Adam Sah

Jazzed for this keynote by Jeff Erhardt at our 12th annual LDV Capital Vision Summit: The Materials Innovation Gap: Why AI Predictions Aren't Enough — And What It Will Take to Transform a $6 Trillion Industry RSVP: https://lnkd.in/eeUHN_m AI is transforming industry after industry, and generating enormous excitement in the world of materials science. Machine learning models can now predict millions of new material candidates with unprecedented speed and accuracy. At the same time, programmable material classes, including those recognized by the 2025 The Nobel Prize in Chemistry, could unlock breakthroughs in some of the most pressing challenges facing civilization, from environmental remediation to the energy transition. But their impact is constrained by something prediction alone cannot resolve: new materials must be physically produced, scaled, integrated, and qualified within the larger systems into which they are deployed, a journey that remains extraordinarily long, expensive, and unpredictable. Jeff will argue that closing this materials innