therapy. This breakthrough highlights the transformative potential of CAR T-cell therapies in oncology, especially for B-cell lymphomas. Such success opens the door for accelerated regulatory pathways, offering hope to patients with limited treatment options. In another key development, JJP Biologics shared positive interim data from its Phase 1b trial of nebaprubart targeting CD89 in linear IgA disease. This monoclonal antibody is promising in treating autoimmune conditions by targeting specific disease mechanisms. Meanwhile, GSK's Velzatinib (IDRX-42) achieved a 61% response rate in Phase 1/1b trials for gastrointestinal stromal tumors, showing efficacy against cases resistant to treatments like imatinib. Johnson & Johnson's Nipocalimab met its primary endpoint in a Phase 2 study for systemic lupus erythematosus, underscoring the potential of FcRn blockade in managing autoimmune diseases. Zenas Biopharma's Phase 3 data for Obexelimab targeting CD19/FcγRIIB in IgG4-related disease further emphasizes the role of targeted therapies in managing complex autoimmune disorders. On the regulatory front, Foundation Medicine's FoundationOne Blood Test received FDA approval as a companion diagnostic for Pfizer's Talzenna (talazoparib) to treat prostate cancer with homologous recombination repair gene mutations. This approval underscores the growing importance of precision medicine and companion diagnostics in tailoring cancer treatments based on genetic profiles. Additionally, Lupin and Natco Pharma secured FDA approval for their generic version of Eribulin Mesylate Injection, essential for reducing healthcare costs and improving patient access to vital therapies. Eli Lilly's collaboration with Ascidian Therapeutics focuses on RNA exon editing for kidney diseases, potentially revolutionizing treatment approaches by correcting genetic errors at the RNA level. This partnership reflects a burgeoning interest in RNA-based therapies and their capacity to address unmet medical needs. Regeneron expanded its pact with CytomX Therapeutics to develop conditionally active bispecific antibodies, emphasizing innovation in oncology drug discovery. Such collaborations combine expertise across companies to expedite cutting-edge therapies' development. In terms of funding, NewLimit's successful $435 million Series C round aims to advance epigenetic reprogramming medicine towards human trials. This initiative highlights the burgeoning field of aging biology and its implications for extending healthy human lifespan through innovative therapeutic approaches. Similarly, Immu Biosciences raised $53 million to enhance its immunology platform using AI/ML technologies, underscoring AI and machine learning's critical role in accelerating drug development processes. Turning our gaze towards China's expanding influence on the global biotech stage, Akeso's presentation at ASCO 2026 marked a significant milestone as it became the first-ever Chinese dataset featured in a plenary session. This achievement underscores China's growing prominence in biotechnology and highlights its commitment to advancing innovative medical solutions globally. Simultaneously, Gilead's strategic partnership with Cencora aims to enhance access to CAR-T therapies like Yescarta and Tecartus by expanding their network of treatment centers. CAR-T therapies represent a paradigm shift in cancer treatment by offering personalized options for certain types of cancer. Despite challenges such as Roche's setbacks with its oral SERD drug giredestrant in breast cancer trials, innovation continues unabated. Zevra Therapeutics' launch of Miplyffa for Niemann-Pick disease type C exemplifies efforts to transform rare disease markets by improving patient outcomes through increased access and tailored treatment strategies. Finally, Eli Lilly's acquisition spree reflects broader trends where pharmaceutical companies increasingly integrate Chinese innovations into their development pipelines. This period marks a transformative phase characterized by collaboration between global pharma giants and Chinese biotechs, signaling an era where innovation is globalized and aimed at addressing critical healthcare challenges worldwide. These advancements reflect a dynamic period of innovation within the pharmaceutical and biotech industries. The focus on personalized medicine, targeted therapies, and groundbreaking technologies like RNA editing indicates a shift towards more precise treatment modalities. As these discoveries transition from research phases to clinical applications, they hold the potential to transform patient care significantly. Strategic partnerships and substantial funding initiatives illustrate a robust ecosystem supporting these innovations' rapid advancement. As regulatory bodies continue approving novel therapeutics and diagnostics, the emphasis on personalized healthcare will likely drive future developments, ultimately leading to improved patient outcomes worldwide. As we continue navigating these developments, it's clear that the pharmaceutical and biotech sectors are on the cusp of transformative breakthroughs that promise to redefine healthcare delivery across multiple domains. Thank you for tuning into Pharma Daily; stay informed and stay ahead.Support the show

2018 als kleines Start-up in Mannheim gestartet, gilt Osapiens heute als einer der spannendsten B2B-SaaS-Anbieter Europas. Das Unternehmen hilft Konzernen wie Metro, Otto, Edeka oder dm dabei, regulatorische Anforderungen – von der CSRD über das Lieferkettensorgfaltspflichtengesetz bis zur EU-Entwaldungsverordnung – rechtssicher zu erfüllen. Nach einer Series-C-Runde über 100 Millionen US-Dollar Anfang 2026 hat Osapiens den Unicorn-Status erreicht.Diesen Wachstumsprozess begleitet CFO Lucas Ziegler sowohl strategisch als auch operativ mit. Seit 2025 als CFO an Bord, hat er das Finance-Team von drei auf mehr als 20 Personen skaliert, die entscheidende Series C begleitet und gleichzeitig die internationale Expansion finanzierungsseitig abgesichert. In dieser Folge spricht er offen darüber, was das wirklich bedeutet – für Strukturen, Prozesse und den eigenen Kopf.Das erwartet Sie in der Podcast-Folge:• Warum der CFO heute Unternehmer und Go-to-Market-Copilot sein muss. • Wie man sich auf eine 100-Millionen-Dollar-Finanzierungsrunde vorbereitet. • Internationalisierung: Worauf es finance-seitig von Anfang an ankommt.• Teamaufbau: Wer zuerst kommt – und warum Konzernprofile oft scheitern.• Welche fünf KPIs im Scale-Up wirklich zählen.• Welche drei Fähigkeiten Finance-Talente entwickeln sollten, um später CFO in einem Scale-up zu werden.Die GesprächsteilnehmerHost: Esra LaubachGast: Lucas Ziegler (CFO Osapiens)Mit diesem Code können Sie sich oder Ihren Mitarbeitenden ein kostenloses Ticket für das Future FINANCE Festival 2026 sichern: FFF26PHier geht's zur Anmeldung: Teilnehmer | Anmeldeformular - Anmeldung zum Future FINANCE Festival am 11. Juni 2026 in Köln

Four stories today — starting with the most importantweek in NIO's 2026 delivery story so far.NIO delivered 37,705 vehicles in May 2026 — up 62.3%year over year and up 28.4% from April's 29,356.Three brands all firing: 20,013 NIO brand, 12,029 Onvo,5,663 Firefly. Year-to-date deliveries hit 150,526 —up 68.7% year over year. Cumulative deliveries reached1,148,118. The ES8 has been the number one sellingvehicle above 400,000 yuan for five consecutive months.The ES9 has over 50,000 pre-orders — with deliveriesonly starting May 28th. June is when the ES9 volumereally arrives.William Li is reportedly preparing to take Mirattery —NIO's battery asset operator — public as his eighth IPO.Mirattery owns the batteries in NIO's 3,846 swap stations.It recently closed a Series C total of nearly 2 billionyuan with state-backed institutional investors. A MiratteryIPO gives the battery swap infrastructure its own publicvaluation — separate from NIO's stock price. NIOshareholders already own a piece of that business.The market hasn't priced it in yet.The May Chinese EV delivery rankings reshuffled.Leapmotor delivered 81,569 units — record high for thesecond consecutive month. Zeekr delivered 34,377 unitsup 81.81% year over year. Huawei HIMA delivered 46,122.Xiaomi exceeded 30,000 for the month. BYD delivered376,990 — ending its 8-month year-over-year declinestreak. The market is forming a barbell — premium brandsand value brands winning, the middle getting squeezed.NIO is firmly on the premium side of that barbell.The Iran 60-day MOU is still waiting for Trump'ssignature. Hormuz remains largely closed. Oil at $92.56.June brings Gen 5 swap stations, Onvo L60 launch, andES9 volume ramping. The setup for Q3 is building now.

Founded in 2020, Axiado deploys hardware-anchored, AI-driven platform security by embedding silicon directly on the rack, protecting AI and cloud infrastructure against cyberattacks in real time. Latham represented Axiado in its oversubscribed US$100+ million Series C+ funding round. In this episode of Connected With Latham, Haim Zaltzman, Global Vice Chair of Latham's Emerging Companies & Growth Practice, sits down with Gopi Sirineni, Founder, President, and CEO of Axiado, to discuss the company's proximity-based security approach, the evolving cybersecurity landscape for AI infrastructure, and India's growing role in the global semiconductor ecosystem.   This podcast is provided as a service of Latham & Watkins LLP. Listening to this podcast does not create an attorney client relationship between you and Latham & Watkins LLP, and you should not send confidential information to Latham & Watkins LLP. While we make every effort to assure that the content of this podcast is accurate, comprehensive, and current, we do not warrant or guarantee any of those things and you may not rely on this podcast as a substitute for legal research and/or consulting a qualified attorney. Listening to this podcast is not a substitute for engaging a lawyer to advise on your individual needs. Should you require legal advice on the issues covered in this podcast, please consult a qualified attorney. Under New York's Code of Professional Responsibility, portions of this communication contain attorney advertising. Prior results do not guarantee a similar outcome. Results depend upon a variety of factors unique to each representation. Please direct all inquiries regarding the conduct of Latham and Watkins attorneys under New York's Disciplinary Rules to Latham & Watkins LLP, 1271 Avenue of the Americas, New York, NY 10020, Phone: 1.212.906.1200

SOND, a startup led by Bose's former head of sleep products, emerged from stealth with $7M in funding for its AI-powered sleep earbuds. Also, WeRoad, the Milan-based group travel startup, has raised a $58 million Series C round led by Airbnb as it prepares for its first major expansion outside Europe Learn more about your ad choices. Visit podcastchoices.com/adchoices

The FBI warns attackers are abusing Microsoft OAuth authentication. India pushes faster patching as AI speeds up cyberattacks. Iranian hackers blend phishing with SEO poisoning. Anthropic's AI finds thousands of open source flaws, while AI also reshapes bug bounties and fuels supply-chain attacks hitting thousands of GitHub repos. Plus, a new LMS zero-day, bulletproof hosting arrests in the Netherlands, FTC action over bogus “active listening” claims, and another busy week for cyber funding and M&A. Our guest is Kurtis Minder, author, joining us to discuss his book "Cyber Recon: My Life in Cyber Espionage and Ransomware Negotiation.” Please disregard all searches for disregard. Remember to leave us a 5-star rating and review in your favorite podcast app. Miss an episode? Sign-up for our daily intelligence roundup, Daily Briefing, and you'll never miss a beat. And be sure to follow CyberWire Daily on LinkedIn. CyberWire Guest Today we are joined by Kurtis Minder, author, joining us to discuss his book "Cyber Recon: My Life in Cyber Espionage and Ransomware Negotiation." Selected Reading FBI warns of Kali365 phishing service targeting Microsoft 365 accounts (Bleeping Computer) India's CERT-In Sets 12-Hour Patch Deadline for Exposed Flaws (Infosecurity Magazine) Iran-Linked Hackers Target US Aviation with Phishing and SEO Poisoning Campaign (Infosecurity Magazine) Anthropic: Mythos Detected 23,000 Potential Vulnerabilities Across 1,000 OSS Projects (SecurityWeek)  HackerOne takes an axe to its bug bounty rewards (The Register) Automated 'Megalodon' Campaign Spreads GitHub Repo Backdoors (GovInfo Security) Hackers Exploited KnowledgeDeliver Zero-Day for Web Shell Deployment (SecurityWeek) Admins of Bulletproof Hosting Service Used by Russian Hackers Arrested in Netherlands (SecurityWeek) FTC to Require Cox Media Group, Two Other Firms to Pay Nearly $1 Million to Settle Charges They Deceived Customers About “Active Listening” AI-Powered Marketing Service (Federal Trade Commission) Socket raises $60 million in Series C funding. (N2K Pro Business Briefing) You can no longer Google the word 'disregard' (TechCrunch) Share your feedback. What do you think about CyberWire Daily? Please take a few minutes to share your thoughts with us by completing our brief listener survey. Thank you for helping us continue to improve our show. Want to hear your company in the show? N2K CyberWire helps you reach the industry's most influential leaders and operators, while building visibility, authority, and connectivity across the cybersecurity community. Learn more at sponsor.thecyberwire.com. The CyberWire is a production of N2K Networks, your source for strategic workforce intelligence. © N2K Networks, Inc. Learn more about your ad choices. Visit megaphone.fm/adchoices

The man who helps finance Europe's defence: Robert de Groot, vice president of the European Investment Bank There is a particular kind of power that comes with someone who decides, quietly, which ideas get funded and which don't. Robert de Groot, and his team, holds that power over an extraordinary range of things: military bridges in Poland, rocket launchers in Spain, satellite-to-smartphone startups in Luxembourg, drone intelligence software in Estonia. As Vice President of the world's largest multilateral lender, the EIB sitting on the Kirchberg plateau, his brief covers security, defence, space, and innovation. It is, as he puts it with characteristic understatement, "quite a new direction" for a bank that, not long ago, wouldn't touch defence at all. That has changed. Dramatically. Since Russia's invasion of Ukraine, the EIB has rewritten its mandate, opening five distinct financing pillars across the defence and security ecosystem, from large-scale infrastructure to venture equity for startups building things that didn't exist five years ago. De Groot has spent the last two years touring every European capital, sitting down with defence, finance, and interior ministers, and asking “What does Europe actually need, and can we finance it?” "The urgency I hear in private is far greater than what you see in public." What he found on the road was a continent with a perception gap. The Baltic states are operating in a different psychological reality from much of western Europe. For Estonia, Latvia and Lithuania, the threat from the east is not geopolitics but geography. However, de Groot is cautiously optimistic. Germany has made a near-complete reversal on defence spending in three years. The Nordics have joined NATO. Ministers of Interior are now showing up to defence finance meetings, because the boundary between military security and civil security has dissolved. Cyber attacks, compromised energy grids, sabotaged undersea cables are happening now. The physical problems, meanwhile, are startlingly concrete. Bridges that cannot carry battle tanks. Ports unable to defend against unmanned underwater vehicles. Roads along NATO transit routes from Antwerp through Germany deep into Poland that haven't been maintained to handle today's military hardware. "It sounds absurd," de Groot says, "until you realise it's a multi-billion euro problem." The financing exists. The fixes are underway. But getting three countries to agree on a shared corridor before one of them goes its own way remains the harder challenge. For innovators and entrepreneurs building the dual-use technologies that now sit at the heart of European defence strategy, de Groot offers a map through the financing ecosystem. Early stage? Venture capital funds backed by the European Investment Fund. Series A and B? Venture debt, a product barely known in Europe five years ago, now scaling fast, with Luxembourg companies OQ Technology and Artec 3D among its beneficiaries. Series C and beyond? The European Tech Champions Initiative, designed explicitly to stop European unicorns from decamping to California. And for defence tech specifically, a new Defence Equity Facility of up to one billion euros: real, patient, European capital, with no American relocation clause attached. "The companies I meet across Europe mostly want to stay. We need to make sure the financing is there when they do." On the day of interview, a loan was signed for the Luxembourg Fire Brigade's logistics infrastructure. Security exists at multiple scales simultaneously, from orbital launch capability to the speed at which a fire engine reaches a crisis. Both matter and both require investment. Both represent the same underlying bet: that Europe, if it chooses to move with enough conviction, is more than capable of defending and financing its own future. De Groot, for his part, seems to believe it. The question, as ever, is whether the institutions can move as fast as the moment requires. Robert de Groot is Vice President of the European Investment Bank, responsible for Security, Defence, Space and Innovation Finance.

The window for government contractors, especially those in defense and space technology, to go public is open again as several listings over the past 12 months show and SpaceX's own offering this year will illustrate. Dave Khalsa, head of mid-cap defense and government technology investment banking at J.P. Morgan, works on transactions of many different types and observes all of them to help companies in the market figure it all out. In starting out this episode, Dave explains what all companies can take away from the handful of initial public offerings over the past 12 months and SpaceX's listing. This is true of whether they plan to go down the IPO path or not. The rest of the conversation between Dave and our Ross Wilkers focuses on how government priorities shape merger-and-acquisition activities by companies under different ownership models, including private equity and venture capital. Public offerings put GovCon in a new spotlight as SpaceX's listing looms HawkEye 360's public offering hauls in $416M AEVEX fetches $320M in IPO proceeds Firefly captures $868M in IPO proceeds York Space Systems raises $629M in public offering Merlin Labs' public offering collects $200M to build an AI autopilot for any aircraft L3Harris to spin off its rocket motor business with the Pentagon as an anchor investor AeroVironment's tech and business blueprints with BlueHalo now in the fold Veritas Capital's ninth fund grows to $15.3B OceanSound Partners hauls in $3.4B for third fund Arlington Capital fetches $6B for its seventh fund Government equity investments open a new frontier for industry Venture investing is part of the M&A conversation too Anduril hauls in $5B for Series H round Shield AI closes $1.5B Series G round and moves on acquisition Saronic wraps up $600M Series C round Sierra Space and Vast detail their Series C investment rounds

The FCC eases restrictions on foreign-made routers. Shiny Hunters hit Canvas and Zara. SailPoint discloses unauthorized access to its GitHub repositories. TrickMo Android banking malware has more tricks up its sleeve. Polish officials warn of increased targeting of ICS and public infrastructure. A federal judge orders $10 million in restitution for stolen zero days. German authorities takedown the Crimenetwork marketplace, again. Monday business breakdown. Dan Lorenc, Chainguard CEO and co-founder, is talking about a recent wave of supply chain attacks. Malware gets signed, sealed and delivered.  Remember to leave us a 5-star rating and review in your favorite podcast app. Miss an episode? Sign-up for our daily intelligence roundup, Daily Briefing, and you'll never miss a beat. And be sure to follow CyberWire Daily on LinkedIn. CyberWire Guest Dan Lorenc, Chainguard CEO and co-founder, is talking about how the recent wave of supply chain attacks is fundamentally different – and more dangerous –than previous incidents, as well as immediate steps organizations should take as this continues to unfold. Selected Reading US: FCC Relaxes Foreign-Made Router Ban to Allow for Security Updates (Infosecurity Magazine) ShinyHunters Escalates Canvas Extortion (Infosecurity Magazine) Zara Data Breach Impacts Nearly 200,000 Customers (Infosecurity Magazine) SailPoint Discloses GitHub Repository Hack (SecurityWeek) TrickMo Android banker adopts TON blockchain for covert comms (Bleeping Computer) Polish ABW warns cyberattacks shifting from espionage and data theft toward physical disruption of critical infrastructure (Industrial Cyber) Trenchant Exec Who Sold Zero Days to Russian Buyer Ordered to Pay $10 Million in Restitution to Former Employers (Zero Day) Resurrected 'Crimenetwork' Marketplace Taken Down, Administrator Arrested (SecurityWeek) XBOW secures an additional $35 million in Series C funding. (N2K Pro Business Briefing) Hackers Trick DigiCert Into Issuing Certificates Used to Sign Malware (Hackread) Share your feedback. What do you think about CyberWire Daily? Please take a few minutes to share your thoughts with us by completing our brief listener survey. Thank you for helping us continue to improve our show. Want to hear your company in the show? N2K CyberWire helps you reach the industry's most influential leaders and operators, while building visibility, authority, and connectivity across the cybersecurity community. Learn more at sponsor.thecyberwire.com. The CyberWire is a production of N2K Networks, your source for strategic workforce intelligence. © N2K Networks, Inc. Learn more about your ad choices. Visit megaphone.fm/adchoices

In 2026, startups age like milk. Josh Foreman's solution is a radical one - step down as CEO, go back to basics, and refound the whole company. Yaniv Bernstein discusses this decision with Josh, founder and (for now) CEO of InDebted - the AI-native debt resolution business he scaled to an $80M revenue run rate, a Series C raise, and operations across 8 markets. Just days before recording, Josh publicly announced he's hiring a new CEO so he can step back into the business as a hands-on operator and refound the company for the agentic AI era.In this conversation, Josh and Yaniv discuss 'refounding' in practice, what it takes to rebuild the company's processes from the ground up, and why technical founders who don't go back on the tools right now are setting themselves up to be outbuilt by a smaller, faster, leaner version of themselves.In this episode, you will:Learn why Josh believes the highest-leverage role for a technical founder in 2026 is no longer CEO, and how to structure a founder-CEO partnership that actually worksUnderstand why 'feature patching' an established business is a losing strategy, and what it really means to rebuild your company function-by-function from a clean slateDiscover how revenue-per-employee has become the metric that matters most when raising capital and competing with AI-native upstartsHear why services-as-software and performance-fee models are suddenly the bull case for investors who hated them 12 months ago - and why the SaaS seat fee is on the way outFind out what it looks like to unbundle your product into agent-ready primitives, and why owning the eval for a narrow domain may be a bigger moat than your full-stack UITimestamps00:00 Coming Up: Refounding00:41 Josh Foreman, CEO (for now)01:41 What Refounding Means04:53 Rebuilding the Factory07:38 Bringing the Team Along10:51 No Choice but Change14:56 Aligning the Board and Investors16:52 Putting Founders Back on the Tools26:24 'Corporate Ozempic' Shrinking Teams32:57 Unbundling and Products for Agents38:39 Hiring a CEO When Refounding44:11 Closing ThoughtsMentioned in this episodeJosh Foreman on LinkedIn: https://www.linkedin.com/in/joshforeman/InDebted: https://www.indebted.co/Scott Galloway on 'Corporate Ozempic': https://www.profgalloway.com/corporate-ozempic/Surviving the AI SaaSpocalypse with Scotty Allen: https://youtu.be/j84LF4aru8I 'Paranoid Optimism' with Yaniv: https://youtu.be/FGqbdzr0-PM The PactHonor the Startup Podcast Pact! If you have listened to TSP and gotten value from it, please:Follow, rate, and review us in your listening appSecure your official TSP merchandise at https://shop.tsp.show/Follow us here on YouTube for full-video episodes: https://www.youtube.com/channel/UCNjm1MTdjysRRV07fSf0yGgGive us a public shout-out on LinkedIn or anywhere you have a social media followingKey linksThis episode of the Startup Podcast is sponsored by .tech domains. Forget weird prefixes and creative misspellings; the availability for .tech domains is simply way better than .com. For a clean name that highlights your tech credentials, get a .tech domain at your favorite registrar.The Startup Podcast website: https://www.tsp.show/episodes/Learn more about Chris and YanivWork 1:1 with Chris: http://chrissaad.com/advisory/Follow Chris on Linkedin: https://www.linkedin.com/in/chrissaad/Follow Yaniv on Linkedin: https://www.linkedin.com/in/ybernstein/Producer: Justin McArthur https://www.linkedin.com/in/justin-mcarthurAssistant Producer: Steph Hefferan https://www.linkedin.com/in/steph-heff/Intro Voice: Jeremiah Owyang https://web-strategist.com/

Progress Software urges customers to patch a critical MOVEit authentication bypass. Washington worries about limited access to advanced AI tools. Paid influencers promote pro-American AI. CISA warns Copy Fail is under active exploitation. The Canvas educational platform suffers a data breach. The Lazarus Group uses ClickFix to target high-value enterprise users. U.S. and Chinese authorities raid scam centers in Dubai. Monday Business Brief. On Afternoon Cyber Tea with Ann Johnson: Tony Sager, Senior VP & Chief Evangelist, Center for Internet Security, joins Ann to discuss the accelerating pace of technology, AI, and global software dependencies. May the Fourth be with your firewall.  Remember to leave us a 5-star rating and review in your favorite podcast app. Miss an episode? Sign-up for our daily intelligence roundup, Daily Briefing, and you'll never miss a beat. And be sure to follow CyberWire Daily on LinkedIn. Afternoon Cyber Tea On this segment of Afternoon Cyber Tea with Ann Johnson: Tony Sager, Senior VP & Chief Evangelist, Center for Internet Security, joins Ann to discuss how the accelerating pace of technology, AI, and global software dependencies are reshaping the cybersecurity landscape. To hear the full conversation, check out the episode and subscribe where you get your favorite podcasts to listen to past episodes. The show is going on hiatus. Stay tuned for the next chapter soon. Selected Reading ⁠Progress warns of critical MOVEit Automation auth bypass flaw⁠ (Bleeping Computer) ⁠What Was Discussed at Google's White House Meeting About A.I. ⁠(The New York Times) ⁠US Military Reaches Deals With 7 Tech Companies to Use Their AI on Classified Systems ⁠(SecurityWeek) ⁠A Dark-Money Campaign Is Paying Influencers to Frame Chinese AI as a Threat⁠ (WIRED) ⁠CISA says ‘Copy Fail' flaw now exploited to root Linux systems⁠ (Bleeping Computer) ⁠Edtech Firm Instructure Discloses Data Breach Amid Hacker Leak Threats⁠ (SecurityWeek) ⁠Lazarus Targets macOS Users With New “Mach-O Man” Malware Kit⁠ (GB Hackers) ⁠US, China partner on scam center takedown in Dubai⁠ (The Record) ⁠Cloudsmith raises $72 million in Series C funding.⁠ (N2K Pro Business Briefing) Microsoft for Startups (N2K Networks) Share your feedback. What do you think about CyberWire Daily? Please take a few minutes to share your thoughts with us by completing our brief listener survey. Thank you for helping us continue to improve our show. Want to hear your company in the show? N2K CyberWire helps you reach the industry's most influential leaders and operators, while building visibility, authority, and connectivity across the cybersecurity community. Learn more at sponsor.thecyberwire.com. The CyberWire is a production of N2K Networks, your source for strategic workforce intelligence. © N2K Networks, Inc. Learn more about your ad choices. Visit megaphone.fm/adchoices

The U.S. military doesn't have enough pilots—and automation may be the only way to scale airpower. At the same time, Skyryse is formally launching its new defense unit, bringing its software-defined flight system, SkyOS, into military applications. On this week's episode of Valley of Depth, we sit down with Mark Groden, CEO of Skyryse, to unpack how the company is building a universal operating system for aircraft that can dramatically simplify flight, reduce pilot burden, and enable fully autonomous operations when needed. The goal is ambitious: turn helicopters and airplanes into flexible, optionally piloted systems that can shift between crewed and uncrewed missions—unlocking a new model for force projection, logistics, and survivability. The conversation spans the tragic accident that inspired Mark to start Skyryse, why aviation's biggest safety problem is really a technology problem, how SkyOS works across platforms from Robinson helicopters to Black Hawks, and why defense demand for autonomy is accelerating faster than most people realize. We cover: How SkyOS transforms aircraft into software-defined systems Why helicopters are so difficult and dangerous to fly today What Skyryse Defense is building for crewed, uncrewed, and autonomous missions How optionally piloted aircraft could reshape military logistics and ISR How Skyryse's Series C positions the company for scale Why the future battlefield requires simpler, more adaptable systems …and much more.   • Chapters • 00:00 – Intro 01:34 – The accident that changed Mark's life and mission 04:10 – A PhD in sensor data fusion 06:54 – The evolution of Skyryse 10:09 – Product stack 15:30 – New business unit 17:12 – Skyryse's partnership with the Army 19:39 – Why even build for humans? 21:35 – The software distribution of SkyOS 26:40 – Guinness World Record for autorotation 30:58 – Training commercial helicopter pilots with Skyryse 33:52 – Commercial picture for Skyryse 37:43 – Addressing the pilot shortage in the military 42:22 – Commercial regulations 45:39 – What certification unlocks for Skyryse 47:19 – Military regulatory process 48:53 – What Skyryse plans to do with their Series C funding 51:27 – How people's lives change if Skyryse is everywhere in 20 years 53:30 – Can you buy the Skyryse helicopter? 54:05 – What Mark does for fun when he's not building helicopters   • Show notes • Skyryse's website — https://skyryse.com/ Skyryse's' socials — https://x.com/skyryse Mo's socials — https://x.com/itsmoislam Payload's socials — https://twitter.com/payloadspace / https://www.linkedin.com/company/payloadspace Ignition's socials — https://twitter.com/ignitionnuclear /   https://www.linkedin.com/company/ignition-nuclear/ Tectonic's socials  — https://twitter.com/tectonicdefense / https://www.linkedin.com/company/tectonicdefense/ Valley of Depth archive — Listen: https://pod.payloadspace.com/   • About us • Valley of Depth is a podcast about the technologies that matter — and the people building them. Brought to you by Arkaea Media, the team behind Payload (space), Ignition (nuclear energy), and Tectonic (defense tech), this show goes beyond headlines and hype. We talk to founders, investors, government officials, and military leaders shaping the future of national security and deep tech. From breakthrough science to strategic policy, we dive into the high-stakes decisions behind the world's hardest technologies. Payload: www.payloadspace.com Tectonic: www.tectonicdefense.com Ignition: www.ignition-news.com

Latin America is the third-largest fintech and payments market in the world — bigger than India, behind only the US and China. Gastón Irigoyen is Co-Founder and CEO of Pomelo, the fintech infrastructure company powering card issuing and processing for banks, fintechs, and global enterprises across eight markets including Brazil, Mexico, Argentina, Colombia, Chile, Peru, Puerto Rico, and Panama. Pomelo is backed by Index Ventures, Insight Partners, Kaszek, Monashees, and most recently Adams Street Partners in their first-ever Latin American investment.In this episode of The J Curve, Gastón unpacks the contrarian playbook behind Pomelo: why the team went regional from day zero on a $10M seed round instead of nailing one market first, how they built a "plug and play" hiring engine that's stayed at 90%+ since founding, why they tripled revenue without adding headcount, and what it actually takes to win enterprise customers like BBVA, Santander, Bci, Bancolombia, Binance, and Bybit when nobody trusts an infrastructure startup. He also shares the Series B-to-Series C lessons most founders never document — including the end-of-year memo that turned rejections into investor trust — and his framework for the AI transformation a five-year-old company is now being forced to run.This is a masterclass on regional-by-design strategy, B2B fintech go-to-market, founder-led fundraising in down markets, and building world-class companies from Latin America for the world.Subscribe to The J Curve Insider newsletter for deeper insights and follow Olga on LinkedIn and Instagram.

The Great private Capital Reset is upon us. Markets are volatile and driving new economic imperatives. Are VC funds still VC funds, even if they raise billions per fund? What happened to the rest of the market? What is driving VC investments? What do Limited Partners think? What is on their minds? This and more, in episode 76 of Tech Deciphered. Navigation: Intro The State of the Reset: The Hangover from the Party? LP Fatigue and VC Differentiation What Really Matters: Performance.. Returns The Mega Fund Question The Case for Smaller… Rightsized Funds What Comes Next? Conclusion Our co-hosts: Bertrand Schmitt, Entrepreneur in Residence at Red River West, co-founder of App Annie / Data.ai, business angel, advisor to startups and VC funds, @bschmitt Nuno Goncalves Pedro, Investor, Managing Partner, Founder at Chamaeleon, @ngpedro Our show: Tech DECIPHERED brings you the Entrepreneur and Investor views on Big Tech, VC and Start-up news, opinion pieces and research. We decipher their meaning, and add inside knowledge and context. Being nerds, we also discuss the latest gadgets and pop culture news Subscribe To Our Podcast Bertrand Introduction Welcome to episode 76 of Tech Deciphered. This episode will be about the great private capital reset. As you know, or you have probably heard, there is significant structural transformation in the world of venture capital, and we are probably witnessing a fundamental reset of the private capital stack. We got a huge bubble in 2020, 2021. Fueled by near-zero interest rates. We got inflated fund size, compressed due diligence, and now a generation of zombie funds and zombie startups. Now that rates have normalized, exits have not been as much as expected. LP patience is a warning sign, and I guess the industry is being forced to confront an uncomfortable truth: most VC funds raised since 2017 might not return what their LPs expected. You know, how do we start?   Nuno This is going to be a relatively nuanced episode. Obviously, there is going to be a lot of haves and have-nots, both in terms of VC funds, also in terms of startups. And so I want to start with that. This is going to be more nuanced than all transformational and disruptive.   Bertrand It’s not the end. It’s not the end.   Nuno State of the Reset: The Hangover from the Party? It’s not the end. There’s still huge mega funds that are raising more and more. It’s clear that the music has stopped, right? So if we’re playing the game of chairs, the music has stopped. Around ’22, ’23, we started seeing the first signals that funds had raised way too much money. Firms collectively raised around $669 billion globally in 2021 alone. If we fast forward now to last year, 2025, depending on the sources, we did some internal analysis at Chameleon. We came up with $75.6 billion was raised last year by 493 funds, right? So That’s a significant drop, right, in terms of fundraising. Other sources would say a little bit more. There’s a little bit of a discussion around how much did the top 30 funds capture. If you believe some of the stats out there, they would say that actually top 30 funds captured 75% of all capital raised last year. We did again some internal analysis at Chameleon, and the conclusion we came to, it was closer to 50 to 55%. So not as dramatic as some of the sources out there, but still pretty dramatic. There’s a lot of capital concentration on the top funds. Again, the top 30 funds would’ve raised 50 to 55% of capital or up to 75% according to other sources. So definitely a tremendous amount of concentration. There was a lot more fragmentation in terms of capital raised if we’re looking at the years from 2010, 2011, all the way through 2021. So 2021 would’ve been sort of the peak of non-concentration if you look at that. And that again, now we are getting more and more concentration. There’s more and more of this arbitrage around, I’ll give money to the top funds, I will not give money to the smaller funds, or I’ll give less money to the smaller funds. There’s a little bit of a movement around concentration. We’ll talk about it later and what that means. Are mega funds really better? Are the small funds still the way to go? We’ll talk a lot about that later in today’s episode. There seems to be a little bit of a bifurcation. We could say it’s either bifurcation around top-tier VCs or larger VC funds versus smaller VC funds. My perspective is the bifurcation that we’re seeing right now is more of a bifurcation between funds that are no longer just stepped into the VC space, but they’re actually becoming more and more private equity firms with full asset management range from early stage all the way to late stage. Think of it almost like a private equity hedge fund, quasi, versus classic VC funds. And I think what we’re seeing is the Andreessen Horowitzes, the a16zs of the world, the NEAs, the Sequoia Capitals, just to name a few, becoming more and more broad asset class managers across private equity, whereas you have more classic VC happening in earlier stages. And so that’s the real bifurcation that I think is actually happening.   Bertrand And maybe not really hedge fund, because they are always still long-only funds. So there is no hedging happening, at least as far as I know.   Nuno Well, some of these guys have become RIAs, like A16z has become an RIA, so they can do secondaries.   Bertrand That’s true. Yeah.   Nuno And they can also sell stuff, etc. So I don’t know how aggressive they’re going to be in terms of secondaries and selling and actually doing other kinds of services you can do if you’re an RIA. But it’s not, I think, out of the realm of possibility that they would sort of acquire and sell stock more rapidly. In that way, to your point, Bertrand, maybe they actually become beyond just long guys, right?   Bertrand Yes. Another trend I have seen is some of the larger VC funds seems to have no problem investing in multiple competitors. This was not possible before. I mean, if you’re a VC fund, you had some sort of duty not to invest in the competitors, but now some invest OpenAI, Anthropic at the same time. Do you see that as part of this evolution?   Nuno For sure. And I think there’s a lot of people like the ostrich putting their heads below the ground and it’s like, “Eh, no, no, nothing to see here.” But that does constitute a conflict of interest. And if I’m a startup raising, this assumption that you will not invest in one of my competitors is no longer there, certainly for the mega funds, because of that notion of deployment of capital. Now, some funds will still hide under the notion, actually formally from a fund perspective, we’re not investing in competitors. It just happens that different types of our funds are investing in competitors. Like maybe my growth fund is investing in a competitor to my early stage fund, right? But our funds are relatively independent. So I think there’s a little bit of hide and seek that will go on if you talk to some of the fund managers. Well, they say, well, we’re not investing out of the same fund into these competitors. But between you and I, as we know, a lot of these partnerships actually do a lot of stuff together at the general partnership level. So are there really actual Chinese walls between the funds? Well, it really depends on the partnership. And to be honest, most of the partnerships don’t have very significant Chinese walls between the funds, right? The managing general partners sometimes actually occupy investment committee roles across different funds. So I think the conflict of interest is there. So that’s why I say there’s a little bit of ostrich behavior. Put your head behind the ground or below the ground and just pretend nothing is happening. Just sharing maybe a couple of interesting stats. Global fund closings for 2025, according to our numbers at Chameleon, 1,098 closed. In 2025. Closed is when you start deploying capital, right? Whereas— so it’s not closed down, it’s closed like we start deploying capital. And that number, 1,098, is dramatically down from 1,600 in 2024. And it’s actually the lowest number of closings that we saw since 2014. So again, this is bad, right? It means there’s less funds doing fund closings and deploying capital in the market than since 2014 and dramatically below the 2024 numbers, right? Where we already saw some market readjustments. The number of active VC firms in the US that did 2+ deals, which is not a huge bar, has dropped 38% back to numbers in 2023. So we don’t have numbers that are a little bit more up to date, but basically in 2023, those numbers are already dramatically dropped. So there’s less and less active funds. So there’s funds that might be in the market, but they’re not actually deploying that much capital, not doing that many investment. They’re sort of either zombie funds or relatively passive funds that have passed their investment period. For those listening to us, the investment period for a VC fund is normally between the first 3 to 5 years of the fund, which is when you build your portfolio, when you can invest in new companies. After that time period, everything that you do up to normally what would be year 10 is follow-ons. You put more money into the companies that you’re already invested in, that you already constructed portfolio with during those 3 to 5 years.   Bertrand Yeah, that’s a pretty scary change. And obviously, I guess we’ll come to it, but the time it takes to fully liquidate investments is getting longer and longer. In the old days, we used to talk about VC funds having a 10-year life, maybe a +1/+1 in terms of extension of the fund life. But it looks like it’s taking 16 to 18 years actually to get full liquidity from a fund investment.   Nuno LP Fatigue and VC Differentiation And I think that’s the scariest piece. I mean, just to share some numbers, we in venture capital talk about vintages, right? Which year did your fund start in? Normally when you did your first close onto the fund, as we were saying before, close is when you get all your investors at that moment in time to come in and you do your first close so the next fund starts running. 2018 vintage funds, right? This is now almost 7 years ago. So you should start having— actually 8 years ago almost at this point in time. You should start already getting distributions or you start getting cash back if you’re a limited partner and investor in those funds, you should start getting cash back. Half of all 2018 vintage funds have returned $0 to their LPs. So they’ve had no distributions to their LPs. 2020 vintage, which was a very hot vintage, only 42% have begun any distribution. So 58% have distributed $0, right? 2021, only 25% have done any distributions. Now, I happen to have a 2018 vintage fund and a 2021 fund. My 2018 fund has already distributed over 3x net of fees in distributions, and my 2021 fund’s already over 10% distributed back in distribution. So we’re very proud of that. But in general, the numbers are awful. There’s no liquidity back to LPs. And to your point, that’s kind of a big deal because some of these funds have been going on for 7, 8 years, and where’s the liquidity going to come from? On the other hand, if you look at TVPI, so DPI is distributions to paid-ins cash on cash. But if you look at TVPI, which is total value to paid-in, which also includes the book value or the value that you’re marking it on your books, basically the paper value as we call it for the company, even on that, the median 2017 fund, so 2017 vintage fund has a TVPI, total value to paid-in, of only around 1.76x, which is well below what should be, which is sort of the 2 to 3x benchmark of a really good performing fund. So the median funds are doing very, very poorly overall. So if you add that to the fact of what’s happening and distributions are taking a long time, back to your point, Bertrand, it’s taking like— this should be a 10-year asset class, maybe 11, 12 years, and now it’s looking a little bit like a 15, to 18-year asset class, which is not what most limited partners sign up for. Part of this dynamic, I think, is that we’ve had tremendously overvalued private companies over the last few years, right? Secondly, these companies have just stayed private longer. And I was having a discussion recently with a friend of mine, it’s like, hey, what’s this thing about companies are staying private much longer? Is there some dynamic around secondaries? And the reality is there is a dynamic around secondaries, right? Because if I’m a very large fund and I can get away with doing secondaries on my portfolio, I will get liquidity at some point, right? But someone else is stuck with private stock, which hopefully will IPO, but who knows, right? And so there’s this funny dynamic right now of because of secondaries, because of a couple of other things that are happening in the market, actually a lot of these startups are staying private for tremendous amounts of times, and some of them will IPO and they’ll be huge deals. Some of them might not and might not warrant the latest private valuations that they’ve exercised. And so there’s this tremendous noise that we’re seeing in the mid to late funnel of privately held companies where some are just waiting to be public. Some of them might not be able to go public at anything that is an up round versus private valuations that they’ve had in previous moments and in previous rounds.   Bertrand And obviously the 2 to 3x returns that funds are targeting, and obviously more 3x than 2x, I mean, that was good and nice if it’s a 10-year fund, but if it’s the same 3x for 15 to 18 years, it’s not at all the same rate of return annualized. So it’s a really, really, really big issue if you keep the return the same, but you extend the duration of the fund. Concerning going IPO, there is a lot of complexity going public, the IPO process itself, but also after that when you’re a public company. It changed how you can run the business. Some would argue that we have had an issue with more companies delisting than companies listing on the public market. So I think there might be also separate issues about the efficiency of the public market and maybe a need for change. We went very strongly in one direction for the public market, have post and run, but was it really ultimately the right thing to do? I’m actually not so sure.   Nuno Yeah, I mean, just to be clear, this is anecdotal, but when we tell prospective LPs at Chameleon about our returns, the last few funds, 2018, 2021, the first reaction is, “You must be lying, right? Surely you can’t have distributions already for 2021,” et cetera, et cetera. So clearly there’s almost a state of disbelief right now from limited partners. And liquidity does matter. So clearly you have to move forward. So how did we get to this point where we had this bubble 2021 all around that time space and now things don’t look so good. Well, the macro conditions have changed dramatically. I mean, rates when they were near zero, safer assets yield nothing or yield nothing. So basically you had to push capital into longer duration risk assets like venture capital. And so you had to push it. So the opportunity cost of capital also has fundamentally shifted. Obviously a 3x VC return in 15 years over 10 actually competes very poorly against 5% annual credit returns over several years. So there’s been a readjustment of stuff. And then the public equities in particular, the tech public equities have had a lot of volatility, but some of them have done extremely well, right? Chipsets, things like NVIDIA, the Amazons of the world, Alphabets, et cetera, et cetera. They’ve done very, very well. So why would I invest in a long-term illiquid asset that takes now longer to give me money back, and in some case doesn’t give me back, if I can invest just in public equities, and a variety of other things. The venture debt costs have increased dramatically. The burn rates that were sustainable back in the day with sort of the addition of venture debt, private credit, et cetera, now are overblown at this moment in time. At the end of the day, there’s been a lot of movements also overall in the pipeline in terms of valuations, et cetera, et cetera. Now, I would put a grain of salt into all the numbers I just told you. There still is a little bit of the haves and have-nots in startup land. Certainly in early stage where if you’re a hot AI company, you can get away with raising a Series C or $480 million. This is actually a true story. Series C, right? Not Series C, a $480 million at $4 billion pre-money valuation. Whereas if you are maybe in a space that’s less hot, you’ll have more difficulty in raising money at this point in time, might not be able to even raise a Series C, right? So there’s a little bit of the haves and have-nots happening on the VC side in early stage that has been really amplified by the macro regime and where we’re at, which is actively zero-rate era is done and now the new regime is quite different. And so I can get better returns by doing something else.   Bertrand Kind of makes sense. I mean, if you have some ways the SaaSpocalypse in the public market because there is that fear that AI is going to completely change the game for especially for the more typical software companies. Good luck raising private money to quote unquote just build traditional software companies. You cannot expect a warm embrace from the private market if the public markets are completely destroying that category. I’m not saying that this is there forever, uh, things might change over time, but for sure what’s happening on the public markets always have a very strong impact on the private market.   Nuno Indeed. So what’s happening in this relationship between limited partners and VCs, the general partners? Again, limited partners are the people that give venture capital firms and venture capital funds their capital to actually deploy. And they are a variety of different players, right? Could be endowments, like university endowments, pension funds, family offices, very high net worth individuals, fund of funds, et cetera, et cetera. I mean, in particular, if you look at the institutional investors, the endowments, the pension funds, the fund of funds, they have allocations that they do to different asset classes typically. And the feedback that we’ve received from the market is they are increasingly frustrated with what’s happening in terms of distributions. They’re not getting capital back. It’s like, I gave you capital 8 years ago, 9 years ago, 2017, 2018 vintages, and I’m not getting any capital back. So what the hell’s happening? On paper, it looks maybe the fund’s doing okay or it’s doing great in some cases, but where’s my money? And so that creates a little bit of wait-and-see kind of game on portfolio allocation. As we’re thinking through their re-ups, putting more capital into funds that they’re already actually put capital or putting in capital into new slots, into new fund managers that they want to put money into. They’re like, well, let’s wait and see. I want to get my money back or get some money back first before I redeploy it. Again, this is a little bit the haves and have-nots because we’ve seen, for example, a couple of top-end LPs in terms of returns that have a little bit the opposite problem, right? Because they are into funds that are performing extremely well. They actually are over that period and they want to actually redeploy. But to be honest, the average in the industry right now is a wait-and-see game. It’s like, I want to wait and see, which leads to what can only be characterized— I was hearing someone the other day, one of the top advisors in the LP community, saying this is the worst fundraising environment ever for venture capital. Not the last 20 years, 30 years, like ever, right? Since this became an asset class more institutionally in the late ’60s, early ’70s, Pulse Robo 2 as it was created, this is the worst fundraising environment ever. Oh, wow.   Bertrand And concerning TVPI, let’s not forget that typically it’s not mark-to-market. So the metrics in terms of TVPI, correct me if I’m wrong, you know, but the metrics in TVPI are based on typically the last fundraise. So if the valuation went down but there was no additional fundraise, we wouldn’t know by looking at the TVPI metrics. It will only be updated if there is a new Financing, equity financing, or an exit.   Nuno Yeah, normally most funds act like that. Some funds are a little bit more aggressive and do do mark-to-market, but normally funds would be conservative and say, hey, I’m being conservative, it’s whatever is the last known valuation of the company. And if there wasn’t a priced round, it’s a little bit more obscure than that, right, Bertrand? Because it might actually be the company has raised money on a note, or either convertible note or a SAFE note, and that wouldn’t count as a priced round. So I would say actually, even if it was a cap that’s below with a significant discount, I won’t recognize the assets as a down round. I won’t recognize the asset with a lower valuation because formally it wasn’t a price round. So it’s on the one hand conservative, on the other hand, it’s only relating to price rounds or exits to your point. So it’s sort of, you can be like, hmm, well, we opt to do that because we think it’s actually the most conservative route. Mark-to-market is extremely difficult to do. And who would do the mark-to-market for you, right? It’s like it’s some valuation firm, et cetera.   Bertrand I’m not saying a mark-to-market is easy, but I’m not sure I would call using the last valuation something conservative in the context that most startups will fail. So it’s not clear.   Nuno Well, in some cases it is, some cases it’s not, right? Depends on the startup situation, to be honest. Yeah, yeah.   Bertrand But yeah, at least that’s how it’s done. So for instance, to evaluate the impact of the SaaS apocalypse, it’s tough to know. We will have on the private market. I mean, we will see that in a few quarters. Because if companies still exist in that environment, if they still do additional truly price rounds after that, that’s when I will start to know.   Nuno I mean, just to share a little bit more data, like VC fund close time stretched to 15 months. Basically, it’s just taking a long time to raise money. It’s taking a long time to do your first close, get your fund running. When entrepreneurs complain to me that their fundraising is difficult, I always say, you have no clue how difficult it is compared to ours. First-time funds have collapsed. We had some numbers that only 77 first-time funds actually closed. I assume this is in 2025 versus 215 in 2023. So that’s a huge number. We did some internal analysis on our side and we did some analysis that emerging fund managers, emerging fund managers are normally people that are in their first one or two funds. Basically emerging fund managers gained some ground until 2017. Reaching by then a slice that was 63.7% of all capital raised in 2017. But since then, the capital deployed to emerging managers has been largely reduced to actually 24.2%, right? So it’s gone from 63.7% in 2017 to 24.2%. So this has been a culling of sorts on emerging managers and almost like a slaughterhouse of emerging managers. Compared to previous situations, which is obviously incredibly concerning if you’re an emerging manager starting your VC firm, et cetera, et cetera. So really tremendously problematic for those. We think capital’s not leaving VC. I think we see a lot of the institutionals saying— there’s some numbers as high as 33% of institutional investors plan to invest more in venture in the next 12 months. So I don’t think capital’s leaving VC. I think it’s really concentrating. We’ll come back to the concentration issue later in the episode. And part of that concentration comes from a topic that has been widely spoken in venture capital recently, which is differentiation. How do you differentiate in venture capital if you’re talking to a limited partner, right? How does my firm differentiate versus the firm next to mine? And that’s incredibly, incredibly challenging. Bertrand, what are your thoughts on that?   Bertrand Differentiation is always a question. I mean, if you’re an entrepreneur, Typically, you think fully about the best possible partner for your stage and for your type of business model. You want a VC who understands fully your business model, because if they don’t, then it’s going to be troubled down the line. But that’s true that another piece of the puzzle is that the best VCs help you get more visibility in terms of achieving potential customer deals, in terms of attracting the best talent. And that’s where VCs’ brand names can help. If you can say you have backing by some of the top, most visible names in the industry, and usually these are the mega funds because others have trouble to be as visible, then they have some sort of unfair advantage compared to others. So I can see that there is some level of concentration happening naturally, especially in the later stage from Series B onwards.   Nuno What Really Matters: Performance… Returns Yeah, I mean, we did some analysis internally about What are the top funds that invested in the top performing companies in early stage, Series C, Series A? And we looked at it by size of fund and the top performing normally are funds below $100 million, but in some cases very closely followed by funds between $100 and $500 million. And actually funds above $500 million, so $500 million to $1 billion and then $1 billion and above are actually tremendously underperforming. So this notion of the industry that says, well, the mega funds still see The top investments early on, because they still deploy in Series C and Series A opportunistically, in some cases even spray and pray if they have their own incubation and acceleration programs, is not true. Actually, we verified that over the last 12 to 13 years. It is not 12 to 13 years in vintage, right? So up to a 2021 vintage fund. So we went basically 12, 13 years back from there. And it’s not true. Actually, the most performing are 0 to 100 and then 100 to 500. And as I said, there’s 100 to 500 in a couple of years actually are a little bit better. Than the $0 to $100 million ones. So that’s the first thing that’s a conclusion. And actually, that’s not shocking. If we remember back in the day, Kleiner Perkins used to raise funds up to $600 million, Benchmark raised their $425 million funds. It seems like the sweet spot for a VC fund would be around $500 million at the top end, like maximum. And now somehow people are saying, well, I’m raising a $3 billion VC fund. It’s like, well, it can’t be a VC fund. The return profile is totally different, right? You can’t deploy that capital just based on early stage investing. And by the way, you’re not seeing the guys at early stage, all that you’re seeing, you’re going to make your returns in mid to late stage, right? Back to what we said at the beginning of the episode. So there’s a little bit of the haves and have-nots there. The big guys are raising more and more money, but they’re no longer venture capital. And I think limited partners that are a little bit more evolved, that are a little bit more conscious of this, that have been in the market longer, are realizing that shift. So it’s like if they want to have the alpha of venture capital, they need to deploy to the sub-$100 million funds or the sub-$500 million funds, right? That’s where they need to actually focus their VC capital. They can still deploy to mega funds, but they’re deploying to a different asset class. They’re deploying to a private equity, mid to late stage asset class, which looks maybe a little bit more like a growth fund or something like that. The second part of differentiation is the honest truth is most VC funds are like, I have proprietary network access, right? I’m ex-Stripe or I’m ex-Google or I’m ex-Facebook or whatever, and I have access to that. I mean, we know proprietary networks from that standpoint are no longer true. The whole thing that created Silicon Valley back in the ’70s of what I used to call the country club deals where there were a few people coming out of the big companies, the Fairchilds of the world, later on the Intels of the world, et cetera, et cetera, that made some money along the way that sort of bootstrapped their next companies, were well-known quantity to the existing VCs and raised money relatively easy on ideas, that doesn’t work anymore. Someone was telling me the other day one interesting thing that I wasn’t quite aware of, a lot of it had to do with the NDAs. I don’t know if you knew this, Bertrand, but like the fact that in California, it was sort of the Silicon Valley community sort of imposed this, we don’t sign NDAs thing and Boston continued signing it. And this whole NDA enforcement issue and non-compete, actually not the NDA thing, but more strongly that California did not enforce non-competes. I could leave Fairchild and start a company that magically was doing something that could be considered competitive to Fairchild. And that was sort of part of the acceleration actually of venture capital in California versus, for example, Boston, which was sort of hand in hand at the beginning.   Bertrand Yeah, I mean, I’m a big, big believer in California success coming from not enforcing or banning non-compete agreements. I think it’s a key part of the game. If you lock people into not doing something similar in the next 6 months to 24 months. And the industry has always been moving fast. So this is a significant time where you are blocked to do something very similar. I think it was really an issue. So I think it’s a key part of the game and it has been there. I don’t know how it started, but I think that non-enforcement of non-compete has been a key part of the success of California. I’m actually pleased to say that Washington State is going in the same direction. They are just signing a non-compete ban. And you might remember that at the federal level, I think in 2024, there was also a ban that was put in place to ban non-compete, but this has been reversed by the courts. So this is not there anymore. So that’s why we see a state like Washington State putting their own ban, and we might see more state by state moving in that direction. I think it was not helping at all, this non-compete. I mean, there is obviously stuff that needs to be done, like you cannot steal secrets, you cannot steal IP.   Nuno Yeah.   Bertrand Even stealing employees, there should be some restraints. We need to find the right balance, but you have to be careful there. That was key for the success of California, and I’m glad to see that this is a trend that’s going to go beyond California. And I hope most states will have a ban on non-compete.   Nuno Maybe just to close on the differentiation process, two things. One, I think there’s this notion When you talk to some LPs, that seems to be a little bit ingrained, some LPs that prefer specialized funds. We’ve also done some significant analysis internally and have talked to a couple of datasets other than our own, or people that own datasets other than our own, and the feedback has actually been not so fast. Actually, generalist funds over time cannot perform specialist funds. There seems to be a little bit of a sweet spot around generalist funds. We like to call ourselves multi-specialized at Chameleon, but ultimately from the perspective of specialized versus Generalist funds, the picture’s not as clear as specialized funds outperform generalists or generalists outperform specialized. We’ve seen there are pockets where actually generalists outperform specialized, in other pockets where specialized of a certain size can outperform generalists. So that’s one topic on differentiation that is a little bit broader. And then the final topic on differentiation, it’s really an industry that hasn’t innovated dramatically on where it creates the most value, which is really the picking stage, right? So it’s having great deal flow, very optimal, productive, efficient due diligence with very few resources and the ability to then get into those deals. That’s where most of the value is created. And then hopefully liquidating the asset if there’s an opportunity to do so at the right time, either through secondary trade sales or an IPO or something else. And what we’ve seen is the industry has innovated very little. I mean, the only thing I could point out in terms of core innovation at the top of the funnel has been the creation of the mega funds, the well-known funds, right? Like a16z, Union Square Ventures, et cetera, et cetera. But there needs to be more innovation on that cycle. And that’s why we certainly at Chameleon believe that the future is to have quant and AI-native VC firms that develop their own tooling, their own platforms. We have Mantis in our case that allow you to have this unfair advantage in how you source deals and how you do due diligence, how you get into the deals, et cetera, and how you take it to the next level. And we think that’s the beginning of the next stage is that the industry becomes more tech-enabled, shockingly enough, an industry that has made all its returns on tech or almost all of its returns on tech. That we need to be more tech-enabled ourselves. But I think the writing is on the wall there, and that will be a source of differentiation certainly over the next 3 to 5 years.   Bertrand One thing the industry has innovated somewhat and maybe could innovate even more is providing liquidity beyond trade sale and an IPO, because it’s clear that if VCs want more liquidity without waiting 18 years, you need that liquidity at different stage, not just when it’s time to do an exit, a full exit for the business. And for employees as well. I mean, it’s one thing to stay for a company for 4 years, which is your typical vesting. Maybe you extend that to 6 years, to 8 years, you have a great time at the company. But to think that maybe you have to stick around for 15 to 20 years in order to get liquidity on your stock options. I mean, that’s too much to ask for most people. I mean, people have a life, they have other things to do, other plans, they might want to move, they come at a different stage of life. So you need to provide them liquidity. The new game is we are not going to exit until 15 to 20 years, else it’s truly unfair. It’s not just unfair, but people will say, you know what, I’m going to go across the street, go work for Amazon or Google. I will have RSUs at best regularly that are liquid, and why bother? I mean, we need to find pathways to liquidity for both investors but also employees. There has been a change in that direction, but I think we need more of this change, and maybe not just reserved for the absolute biggest, most successful companies like OpenAI or SpaceX, but also us as well. Hopefully we can find a way.   Nuno Well, now we have these AI companies that actually grow so fast that they will IPO in one year. Now, isn’t that what’s going to happen? They raise They raised $500 million in Series C or $1.4 billion in Series C, and they’re going to IPO in 2 years. No? Is that not the new reality? I’m being facetious.   Bertrand At the same time, I mean, there are rumors that some of them are going to IPO this year. I mean, we talk about OpenAI, about Anthropic. I mean, OpenAI is quite old, but Anthropic is a relatively new business, quote unquote. So I think it’s a good time.   Nuno The Mega Fund Question So maybe it will be true after all. Moving to the next section, are mega funds still venture capital, Bertrand? Are they still venture capital funds?   Bertrand Yeah, I guess venture capital is a term that can encompass from small to very big funds. I truly don’t know. I mean, once you reach a growth stage, are you truly a VC fund? I don’t know. I think some of these definitions are kind of arbitrary from my perspective. What is clear is that you as a business need different providers of capital. And as we just discussed, you as a business, probably need to keep going and stay private for longer. One reason being, again, there is a tremendous cost to being a public company. There are some true strategic disadvantages. And at the same time, just practically, I mean, you need to get bigger and bigger in order to have a chance of a successful IPO. So you cannot just go IPO at a $500 million valuation. I mean, that’s like committing suicide, at least in the US market on NASDAQ. So my point is, you truly have no choice. You need to extend and If you need to extend, then you need to have capital providers that are there at later stage and therefore have more money. Is it still true venture capital? Is it true venture? I don’t know. At some point, it makes sense that from the startups to the capital providers, everyone adjusts to a reality where the life cycle is getting longer.   Nuno We don’t think it is. We don’t think mega funds are venture capital. We have actually some data that shows that they’re not in terms of actual returns. The alphas you can generate, the IRR that you can generate is actually not comparable. We did some analysis again with some of our datasets and from 2012 to 2022, so that’s the datasets that we used so that we had actual distributions and stuff we could take into account and so on and so forth. And looking at IRR, just to share some numbers in terms of IRR over those 10 years on sub-$100 million funds versus above $1 billion funds, the differences are incredibly stark. And this is true for global and US IRR, right? So just to quote some numbers in terms of average, sub-$100 million funds, global IRR of 22.9%, US IRR of 21.6% versus above $1 billion, 9.1% and 9.0%. Median IRR, if we just looked at median, 7.3% and 16.6% for sub-$100 million funds, 7.5% and 8.1% above $1 billion. Top quartile IRR, sub-$100 million, 31% versus 30.4% US IRR. And then above $1 billion funds, 14.7%, 15.5%. So it’s very clear if you sort of cut this in different ways, averages, medians, top quartiles, et cetera, over all these years that sub-$100 million funds are in a very different asset class than above $1 billion funds. They’re in different alpha that you can generate and so on and so forth. Now to the point you made, Bertrand, I don’t fully disagree with the point you made of the bigger funds should become bigger. I just think they’re becoming different things. Now, again, some of these funds will hide under the facts like, well, wait a second, we have all these assets under management, but they’re over different funds. Sequoia, we’re still raising small early-stage funds, $500, $600 million funds. And then we have larger funds for growth, et cetera, et cetera. Andreessen Horowitz, a little bit less clear what they’re actually doing. We heard that they’ve raised $15 billion across funds. I’m not sure if that’s the exact number at the end of the day. But the point is, if I’m a multi-asset class manager, like early growth, et cetera, et cetera, then it still applies what Nunu is saying. I’m still going after the $500 million, $600 million early-stage funds. Well, not so fast, right? Because you still have all this capital with managing general partners that are maybe across funds for which their incentives in particular, both carry and management fees are coming from the larger funds. Et cetera, et cetera. So there’s necessarily conflicts of interest. In many cases, the funds are just straight up big, right? And so they are above a billion. And so I don’t think a lot of these guys are in early-stage investing anymore, right? It may appear that they are, but I don’t think that’s where the returns necessarily are going to come from. And so if you are a limited partner, if you’re looking at your asset class allocation, again, you’re absolutely free to put money into mega funds because that’s the kind of asset class you want to play in. In terms of a blended private equity asset class that has a little bit of growth, a little bit of whatever, or actually a lot of growth, a lot of late stage, and maybe a little bit of early stage. And I want something that’s a little bit more blended, right? But if I still want the alpha venture capital, I need to deploy to funds that are early stage, right? And that’s like up to $100 million, up to $500 million. I think that’s my two cents on that topic. We see crossover things coming around, like guys who do both public and private markets. Again, that starts feeling a bit like a hedge fund. A lot of these funds have also become RAs, as we discussed earlier. So I feel the writing’s on the wall. The mega funds are going more and more after either some mechanism of edging or a mechanism that’s a little bit more blended in terms of private equity than classic venture capital.   Bertrand Yes, I think a few things. One, if you’re an LP, I can imagine that dealing with multiple $100 million funds might be more difficult. You, you need to know the partners, you need to have some background, uh, visibility. You need potentially to change regularly of VC investments. So I can see some level of simplicity if you just focus on the bigger ones, especially if you have a lot of assets you have to put to work. Another piece of the puzzle, I would guess that the bigger funds are able to return money faster because they are at later stage of the cycle. So instead of that 15 to 18 years, maybe they are more in a 5 to 10 year range, while the smaller funds being there more early might be the one who are taking longer to deliver. So I can see that Yes, there is an IRR picture, but there is also time to liquidity that is not the same. So that can probably also influence. And in terms of crossover PE hybrid model, I mean, for sure we have seen some of the public equity investors doing crossover, meaning going into private equity firms like Coatue, like Tiger Global and others. And for companies that are preparing for IPO, there is a lot of value to work with these firms because they have very good visibility and understanding of the public markets. And their presence in the cap table is also a sign of quality, typically for public market investors. So there is a lot of value and logic for them to be there on both sides of the puzzle. But again, the fact that firms keep delaying IPOs, that the market is not so much startup-friendly, makes this model a bit more difficult. But personally, I think there is value there.   Nuno Yeah, I think on the mega fund, just so that I’m not boo-booing everything, I mean, but there’s definitely angles in terms of the asset class that make a lot of sense. And there’s the scalability of the model. The ability to go after Series B, Series C, as well as mid-stage, as well as late-stage, even secondaries over time, to your point, in some cases even public equities. And that level of skill I think matters. We’ve also seen, as we’ve known, we won’t mention any brands, but people will know who they are, that late-stage hedge funds and investors, even if they’ve done okay-ish in growth in private equity, don’t necessarily do well in venture. So it’s clearly a very different asset class, right? So once you start getting venture teams together, The returns are not quite the same. Actually, sometimes they’re not even quite the same as the growth investments. So clearly they’re very good at the growth side, but not so good in early stage. But definitely there is a case for it. The Case for Smaller…Rightsized Funds But if we switch gears maybe to the small, or I would call right-sized funds, maybe just to quote a couple of numbers and then open up the discussion. Small funds do seem to outperform larger funds. There’s a lot of data in the market that shows some of that dynamic outperformance frequency. All the Very historical numbers from Cambridge Associates from 1981 to 2010. 19 out of 30 vintages were won by sub-$150 million funds. We did our own analysis as I was sharing before. Funds between $0 and $100 won most years between around 2010 and 2021. And the years that they didn’t outperform in terms of investing in the top-performing companies in early-stage Series C, Series A, they were outperformed by the $100 to $500 million funds. The $500 to $1 billion funds and $1 billion or above were never even in the same league in terms of performance, of having identified those top performers in terms of quantity over those early-stage investments. Top 10 funds by vintage, 2004 to 2006, 2016 numbers. Top 10 funds, 73% were sub-$100 million. 2004 to 2016, top 10 funds by vintage, 73% of those were sub-$100 million. So there seems to be a little bit of a case that actually smaller funds, sub-$100 million, sub-$500 million in some cases, are outperforming the larger funds over time. Now, these funds are complex in and of itself. The positive of it is small fund GPs like myself, we are deeply invested in our own funds. We’re not there to just make management fee monies. I mean, we’re not making $1 million, $2 million a year in management fees of salary ourselves, like some of the larger funds. So we are there to really get the carry and be less focused on management fees. And so I think there’s a little bit of alignment around that and really taking that kind of perspective on portfolio construction and liquidation, being also more aggressive on the individual time that we spend with our startups. On the negative side, obviously a lot of these smaller funds, not the case of Chameleon, but others out there are single GPs, very little teams or very small teams. And so it’s sometimes difficult to actually do a lot for portfolio companies as well. And this is where the mega funds, for example, a16z notably would say, hey, we have 600+ people that can support you, right? On market development, business development, communications, talent recruiting, all this stuff. Question mark whether that’s the right way to do it in terms of operating model, if technology is not a better way of supplying that value back to your portfolio companies, or if there’s no better way of doing it. But still, that’s one of the appeals of actually dealing with a larger mega fund if you’re a startup, right? That they will have the resources, also the financial resources to put more capital in you. But also, again, if there’s entrepreneurs listening to this right now, and hopefully there are, it’s a two-edged sword, right? Because if you have Andreessen Horowitz putting money in you, or NEA, or General Catalyst, or whatever, putting money in you on a Series C and then not doubling down on the Series A or the Series B, there will be questions, right? Because like they have the capital, they have other funds, so why the hell are they not putting more money in? Um, so, so it’s a little bit of a two-edged sword.   Bertrand Yeah, I think that one is a pretty big one. And on top of it, as we discussed, some of these big firms have multiple funds managed technically by different teams. So you might have convinced the early-stage teams, they have investors, they’re happy, but you don’t convince the growth-stage firm. As you say, it might raise questions because people might think that there is some communication between the early-stage team and the growth-stage team. So why the heck are they not deciding to invest? And as we also discussed, even worse possible situation, what happens if the growth-stage team has invested in your competitor? It’s even more trouble. So I think trying to understand how firms behave, what’s the reputation of the firm, what’s the reputation of the partner you are working with, I mean, can have tremendous importance and impact. When it’s time for you to work with a firm.   Nuno Indeed. I mean, at the end of the day, we still believe that the smaller fund— we at Chameleon discuss the notion that our limit should be $500 million per fund, right? And that’s the logic of it. We think that model is the model that works well in venture capital. We do recognize, as I said before, why mega funds keep raising more and more money, right? It becomes a harm’s race at that end of the market. As I said, probably a slightly different asset class, or if not a significantly different asset class as well. So seeing a little bit both sides of the market, I mean, we often compete with the mega funds, but honestly, a lot of the mega funds are kind to us and they let us in. And this whole notion of elbows out, we haven’t felt it that much in the market. And people see our value at the table. And in many cases, I, I do see the larger funds more and more seeing the value of smaller funds coming in on the same rounds and even in some cases co-leading early stage rounds like Series C. So it’s not like elbows are out everywhere across the board. So I don’t mean to say this is like an all-out war between small funds and big funds and the small funds need to win or the big funds need to win. I think actually there’s a lot of potential for coexistence. My point is more that the asset classes and the returns are quite different over time, and that’s how I would think through it. And if you’re an entrepreneur, you should think about that as well, right? What are the implications of taking money from certain funds versus others in terms of the expected returns, expected time allocated to you? For example, if you’re not doing very well as a as a company, right? Will the big funds spend the same amount of energy on you if you’re not doing great and all of that? So it’s a little bit sort of a beware, open your eyes, both for limited partners and for startups. What do you actually want, right? What do you want from your VC firm if you’re a startup? And what do you want from your VC firm if you’re an LP?   Bertrand I must say, as an entrepreneur, uh, a board member, I have seen some situations where the bigger funds are actually trying sometimes to elbow out the existing investors. Like, uh, we have that much money to put to work, we cannot do less. And you’re like, yeah, but I don’t need that much money. And then they’re like, okay, just don’t let your existing investors do their pro rata. I don’t think it’s great because an entrepreneur, if your investors, your VCs, trusted you earlier stage when it’s more risky, and when it’s becoming less risky, you don’t give them the right to their pro rata because you have to let this big guy come in. That’s not great. Or even if there is not this pro rata issue, when an investor tries to put more money to work than it’s really necessary, it’s also not a good idea as an entrepreneur to take more capital than you could use. It will dilute you more, it will set higher expectations in terms of valuation, it will push you to use that capital faster than maybe would be reasonable. So I think that’s something you want to be careful with the bigger funds. So don’t talk to funds that are in some ways beyond your stage and try to make it work in that context. Or don’t accept to have your strategy change dramatically for no good reason by funds that just want to put too much money to work in your business. And that for me is surprising because it should also be in their best interest not to invest in businesses that are not ready to accept that much capital. But as we have seen, there were in the past some funds that believe that capital is a moat. Was a good idea. So hopefully, I guess we’re a bit behind that. But yeah, I would say entrepreneurs, be careful, find partners that are the right partners for you at your current stage. Sometimes some big names look great, but at the same time, if it comes with a lot of issues, from too much capital to also taking the risk that these partners don’t understand the stage of the business you are in or your industry, Just be careful. There is a lot of value to have firms that are very focused on your stage, on your industry, are finely attuned to that situation.   Nuno What Comes Next? Maybe to end in terms of sections, what comes next? And maybe we can come up with some predictions that are a little bit provocative on what’s going to happen to the market. You, if you’re listening to us, feel free to interact with us on LinkedIn, on X. If you have our email address, shoot us an email as well. We’d love to hear from you if you think these are the right predictions or if we’re totally off. Maybe I’ll throw in the first one, Bertrand, and we’ll go one by one. So we’ll each put one at the table and see where we head. My first one is that we’ll have a huge culling of VC investors. We had this rapid expansion of the VC asset class with arguably at least tens of thousands of firms globally, maybe even over 10,000 in the US. I think we’ll have a culling and the culling will continue and we’ll have several firms sort of getting eliminated over the next couple of years that will have either because they’re having tremendous difficulty doing their first close in their next fund, or the returns are not there, or it’s a firm that has done 3, 4 funds, but for some reason the returns have just gone out of whack in the last few years during the bull years. And so therefore, actually they can’t justify to raise more funds out there. So I predict there will be a significant elimination of active firms in the next at least 2 to 3 years. So maybe by 2028, and we’ll be below, I don’t know, 30% of number of active firms that we are today. The other side of it is I do think if we look beyond that, 2029, 2030, and so on, we’ll have the reemergence of not micro funds, but nano funds where people will start deploying capital very, very early and writing small angel checks, but doing it in a way that it’s sort of not this cottage industry that we’ve had of angel investors. So I think angel investment will be disrupted by people that will use more and more of the AI toolification out there to actually manage their portfolios of 10, 15, 5K investments in a way that is a lot more professional, creating sort of an advent of nano funds.   Bertrand Yeah, makes sense. On my side, in terms of prediction, I think there is a possibility that the mega fund model keeps expanding and looks more similar over time to some PE models. So do we have the top 10 VC firms that look more like a Blackstone than a Kleiner Perkins or Sequoia used to be? That for me will be an interesting question and development. I think that there is some possibility that it keeps going in that direction. A lot of incentives are pushing things that way.   Nuno My next prediction is that DPI, distributions to paid-in cash on cash, just cash back, will become essential for limited partners. I think TVPI, total value to paid-in, that also has in there, as we just said, paper valuations. There’s a lot of disbelief now around the TVPI metric if there isn’t distributions going alongside it. For those who, again, don’t know what TVPI is, it’s total value paid in, but it also includes DPI. So it’s cash on cash component plus a remaining valuation to paid in, an RVPI. And the problem is the RVPI really, in reality, it’s that kind of on-paper valuation that never gets attributed. I think LPs, they’ve seen the writing on the wall and they’re like, dude, just show me your DPI numbers. I don’t care about TVPI. Some LPs will still ask about TVPI just to make sure that the rest is sort of looking in order. Like, show me the money, show me the cash. Actually, it’s not money, show me the cash, right? I want money back.   Bertrand But that’s an issue. I mean, if you’re supposed to raise financing every 3 or 4 years, good luck getting DPI to show for that. So you need to be at least on your third fund in order to be able to show DPI, I guess.   Nuno I mean, my corollary to that, Bertrand, is if you allow me just to have a corollary kind of prediction, is that we’ll see certainly for funds like $50 million and above, $100 million, $200 million, et cetera, even increased concentration, right? I really need to have anchors that believe in me over time. And we might start having, again, the advent— we had it some decades ago, the advent of cap table kind of VCs, right? Like Sutter Hill Ventures, right? Where they’re not really raising funds anymore. And so we might have the advent of that, that we’ll have structures that are created that have more permanent capital allocated to them, or at the very least more concentrated capital by very few players.   Bertrand Interesting. Me on my side, as I shared before, I believe secondaries are, are important and here to stay. Um, in the past, some could argue, is it a distress signal or something? I, I don’t think it’s true anymore. In a world where your average startup might take 15 to 18 years to exit through M&A or IPO, we need to have other options. For funds, for employees, they cannot be expected to stick around for so long and have no liquidity. I mean, it’s just pure madness. It’s just bad alignment at some point to do that. So I think secondaries are becoming the third liquidity pathway for VCs, for employees, and it should be more and more a key part of the game, a key infrastructure in the VC/startups tech industry.   Nuno I mean, on specialized versus generalist funds, I believe we’ll continue seeing the coexistence of those two models where the specialized funds will in many pockets actually outperform generalist funds, but where we’ll continue seeing that the large franchises, the tier one franchises will likely be generalist funds. I mean, we just saw it in the cycle. The AI cycle went upon us. We had a 2021 fund. We could easily adapt and go into AI and figure out that AI was growing very fast. I mean, if you have an ultra-specialized fund and that’s your remit and that’s the only thing you can invest on, very difficult to change even during our investment period. I will put a caveat on that. We don’t call, for example, ourselves at Chameleon generalist. We call ourselves multi-specialized because our scoring models for the verticals that we track are specialized within Mantis. Because the partnership is specialized, we all focus on different areas. And because we have the Kin network that allows us to tap into that level of expertise, Again, I think the world will be specialized coexistence. Some pockets specialized will do very well, certainly on the smaller fund size, but the big franchises will likely look a little bit more generalist. And as I said, multi-specialized from our perspective is the future. We’ll start seeing more and more funds that are multi-specialized like ourselves. Do you want to talk about AI and how it’ll distort the metrics? No.   Bertrand Yes. I think AI is an exciting moment in the tech industry. It feels in some ways that the same way we had a big distortion coming with COVID and work from home in 2020, 2021. 2021, where suddenly everyone and their mother will build a SaaS company or invest in a SaaS company. AI feels a bit of the same. I mean, to be clear, I truly believe it’s deserved. I mean, we are facing a dramatic shift in how computing is being done in terms of value you can get from software. So at the same time, AI will probably distort this matrix for a long time. We clearly see a split where investments are going, in what startups are being created. So I think, yeah, we will see some distortion. And we know that maybe 50% of all deal value is going to AI in 2025. We have seen single rounds reaching 40 billion, like to OpenAI. We have seen, as you discussed, some seed stage investment of 400 million. So AI investing and AI startups are definitely a beast on their own. And will distort VC metrics for a long time. And we might need two sets of metrics in parallel, you know, AI versus everything else. So that would be an interesting bifurcation in the industry in some ways. I would say it’s fair to separate AI versus non-AI. We reach a point where it’s two different beasts.   Nuno Conclusion So in conclusion, AI has changed the world and it’s changing VC as well, as we discussed earlier in the episode. We have a tremendous momentous occasion for the asset class where venture capital is really bifurcating into very large funds, which no longer are in venture capital or seemingly may be distributed between different asset classes, and the smaller funds, sub-$500 million and sub-$100 million, that keep having the better returns, but also with much smaller scale. We’re seeing a culling of the industry where the industry is definitely getting smaller and smaller and more concentrated at both ends, number of VC firms, as well as a number of limited partners per fund and the interest that some of these limited partners have of being more and more concentrated in their own portfolio allocations. And last but not the least, the discussion around specialized versus generalist, where it seems like there’s some clear winners on some asset classes, on some sizes, in some industries, but on others, there’s other kinds of winners. And so maybe the future is multi-specialized, as I framed at the end. Thank you so much for listening. If you want to check us out and if you want to comment, feel free to send us messages on X, LinkedIn, to both myself and Bertrand, as well as send us an email. Thank you so much, Bertrand.   Bertrand Thank you, Nuno.

Nemetschek just dropped $2.4B on a software company nobody's heard of. This week on Bricks, Bucks & Bytes, Owen, Patric and Martin dug into the biggest AEC deal of the year, plus two robotics founders changing how we build. Tune in to find out about: ✅ Why HCSS is the smartest AEC acquisition of 2026✅ Why Kawazu ditched construction for oil and gas✅ Mesh Robotics on solving rebar, the ugliest problem in construction✅ Martin Fisher's rule: 80% data, 20% AI #aec #construction #constructiontech #bricksbucksandbytes #bricksandbytes #ai #vcOur Sponsors:BreadCrumb- 50,000+ projects globally. All running safer, faster, with Breadcrumb. - breadcrumb.coAphex is the multiplayer planning platform where construction teams plan together, stay aligned, and deliver projects faster – check out aphex.coArchdesk - “The #1 Construction Management Software for Growing Companies - Manage your projects from Tender to Handover” check archdesk.comChapters00:00 Intro00:56 Celebrating Milestones and Future Aspirations03:25 Insights from the Martin Fisher Event06:33 The HCSS Acquisition by Nemetschek12:09 Strategic Implications of Acquisitions in Construction Tech30:04 The Genius of Brad Jacobs30:40 Meditation and Success31:49 Exploring AI Tools and Credits35:44 The Future of Urban Mobility: Glideways45:01 Innovations in Nuclear Power50:14 Understanding Labor and Scaffold Costs52:36 The Evolution of Robotics in Construction54:29 Data-Driven Insights in Material Flow56:44 The Role of Transparency in Construction57:58 Scaling Robotics Solutions in Construction1:01:09 The Intersection of Robotics and Rebar1:05:31 Bridging the Gap from Lab to Market1:11:46 The Future of Assembly in Robotics

Users are fed up with misinformation and AI slop cluttering their feeds. SaySo is a new short-form video app that delivers news from vetted creators and journalists. Also, Loop, the San Francisco startup, closed a Series C funding round led by Antonio Gracias' firm Valor, which is a major backer of xAI. Learn more about your ad choices. Visit podcastchoices.com/adchoices

Shoot us a Text.Episode #1319: Dealers prove growth doesn't require more rooftops, Amazon inches into car sales with real-world friction, and Slate Auto raises $650M to bring its affordable EV vision closer to production reality.Show Notes with links:Forget “grow or die”—2025 proved you can win without adding rooftops. Many Top 150 groups drove serious gains through operational discipline, not acquisitions, signaling a shift toward smarter, not just bigger, dealership strategies.52 groups grew new-vehicle sales with zero footprint change, pointing to stronger same-store execution.High performers leaned into used-car ops, inventory availability, and internal GM development.Great Lakes Auto Group climbed 19 spots to #88, boosting volume 28% while holding steady at nine stores.Late-year acquisitions (Q4 closings) meant organic performance—not M&A—drove most gains.“We think that scale helps… but I don't think it's absolutely necessary,” said Hudson Automotive (#11) CEO David Hudson.Amazon is upping its new-car retail platform, and yes, you can now buy a Corvette there. What started with Hyundai has expanded to include multiple brands, bringing digital-native shopping into a $1.3T dealership market.Amazon Autos now features Hyundai, Kia, Mazda, Subaru, Chevrolet, and Jeep in 130+ cities.Customers can browse, price, finance, and start paperwork online, reducing time in-store—not replacing it.Dealers pay to list inventory, gaining high-intent traffic from Amazon's massive audience.Early friction like inventory sync issues and incomplete deal structures highlights the complexity of auto transactions.“Customers have a level of comfort with Amazon… but it's definitely just in the starting phase,” said dealer Matthew Phillips.Slate Auto just locked in $650M in Series C funding, keeping its low-cost EV truck plans on track—and putting a spotlight on its next big milestone: production.The funding supports next-stage development and production ramp at its Indiana facility.Slate just crossed the 160K reservation mark and still targets late 2026 deliveries, with preorders expected to open in June.The truck will start at a mid-$20K starting price, using a stripped-down base model with modular add-ons that let customers upgrade into things like a 5-seat SUV or fastback configuration.The company plans to invest $400M in its plant, creating 2,000+ jobs.“We will deliver Slate Trucks at nearly half the cost of the average new vehicle—as promised,” said President Chris Barman.Join Paul J Daly and Kyle Mountsier every morning for the Automotive State of the Union podcast  as they connect the dots across car dealerships, retail trends, emerging tech like AI, and cultural shifts—bringing clarity, speed, and people-first insight to automotive leaders navigating a rapidly changing industry.Get the Daily Push Back email at https://www.asotu.com/JOIN the conversation on LinkedIn at: https://www.linkedin.com/company/asotu/

On today's Technology Report, Ed Mehr, the co-founder and CEO of Machina Labs, joins Defense & Aerospace Report Editor Vago Muradian to discuss the innovative Chatsworth, Calif., firm's AI-driven intelligent factories approach to increase commercial and defense production volumes at lower costs that attracted $124 million in Series C funding from investors including Lockheed Martin Ventures and Toyota to underwrite a new 200,000 square foot facility; the imperative to adopt AI as well as cutting edge machining technology and greater robotization to increase production; scaling the business as a manufacturer and technology provider; importance of designing for production and greater component flexibility to speed manufacture; how to better manage supply chains; how to manufacture at the edge; and the critical role of allies and partners in delivering capability.

What does it actually take to make fractional work sustainable, not just for a few months, but as a real career path?In this special solo session, Ben Erez shares everything he's learned about fractional product work from two distinct perspectives: as Head of Product at Continuum (a Series A marketplace for fractional executives that eventually shut down) and from his own two years doing fractional work. He walks through the foundations that set you up for success, how to actually find work and structure engagements, and most importantly, how to make it sustainable and not die. Ben is ruthlessly honest about what worked, what failed, and why intentionality is the difference between loving fractional work and hating it.He covers the counterintuitive truth that narrow positioning beats broad positioning (using Phil Carter as a case study), why you need to learn to enjoy marketing yourself or you'll burn out, and the three non-negotiable requirements for any engagement (relevant expertise + urgent need + budget). Ben shares real revenue data showing the spiky, unsteady income reality, explains the “feast or famine” trap that kills most fractional careers, and why he's skeptical that demand aggregators will ever crack this market at scale. Plus, tactical advice on pricing, the transition from being helpful for free to asking for money, and why referrals beat every other channel.If you're exploring fractional work as a side hustle, between full-time roles after a layoff, or considering it as a long-term path—this session is for you.All episodes of the podcast are also available on Spotify, Apple and YouTube.New to the pod? Subscribe below to get the next episode in your inbox

C is Home. Root. River. Flowing as far back as you can, then further, and further still. Breaking the banks, spreading across the fields to the foot of the mountain. You inhale until you can hold no more air. Then, you slowly exhale until you can release no more, revealing the hidden life, the reeds and the muck. This way, you find your bed, your walls, and your ceiling. You feel every shake and every glitch as bare electrical cables, peeling wallpaper, dust, dirty pans in the sink, all mounting as you write by candlelight. In January 2026, Mikey Kirkpatrick played a C every day on a flute. These are all of the Cs and their ambiences stitched together into one continuous flow. This year he will step up the chromatic scale, one month at time, playing and recording a note a day; 12 months, 12 notes. In January 2027, he will arrive at the octave of C, the same note, but different. This is part of a residency- a residency that he invented for himself. He creates every day, even if he only has two minutes spare. This note a day project has grown from this. Everything is documented on a simple substack page (http://mikeykirkpatrick.substack.com). The Residency is open to all. He writes: You are invited to be a resident artist; resident in your body, your home, your streets and parks, seas and dreams. Gather your tools, instruments and materials and begin. Mikey Kirkpatrick, based in London UK, is a flautist, singer, composer, improviser, writer and educator. He has performed on hundreds of streets and stages internationally, composed for film, stage and radio, and collaborated extensively with musicians and artists across different disciplines. He is the founding director of Alchemy at Goldsmiths University, where he is also an associate lecturer. Mikey's practice both as an artist and an educator focuses on how deeper listening and musicality can be beneficial and healing for individuals, relationships, communities, society, and the environment at large.

It's been a week since missing episodes of The Daleks' Master Plan revealed themselves to the public to great excitement and joy and indeed a Film of Fabulous! screening of said episodes on April 4 sold out near-instantly upon its announcement. We also have news of an ex-Google chief possibly taking the helm at the BBC, The War Between still being MIA from Disney+, the usual Big Finish over-examining, and our feature interview with The Long Game and Pull to Open author Paul Hayes, here to discuss the production of 1956 CBC production Flight Into Danger, starring one James Doohan and produced by future (at the time) Doctor Who producer Sydney Newman! CanCon, baby! Links: Support Radio Free Skaro on Patreon Film is Fabulous statement about the Doctor Who recovery Film is Fabulous short restoration clips Film is Fabulous Daleks' Master Plan screening Apr 4, sold out Gallifrey One 2027 tickets on sale BBC Close To Hiring Ex-Google Chief Matt Brittin As Next Director General The War Between the Land and the Sea not on the Disney+ April schedule Paul Cornell's Saucer Country comic optioned for TV series Big Finish: David Tennant returns as the Tenth Doctor, new 12-part series coming 2027 Big Finish: The First Doctor Adventures: Beware the City of Illusions out now Big Finish: UNIT – Brave New World: Knightfall out now Blake's 7 Series C available to preorder, release date TBD Interview: Paul Hayes Flight Into Danger Doctor Who: Star Flight by Paul Hayes

This week's Espresso covers news from Meddi, MetaBix, Handle, and more!Outline of this episode:[00:30] – Azos raises $24M Series C[00:44] – Meddi raises $7.8M from GNP Seguros[00:59] – Deeptech startup metaBIX raises $1.3M[01:10] – Handle raises $6M seed round led by Andreessen HorowitzResources & people mentioned:Startups: Azos, Meddi, metaBIX Biotech, Handle,VCs: Kaszek, Kevin Efrusy, GNP Seguros, Médica Móvil, Dalus Capital, EWA Capital. Andreessen Horowitz

Has your hiring process kept up with the industry's AI leaps, or are you still interviewing like it's 2022?Today's AI-driven landscape means the skill gap between a good engineer and a great one widens every day, and the great ones can be difficult to find. But how do you choose – and hire – the best in the business? How do you find those elusive engineers who can skilfully handle multi-agent workflows, ship in hours what used to take weeks, and add an AI-focused competitive edge to your startup?In this episode, Chris Saad and Yaniv Bernstein are joined by Matt Cook, an expert in hiring the best engineers in the business, and co-founder of Scouut – one of Australia's most respected engineering recruiters for early-stage startups. Matt works with pre-seed through to Series C companies and has a front-row seat to how AI is reshaping what ‘great' looks like in engineering, and how founders and hiring teams alike can keep pace.They cover what's fundamentally changed in engineering hiring, what hasn't, and how to build a small, elite team that punches well above its weight.In this episode, you will:Understand why AI is widening the gap between great and average engineersDiscover the three questions to ask any Big Tech candidate to determine if they'll thrive in a startupFind out why much more is expected of ‘senior' engineers, and what that means for foundersHear how the best startups are now testing for AI competency in interviews, not just coding abilityLearn why smaller teams, higher salaries, and generous token budgets are the new arbitrage for attracting elite engineersUnderstand why you should be designing the role for the people you want, and making your AI-forward culture clearly visibleResources mentioned in this episodeScouut (Matt Cook's engineering recruitment firm for early-stage startups): https://scouut.com.au/Matt Cook on LinkedIn: https://www.linkedin.com/in/matthewmarkcook/Loki (Australian startup referenced for their public AI-first hiring stance): https://www.itsloki.com/Solid (lightweight vibe-coding tool mentioned by Chris): https://trysolid.com/The Pact Honor the Startup Podcast Pact! If you have listened to TSP and gotten value from it, please:Follow, rate, and review us in your listening appSubscribe to the TSP Mailing List to gain access to exclusive newsletter-only content and early access to information on upcoming episodes: https://thestartuppodcast.beehiiv.com/subscribeSecure your official TSP merchandise at https://shop.tsp.show/ Follow us here on YouTube for full-video episodes: https://www.youtube.com/channel/UCNjm1MTdjysRRV07fSf0yGg Give us a public shout-out on LinkedIn or anywhere you have a social media followingKey linksThis episode of the Startup Podcast is sponsored by .tech domains. Forget weird prefixes and creative misspellings; the availability for .tech domains is simply way better than .com. For a clean name that highlights your tech credentials, get a .tech domain at your favorite registrar.This episode of the Startup Podcast is sponsored by Vanta. Vanta helps businesses get and stay compliant by automating up to 90% of the work for the most in demand compliance frameworks. With over 200 integrations, you can easily monitor and secure the tools your business relies on. For a limited time offer of US$1,000 off, go to ⁠⁠⁠⁠https://⁠www.vanta.com/tsp⁠⁠⁠⁠⁠Get your question in for our next Q&A episode: https://forms.gle/NZzgNWVLiFmwvFA2A The Startup Podcast website: https://www.tsp.show/episodes/Learn more about Chris and YanivWork 1:1 with Chris: http://chrissaad.com/advisory/ Follow Chris on Linkedin: https://www.linkedin.com/in/chrissaad/ Follow Yaniv on Linkedin: https://www.linkedin.com/in/ybernstein/Producer: Justin McArthur https://www.linkedin.com/in/justin-mcarthurIntro Voice: Jeremiah Owyang https://web-strategist.com/

Silicon Valley loves billion-dollar moonshots and AI darlings. Sam Gerstenzang and Dan Friedman are doing something different—they're starting medical spas and funeral homes.On this episode of AI & I, Dan Shipper sat down with Gerstenzang and Friedman, partners at Boulton and Watt, which they call the "world's slowest startup incubator." Their model: Come up with an idea, achieve five or 10 million dollars in revenue themselves, then hand it off to a CEO who can take it to the next stage. They've used this playbook to build Moxie, a Series C company that helps nurses open their own medical spas, now with 600-plus customers and a 200-person team globally. Their second company, Meadow Memorials, is a contemporary funeral home with no physical real estate. It has become the largest provider of funeral services in California.Both businesses launched right around the arrival of ChatGPT—and neither was built with AI in mind. So how are they thinking about AI inside companies where the core work isn't going to change? In this conversation, Gerstenzang and Friedman share how they built an AI agent called Matthew Bolton to power their customer discovery process, why synthetic customer calls completely failed for them, and why they believe you shouldn't give anyone credit for using AI.If you found this episode interesting, please like, subscribe, comment, and share!Want even more?Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt. It's usually only for paying subscribers, but you can get it here for free.To hear more from Dan Shipper:Subscribe to Every: https://every.to/subscribeFollow him on X: https://twitter.com/danshipperIntent is what comes after your IDE. Try it yourself: augmentcode.com/intentHead to granola.ai/every to get 3 months free.Ready to build a site that looks hand-coded—without hiring a developer? Launch your site for free at www.Framer.com, and use code DAN to get your first month of Pro on the house.Timestamps00:00:00 — Introduction and how Sam and Dan's paths first crossed00:01:40 — What it means to be “the world's slowest incubator”00:04:50 — Why Bolton and Watt runs companies to several million in revenue before handing off to a CEO00:07:30 — How specialization across the founding journey creates advantages00:10:40 — Building AI-durable businesses versus AI-native ones00:16:10 — How an AI agent transformed their customer discovery process00:19:30 — Where synthetic customer calls completely fail00:29:30 — Deploying AI inside established companies00:32:30 — Why newer projects see huge gains from AI while mature companies see 10 percent00:37:00 — A preview of what's next for Bolton and Watt

Oxide raised a truckload of capital a few weeks ago to fund the business for the foreseeable future. Bryan and Steve describe the raise, and Adam poses the best the best (and worst) questions scraped from Hacker News.In addition to Bryan Cantrill and Adam Leventhal, we were joined by Oxide CEO, Steve Tuck.Previously on Oxide and Friends:OxF s01e25 - Tales from the Bringup LabOxF s04e30 - Intel after GelsingerOxF s05e24 - Oxide's $100M Series BOxF s02e18 - Silicon Valley Bank with Eric VishriaOxF s05e28 - Systems Software in the LargeMentioned during the show:Oxide Blog: Our $200M Series COxide is hiring!If we got something wrong or missed something, please file a PR! Our next show will likely be on Monday at 5p Pacific Time on our Discord server; stay tuned to our Mastodon feeds for details, or subscribe to this calendar. We'd love to have you join us, as we always love to hear from new speakers!

Today, I'm speaking with Richard Craib, the CEO and founder of Numerai.If you've heard of Numerai before and thought of it as an interesting experiment at the intersection of data science and crypto, it's worth updating that mental model. Over the last few years, Numerai has quietly grown from roughly $60 million in assets to over $600 million. JPMorgan has invested and secured $500 million of capacity, and Numerai recently raised a Series C at a $500 million valuation led by top university endowments. This is no longer a toy project. It is a real, institutional-scale market-neutral hedge fund with a very unconventional engine.In this conversation, we go deep into how Numerai actually works. Richard walks through the core insight behind Numerai's design: that crowd-sourced alpha only works if incentives are aligned, not just participation. Simply opening up data and ranking models creates incentives to game the system, not to produce durable signals. That realization led to the introduction of the Numeraire token. By forcing researchers to stake real capital behind their predictions, Numerai shifts from a leaderboard-driven experiment to a capital-weighted signal engine. Instead of rewarding activity, the system rewards conviction, accountability, and uniqueness, creating a self-filtering model that naturally reduces noise and discourages the behaviors that caused earlier crowd-sourced platforms to fail.We also talk about portfolio construction and risk management, including how Numerai neutralizes common factor exposures, what went wrong during the 2023 drawdown, and how those lessons reshaped their approach to diversification and concentration. Finally, we look forward, covering the limits of crowd-sourced modeling, the next frontier for Numerai's research ecosystem, and how Richard sees AI agents reshaping model development.Please enjoy my conversation with Richard Craib.

The Information's Sri Muppidi talks with TITV Host Akash Pasricha about OpenAI finalizing its massive $100 billion funding round and what the new Series C structure means for a 2026 IPO. We also talk with Ann Gehan about Hungryroot's potential public offering and the resurgence of high-quality consumer IPOs, Avery Marquez about the "race to go public" between AI titans, and we get into AMD's strategic financial backstop for Crusoe with Miles Kruppa. Finally, we discuss the "SaaS is dead, long live SaaS" shift with Dallas Dolen from PwC.Articles discussed on this episode: https://www.theinformation.com/articles/openai-finalizing-first-commitments-100-billion-mega-roundhttps://www.theinformation.com/articles/hungryroot-posts-55-revenue-growth-eyes-potential-2026-ipohttps://www.theinformation.com/articles/amd-backstop-300-million-crusoe-loan-following-nvidia-playbookSubscribe: YouTube: https://www.youtube.com/@theinformation The Information: https://www.theinformation.com/subscribe_hSign up for the AI Agenda newsletter: https://www.theinformation.com/features/ai-agendaTITV airs weekdays on YouTube, X and LinkedIn at 10AM PT / 1PM ET. Or check us out wherever you get your podcasts.Follow us:X: https://x.com/theinformationIG: https://www.instagram.com/theinformation/TikTok: https://www.tiktok.com/@titv.theinformationLinkedIn: https://www.linkedin.com/company/theinformation/

Jeremy Au and Mike Mate connect personal career risk with the structural limits of the Philippine startup ecosystem. Mike explains how jumping from law into finance shaped his tolerance for uncertainty, why venture capital requires emotional endurance, and how AI mirrors past industrial shifts. They examine why Southeast Asia imports frontier technology, why Philippine consumer startups struggle to scale, and why grit remains the region's unfair advantage. The discussion ties personal courage to ecosystem maturity, arguing that founders and investors both survive through disciplined risk. 05:10 Investment banking exposed the limits of deal work: Closing transactions felt empty because he never saw what happened to companies after. 10:00 Venture capital demands emotional endurance: Allocating high risk capital requires custody of LP money and deep founder trust. 12:20 AI mirrors the steam engine moment: Technology removes intellectual limits the way railroads removed physical limits. 14:50 AI growth is exponential not gradual: Decades of change now compress into a few years. 18:30 Consumer startups define the Philippine opportunity: Strong demographics exist but iconic exits remain missing. 21:20 The valley of death blocks late stage growth: Series C companies stall without foreign capital. 25:00 Filipino founders survive through grit: Cultural obligation to family and employees fuels persistence under pressure. Watch, listen or read the full insight at https://www.bravesea.com/blog/engineering-soft-landings Get transcripts, startup resources & community discussions at www.bravesea.com WhatsApp: https://whatsapp.com/channel/0029VakR55X6BIElUEvkN02e TikTok: https://www.tiktok.com/@jeremyau Instagram: https://www.instagram.com/jeremyauz Twitter: https://twitter.com/jeremyau LinkedIn: https://www.linkedin.com/company/bravesea English: Spotify | YouTube | Apple Podcasts Bahasa Indonesia: Spotify | YouTube | Apple Podcasts Chinese: Spotify | YouTube | Apple Podcasts Vietnamese: Spotify | YouTube | Apple Podcasts #VentureCapital #StartupEcosystem #PhilippinesStartups #FounderGrit #StartupFunding #SoutheastAsiaTech #AIandInnovation #StartupGrowth #InvestorMindset #BRAVEpodcast

Aman Verjee, Founder and General Partner at Practical Venture Capital, shares his view of how venture capital has evolved over the past two decades and why secondary markets now play a critical role in the ecosystem. Drawing from his time at PayPal, eBay, and Sonos, Aman explains how companies today stay private far longer than they used to, what that means for early investors and employees, and how thoughtfully structured secondary transactions can reduce friction and misalignment on the cap table. He also challenges popular narratives around tech bubbles, walking through historical examples to explain why today's AI-driven market looks fundamentally different.In this episode, you'll learn:[01:11] Aman's journey from Wall Street to Practical VC[03:40] What made the early PayPal team exceptional[06:32] Follow the customer, not the original plan[10:44] Why are startups staying private longer today?[11:17] What secondary transactions actually are[18:41] How founders should handle secondary requests[26:11] Are we in a tech bubble today?The nonprofit organization Aman is passionate about: AYSO (American Youth Soccer Organization)About Aman VerjeeAman Verjee is the Founder and General Partner of Practical Venture Capital, a secondary-focused fund providing liquidity to early investors in late-stage private companies. Before launching Practical VC, Aman spent over a decade in finance and operations roles at PayPal and eBay, joining PayPal in 2001 before its IPO and witnessing its transformation from a money-beaming mobile app to the dominant payment platform for eBay. Earlier, he worked in investment banking in New York after studying economics at Stanford and constitutional law at Harvard Law School. Aman was recruited to PayPal by Peter Thiel and worked directly for David Sachs during the company's pivotal early years. Now partnering with Dave McClure, he focuses on Series C and D investments in SaaS and FinTech companies with $200M+ in revenue and clear paths to liquidity within 5-7 years. He's also writing a book on the history of financial bubbles and co-hosts the Trading Places podcast, analyzing private company valuations.About Practical Venture CapitalPractical Venture Capital is a secondary-focused venture firm that provides liquidity solutions for early investors, employees, and funds. Operating with a 7-year fund structure instead of the traditional 10-15 years, Practical VC targets 20-40% discounts to last-round valuations in Series C and D companies with $200M+ in revenue and clear paths to exit. The firm specializes in SaaS and FinTech but has made exceptions for exceptional opportunities like SpaceX, now their biggest winner despite violating their typical investment criteria. Founded by Aman Verjee and Dave McClure, Practical VC evaluates roughly 50 companies at any given time, making 5-10 investments annually. The firm also offers SPVs for deals that don't fit their main fund and covers LATAM opportunities through an operating partner in Argentina. Their approach recognizes that modern venture capital requires new liquidity solutions as companies like SpaceX (23 years private), Airbnb (17 years), and Palantir (20 years) redefine what "patient capital" means.Subscribe to our podcast and stay tuned for our next episode.

Maxima is building AI agents that automate enterprise accounting while maintaining the auditability and control standards finance teams require. In a recent episode of BUILDERS, we sat down with Yogi Goel, CEO and Co-Founder of Maxima, to explore his eight-year journey at Rubrik from Series C through IPO, and how those lessons shaped his approach to solving the 70-80% of finance time currently wasted on manual work. Topics Discussed: Why Rubrik's approach—entering stagnant markets with first-principles thinking—became Maxima's blueprint Securing $3K-$5K POC commitments from Figma mockups before writing code Why Scale AI and Rippling rejected a point solution and demanded 3-4 modules from day one The compound startup model: building multiple products simultaneously to meet buyer expectations How 17% of CFOs are adopting AI tools today (vs 51% in software development) Why finance teams view AI agents as "digital college freshmen" who need proof of work Hiring from YouTube Studios, Apple, and Robinhood instead of legacy finance software companies How NetSuite World conference booth sizes revealed the data integration infrastructure gap The $3K-$5K validation threshold that proved finance pain was urgent enough to pay pre-product GTM Lessons For B2B Founders: Demand generation unlocks engineering potential: Yogi learned from his Rubrik mentors: "focus on demand and if you have great engineers then they will solve the problems." Maxima built products in 2-3 months they didn't initially know were technically feasible—because customer demand pulled the engineering team forward. For founders with strong technical teams, customer demand should drive the roadmap, not engineering's comfort zone. Trust your engineers to solve hard problems when customers are waiting. $3K-$5K is the pre-product validation threshold: Before writing any code, Yogi secured POC commitments at this price point based solely on Figma mockups. This isn't about revenue—it's about proving urgency. Verbal interest means nothing. Small pilot commitments mean "we'll try it someday." But $3K-$5K pre-product means "this problem is urgent enough to pay before seeing a working solution." Use this threshold to separate real pain from polite interest. Sophisticated buyers will reject your narrow MVP: Scale AI and Rippling told Maxima explicitly: "If you will only build this one thing, we will not buy. You have to commit to building three, four modules." Conventional wisdom says start narrow, but enterprise buyers with complex workflows won't adopt point solutions that create new integration headaches. When sophisticated buyers articulate their real buying criteria, ignore the startup playbook. Yogi built a "compound startup" with 4-5 modules from day one because that's what the market demanded. Target acute pain over easy access: Early-stage companies (10-30 people) were easier to reach but finance wasn't urgent enough. At that scale, it's "build product, ship product"—finance operations aren't broken enough to warrant urgent attention. Companies at 500-1,000+ employees have finance teams drowning in manual work that prevents strategic contribution. Target where pain justifies urgent action and budget exists, not where calendar access is easiest. Hire intensity and first-principles thinking over domain knowledge: Maxima deliberately hired zero engineers from legacy finance software companies. Their frontend engineer came from YouTube Studios. Others came from Apple, Robinhood, Netflix—none with financial product experience. Yogi's three hiring criteria: "incredible intensity, huge confidence in themselves, and fast thinking mode." Domain expertise creates pattern-matching to old solutions. First-principles thinking creates breakthrough products. One team member didn't finish high school but is "one of the best out there." Make AI explainable or finance teams won't adopt: Finance teams adopted faster than expected because Maxima showed every calculation step. "If they can prove by looking at the Math, you know, 18 plus 88 plus 36 is X. And I can see the step of the work, they are willing to give it to them." This isn't about fancy UX—it's about auditor-grade proof of work. Finance professionals won't trust black box outputs. Build transparency into the product architecture, not as an afterthought. This explainability became Maxima's competitive moat. Conference booth sizes reveal infrastructure gaps: At NetSuite World, the largest booths weren't ERP vendors or payment processors—they were data integration companies. This single observation validated that enterprises are desperately solving data fragmentation problems. Companies manually download from Stripe, Snowflake, Salesforce weekly to build Excel pivots. Maxima invested in upstream integrations as core infrastructure from day one. Use industry conferences to validate where companies are spending money on workarounds—that's where infrastructure gaps exist. // Sponsors: Front Lines — We help B2B tech companies launch, manage, and grow podcasts that drive demand, awareness, and thought leadership. www.FrontLines.io The Global Talent Co. — We help tech startups find, vet, hire, pay, and retain amazing marketing talent that costs 50-70% less than the US & Europe. www.GlobalTalent.co // Don't Miss: New Podcast Series — How I Hire Senior GTM leaders share the tactical hiring frameworks they use to build winning revenue teams. Hosted by Andy Mowat, who scaled 4 unicorns from $10M to $100M+ ARR and launched Whispered to help executives find their next role. Subscribe here: https://open.spotify.com/show/53yCHlPfLSMFimtv0riPyM

SpaceX soll mit XAI verschmolzen werden. Anthropic überrascht und sammelt statt 10 Milliarden gleich 20 Milliarden ein, prognostiziert 55 Milliarden Umsatz für 2027. OpenAI plant sein IPO für Q4 2026 und arbeitet an einer 100-Milliarden-Runde mit Amazon, Nvidia und Softbank. Microsoft: 45% der zukünftigen Cloud-Umsätze kommen allein von OpenAI. Tesla meldet eines der schwächsten Quartale – in Europa brechen die Verkäufe um bis zu 70% ein wegen Elon Musks politischem Engagement. SAP stürzt 16% ab nach enttäuschenden Cloud-Zahlen. Der CISA-Chef hat vertrauliche Homeland-Security-Daten in ChatGPT hochgeladen. Meta testet Premium-Abos für Instagram, Facebook und WhatsApp. Unterstütze unseren Podcast und entdecke die Angebote unserer Werbepartner auf ⁠⁠⁠⁠⁠doppelgaenger.io/werbung⁠⁠⁠⁠⁠. Vielen Dank!  Philipp Glöckler und Philipp Klöckner sprechen heute über: (00:00:00) Intro & Dry January (00:02:48) Christian Lindner blockt Pip (00:05:58) Elon Musk: SpaceX-XAI-Tesla Merger? (00:12:10) Apple kauft QAI für 2 Mrd. (00:15:02) Meta testet Premium-Abos (00:17:53) KI in der Produktentwicklung (00:21:04) OpenAI IPO-Pläne & 100 Mrd. Runde (00:23:58) Aleph Alpha (00:55:26) Neom/The Line (00:59:27) Microsoft Earnings (01:05:22) Meta Earnings (01:11:03) Tesla Earnings (01:18:10) SAP Earnings (01:19:29) Samsung Rekordquartal (01:21:30) Verbraucherschutz: Signal-Phishing bei Journalisten (01:24:29) Tokenisierte Wertpapiere und Regulierung (01:25:33) Grok/X: Visa & Mastercard ignorieren CSAM (01:28:23) Meta: Zuckerberg ignorierte KI-Chatbot-Warnung (01:30:44) Data Center Werbekampagnen (01:32:03) Shein EU-Untersuchung & Meta WhatsApp-Zugriff (01:34:28) Das Letzte: Melania-Film & CISA-Chef ChatGPT-Fail (01:40:45) Politische Spenden und deren Einfluss (01:45:31) RobCo 100 Mio. Series C (01:53:28) Apple Earnings (01:55:51) US Trade Deficit auf Höchststand Shownotes SpaceX erwägt Fusion mit Tesla oder xAI - bloomberg.com Apple Q.AI Übernahme - ft.com Meta to test premium subscriptions - techcrunch.com Anthropic Funding - ft.com Anthropic Hikes 2026 Revenue Forecast 20% - theinformation.com OpenAI und Anthropic im Rennen um IPO - wsj.com Nvidia, Microsoft, Amazon in Talks to Invest Up to $60 Billion in OpenAI - theinformation.com Amazon plant Investition in OpenAI - wsj.com OpenAI plant biometrisches soziales Netzwerk gegen X-Bot-Problem - forbes.com Bosch verkauft Aleph Alpha Anteile an Schwarz Gruppe - faz.net Neom Projekt - ft.com Microsoft Marktkapitalisierung und Gewinne - cnbc.com Microsoft Earnings - ft.com LinkedIn erreicht erstmals 5 Mrd. Quartalsumsatz - geekwire.com Produktionsende für Tesla Model S und X - cnbc.com SAP Earnings - ft.com Signal Scam - linkedin.com Signal Scam- linkedin.com SEC klärt Regeln für tokenisierte Aktien, verschärft Überwachung. - coindesk.com Zahlungsmethoden und CSAM-Erkennung - theverge.com Gericht blockiert Chatbot-Beschränkungen für Minderjährige - reuters.com Meta startet Werbekampagne für Rechenzentren - theverge.com EU-Untersuchung gegen Shein möglich - reuters.com US-Untersuchung zu WhatsApp-Datenschutz - bloomberg.com Melania Trump-Dokumentarfilm verspottet wegen leerer Kinovorführungen. - gbnews.com Amazons 35-Millionen-Dollar-„Melania“ - nytimes.com Wie viel kann ein Präsident verdienen? - nytimes.com Party donations 2025: AfD large donations skyrocket - lobbycontrol.de Diese Entwicklung sollte jeden Amerikaner beunruhigen. - x.com Trump Cybersicherheit Sensible Materialien ChatGPT - independent.co.uk Marktführer Robitik Robco- handelsblatt.com US-Handelsdefizit wächst am stärksten seit 34 Jahren - reuters.com im Loop Podcast - finanzfluss.de Tim Gabel Podcast mit Pip – Youtube

i6 Group is connecting the fragmented aviation fuel ecosystem-airlines, fuel suppliers, and service providers-through a real-time digital platform that eliminates paper-based processes at over 260 airports worldwide. After launching with British Airways at Heathrow in 2015 and recently closing their Series B with German PE firm Itrium, i6 is proving that even heavily regulated, risk-averse industries can achieve step-function operational improvements through software. In this episode of BUILDERS, Alex Mattos, CEO and Managing Director of i6 Group, breaks down how they navigated decade-long enterprise sales cycles, leveraged strategic customers as Series A investors, and are now building toward profitability to maximize exit optionality. Topics Discussed: The surprising analog nature of aviation fuel operations despite advanced aircraft technology i6's pivot from defense fuel system testing to commercial aviation digitization The multi-party fuel ecosystem: airlines, suppliers, service providers, and logistics chains Strategic approach to landing British Airways and Virgin Atlantic as launch customers Fundamental differences between European fuel optimization vs. US supply chain management models Multi-stakeholder enterprise sales involving fuel teams, flight ops, pilot unions, and CFOs Strategic Series A with customer-investors: British Airways, JetBlue, Shell, and World Fuel Services Series B transition from strategic to PE backing focused on scaling operations and go-to-market Network effects driving compounding value as airport coverage expands Path to self-sustainability and exit strategy considerations GTM Lessons For B2B Founders: Target brand DNA, not just budget, for early enterprise customers: i6 deliberately approached Virgin Atlantic because of Richard Branson's reputation for "being entrepreneurial, taking a risk, doing something different." This wasn't naive brand worship—it was strategic targeting based on organizational risk tolerance. When selling complex infrastructure to enterprises pre-product-market fit, a prospect's innovation track record matters more than their budget size. Map your early pipeline based on cultural willingness to partner with startups, not just technical fit. Invest in non-paying reference customers as currency for tier-one deals: Virgin Atlantic became i6's first operational deployment without payment. This wasn't charity—it was strategic capital allocation. The working reference at Virgin directly unlocked British Airways: "we turned up, demonstrated what we were doing...we've done this trial with Virgin and here's the results, and it went really well." For founders selling to conservative enterprises, one live deployment at a credible brand is worth more than a dozen pitch decks. Budget 6-12 months of runway for strategic pilots that generate proof points, not revenue. Create forcing functions with specific follow-up commitments: When British Airways said "if you're still here in six months, come back," most founders would hear soft rejection. Alex heard a calendar commitment and returned "to the day" with results. This precision signaling—we take your requirements seriously enough to track them to the day—separates serious vendors from tire-kickers. When enterprises set conditional bars, treat them as binding contracts and demonstrate execution discipline through exact follow-through. Position for market disruption by maintaining warm enterprise relationships: i6 benefited when an incumbent competitor liquidated, creating urgent procurement needs at British Airways. But luck favors the prepared—they had already established credibility through their Virgin deployment. Maintain enterprise relationships even when deals seem stalled. In concentrated B2B markets, competitive exits, budget releases, and trigger events happen regularly. Your position in the consideration set when disruption hits determines whether you capture the opportunity. Engineer word-of-mouth in concentrated industries through excellence, not marketing: Four months after Heathrow deployment, Dubai airport approached i6 unsolicited: "we've heard great things." In the aviation fuel community—which Alex describes as "surprisingly small"—exceptional execution travels faster than any outbound motion. This changes GTM strategy: in concentrated industries, over-invest in customer success and operational excellence at early deployments rather than spreading thin across many accounts. Your first customers are your sales team. Segment GTM by operational model, not just geography or company size: i6 discovered European airlines optimize for fuel efficiency and real-time decisions, while US airlines (controlling their own supply networks since the late 1980s) prioritize supply chain visibility: "how much fuel did we put in the plane, how much have we had delivered, how much have we got left." These aren't feature preferences—they're fundamentally different jobs-to-be-done driven by market structure. Don't assume global enterprises have unified needs. Segment by operational model and regulatory environment, then customize messaging and roadmap accordingly. Stage investor expertise to match company evolution, not just valuation milestones: Series A brought strategic investors who were actual users (British Airways, JetBlue, Shell, World Fuel Services) for product validation and network access. Series B brought PE firm Itrium for "scaling the business...building and growing our sales and revenue teams." This wasn't opportunistic—it was deliberate staging of capital sources to match capability gaps. Don't optimize fundraising purely on valuation or dilution. Map your next 18-month bottleneck (product validation vs. operational scaling vs. market expansion) and raise from investors who've solved that specific problem. Build for profitability to control your exit timing and terms: Alex's goal is avoiding Series C entirely: "we build and establish a fully self-sustaining business...the business becomes fully sustainable in the next couple of years." This isn't conservatism—it's strategic optionality. Reaching profitability eliminates the forced march toward subsequent rounds, letting you choose between IPO or M&A based on market conditions rather than cash position. For infrastructure plays with long implementation cycles, factor sustainability into your growth model early, even if it moderates topline growth rates. // Sponsors: Front Lines — We help B2B tech companies launch, manage, and grow podcasts that drive demand, awareness, and thought leadership. www.FrontLines.io The Global Talent Co. — We help tech startups find, vet, hire, pay, and retain amazing marketing talent that costs 50-70% less than the US & Europe. www.GlobalTalent.co // Don't Miss: New Podcast Series — How I Hire Senior GTM leaders share the tactical hiring frameworks they use to build winning revenue teams. Hosted by Andy Mowat, who scaled 4 unicorns from $10M to $100M+ ARR and launched Whispered to help executives find their next role. Subscribe here: https://open.spotify.com/show/53yCHlPfLSMFimtv0riPyM

Read the full article at https://www.blockchaingamer.biz/features/41347/top-50-blockchain-game-companies-2026/[6:55] One of the trends is the rise of games with degen mechanics, notably YGG at #3.[8:35] YGG's positioning of LOL Land as a "Casual Degen" game - a casual game with degen rewards.[10:30] These games have been a (relative) success. Mass-market web3 games haven't been.[12:55] Also in the Top 50, Cambria, Gigaverse and Fishing Frenzy use this degen mechanic.[15:37] Even Mythical Games is moving towards this with its Pulse Arena platform. [19:35] Sky Mavis' Ronin blockchain has a lot of these degen games running on it.[27:02] Dacoco's Alien Worlds is an underappreciated web3 game for its DAO focus. [35:00] CCP's EVE Frontier is the prime example of a bottom-up, fully onchain MMORPG.[41:18] Roblox-with-blockchain Hytopia launched in 2025 with 2.5 million mobile downloads.[44:27] The #1 company is MapleStory Universe dev Nexpace (Nexon).[52:04] Also ... Bitcoin game payment company ZBD announced a $40 million Series C funding round.

Bobby and Alex open by discussing (you guessed it) the Los Angeles Dodgers signing Kyle Tucker to a jaw-dropping four year, $240 million contract. They discuss whether this has any impact whatsoever on the impending lockout and what, if anything, we should read into this about the state of baseball economics. Then, they're joined by Good Friend of the Show Jen Ramos Eisen to discuss, among other things, private equity in minor league baseball, the Joker, Series C funding, and the nascent Women's Professional Baseball League.Links:Follow Jen on BlueskyThe WPBL Has Teams and Players Now⁠⁠⁠⁠⁠⁠⁠Join the Tipping Pitches Patreon ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Tipping Pitches merchandise ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Call the Tipping Pitches voicemail: 785-422-5881Tipping Pitches features original music from Steve Sladkowski of PUP.

Lawmakers in the US have reintroduced a bipartisan bill to help commercial satellite owners and operators defend against growing cybersecurity threats to satellites. Satellite manufacturer K2 Space has raised  $250 million in Series C funding at a $3 billion valuation. Fortastra has emerged from stealth and has raised over $8 million in seed financing, and more. Remember to leave us a 5-star rating and review in your favorite podcast app. Be sure to follow T-Minus on LinkedIn and Instagram. T-Minus Guest Elysia Segal brings us the Space Traffic Report from NASASpaceflight.com. Selected Reading Lawmakers revive satellite cybersecurity bill to shield commercial space systems from rising threats - Industrial Cyber K2 Space Raises $250M at $3B Valuation to Roll Out a New Class of High-Capability Satellites Strength Among the Stars: Fortastra Emerges to Build Maneuverable Spacecraft for National Security New Orders Signed With MBDA For Over 35m€- Avio NASA Selects Two Heliophysics Missions for Continued Development NASA JPL Unveils Rover Operations Center for Moon, Mars Missions Share your feedback. What do you think about T-Minus Space Daily? Please take a few minutes to share your thoughts with us by completing our brief listener survey. Thank you for helping us continue to improve our show.  Want to hear your company in the show? You too can reach the most influential leaders and operators in the industry. Here's our media kit. Contact us at space@n2k.com to request more info. Want to join us for an interview? Please send your pitch to space-editor@n2k.com and include your name, affiliation, and topic proposal. T-Minus is a production of N2K Networks, your source for strategic workforce intelligence. © N2K Networks, Inc. Learn more about your ad choices. Visit megaphone.fm/adchoices

In this episode of Truck Tech, we catch up with John Harris, co-founder and CEO of Harbinger to dive into their recent Series C round, their big order with Fedex, and what's been going on in the medium duty EV truck space. ⁠Follow the Truck Tech Podcast⁠ ⁠Other FreightWaves Shows Learn more about your ad choices. Visit megaphone.fm/adchoices

CrowdStrike fires an insider who allegedly shared screenshots with hackers. Google agrees, it wasn't Salesforce. Cox Enterprises confirms Oracle EBS breach. Alleged Transport for London hackers plead not guilty. Hackers exploit new WSUS bug to deploy ShadowPad backdoor. Iberia discloses breach of customer data. Harvard discloses voice-phishing breach exposing alumni and donor data. We have our Monday Business Briefing. Our guest today is Brandon Karpf, friend of the show discussing maritime GPS jamming and spoofing. And the launderers who wanted a bank for Christmas. Remember to leave us a 5-star rating and review in your favorite podcast app. Miss an episode? Sign-up for our daily intelligence roundup, Daily Briefing, and you'll never miss a beat. And be sure to follow CyberWire Daily on LinkedIn. CyberWire Guest Today we are joined by Brandon Karpf, friend of the show discussing maritime GPS jamming and spoofing. Selected Reading CrowdStrike fires 'suspicious insider' who passed information to hackers (TechCrunch) Google says hackers stole data from 200 companies following Gainsight breach (TechCrunch) Cox Confirms Oracle EBS Hack as Cybercriminals Name 100 Alleged Victims (SecurityWeek) Teens plead not guilty over TfL cyber-attack (BBC) Attackers deliver ShadowPad via newly patched WSUS RCE bug (Security Affairs)  Iberia discloses customer data leak after vendor security breach (Bleeping Computer)  Harvard University discloses data breach affecting alumni, donors (Bleeping Computer)  Doppel secures $70 million in a Series C round. (N2K Pro Business Briefing)  Russia-linked crooks bought a bank for Christmas to launder cyber loot (The Register) Share your feedback. What do you think about CyberWire Daily? Please take a few minutes to share your thoughts with us by completing our brief listener survey. Thank you for helping us continue to improve our show. Want to hear your company in the show? N2K CyberWire helps you reach the industry's most influential leaders and operators, while building visibility, authority, and connectivity across the cybersecurity community. Learn more at sponsor.thecyberwire.com. The CyberWire is a production of N2K Networks, your source for strategic workforce intelligence. © N2K Networks, Inc. Learn more about your ad choices. Visit megaphone.fm/adchoices

Learn how EV maker Harbinger secured significant capital and a key initial fleet order in Harbinger lands $160M Series C, inks initial FedEx deal for 53 electric trucks. The electric vehicle manufacturer raised $160 million in Series C funding, bringing its total to $358 million, and simultaneously received an initial order for 53 Class 5 and Class 6 electric vehicles from FedEx. Harbinger's proprietary electric platform offers competitive acquisition costs and modular batteries, ranging from 140 to over 200 miles, positioning the company to lead the mass adoption of medium-duty electric trucks. Next, we dive into the contentious rail industry merger detailed in Rail merger could raise prices, hurt US ability to compete, say GOP legislators. Dozens of Republican state legislators have warned regulators that the proposed Union Pacific and Norfolk Southern rail mega-merger threatens to raise consumer costs on essential goods and hinder the competitive ability of U.S. companies. Legislators argue that the combined system would control nearly 45% of U.S. rail tonnage across 43 states, creating "captive shippers" and risking widespread service disruptions and supply chain instability. Finally, discover the major strategy shift at the national carrier, covered in US Postal Service makes U-turn on last-mile delivery. New Postmaster General David Steiner announced the U.S. Postal Service must grow revenue by leveraging its unique national network to provide last-mile delivery service for large shippers, reversing the strategy of his predecessor. This reversal has led to a tentative agreement with UPS for its budget Ground Saver service, although critics like parcel industry executives worry that offering last-mile services to competitors could cannibalize existing USPS parcel products. Learn more about your ad choices. Visit megaphone.fm/adchoices

The FCC plans to roll back cybersecurity mandates that followed Salt Typhoon. The alleged cybercriminal MrICQ has been extradited to the U.S. Ransomware negotiators are accused of conducting ransomware attacks. Ernst & Young accidentally exposed a 4-terabyte SQL Server backup. A hacker claims responsibility for last week's University of Pennsylvania breach. The UK chronicles cyberattacks on Britain's drinking water suppliers. Monday business brief. Our guest is Caleb Tolin, host of Rubrik's Data Security Decoded podcast. Hackers massage the truth.  Remember to leave us a 5-star rating and review in your favorite podcast app. Miss an episode? Sign-up for our daily intelligence roundup, Daily Briefing, and you'll never miss a beat. And be sure to follow CyberWire Daily on LinkedIn. CyberWire Guest Today we are joined by Caleb Tolin, host of Rubrik's Data Security Decoded podcast, as he is introducing himself and his show joining the N2K CyberWire network. You can catch new episodes of Data Security Decoded the first and third Tuesdays of each month on your favorite podcast app. Selected Reading FCC plans vote to remove cyber regulations installed after theft of Trump info from telecoms (The Record) Alleged Jabber Zeus Coder ‘MrICQ' in U.S. Custody (Krebs on Security) Chicago firm that resolves ransomware attacks had rogue workers carrying out their own hacks, FBI says (Chicago Sun Times) Ernst & Young cloud misconfiguration leaks 4TB SQL Server backup on Microsoft Azure (Beyond Machines) Penn hacker claims to have stolen 1.2 million donor records in data breach (Bleeping Computer) Hackers are attacking Britain's drinking water suppliers (The Record) JumpCloud acquires Breez. Chainguard secures $280 million in growth financing. Sublime Security closes $150 million Series C round. (N2K Pro) Hackers steal data, extort $350,000 from massage parlor clients (Korea JoongAng Daily) Share your feedback. What do you think about CyberWire Daily? Please take a few minutes to share your thoughts with us by completing our brief listener survey. Thank you for helping us continue to improve our show. Want to hear your company in the show? N2K CyberWire helps you reach the industry's most influential leaders and operators, while building visibility, authority, and connectivity across the cybersecurity community. Learn more at sponsor.thecyberwire.com. The CyberWire is a production of N2K Networks, your source for strategic workforce intelligence. © N2K Networks, Inc. Learn more about your ad choices. Visit megaphone.fm/adchoices

Elizabeth Gore has always believed that all entrepreneurs deserve equitable access to capital. But pursuing that mission landed her in the middle of a federal class action lawsuit. This week on SUPERWOMEN, I'm joined by Elizabeth Gore, co-founder and president of Hello Alice, which has helped 1.6 million small businesses and distributed over $65 million in grants. She talks candidly about being sued for funding Black business owners, surviving the collapse of Silicon Valley Bank, and a health scare that nearly took her out. Through it all, Elizabeth shares the lessons she's learned about resilience, leadership, and staying rooted in purpose—even when the stakes couldn't be higher. Episode Guide: (00:00) Meet Elizabeth Gore, co-founder and president of Hello Alice (06:37) Building an AI platform before it was mainstream (09:29) Getting sued for funding Black entrepreneurs (13:18) How they lost ⅔ of their Series C funding (15:03) The health crisis that stopped her cold (20:23) Learning to truly rest as a founder (28:48) What's next for women and AI in business (31:37) The decision that nearly broke Hello Alice (33:43) Raising millions without a CFO (36:03) What Elizabeth would do if everything fell apart Learn more about your ad choices. Visit megaphone.fm/adchoices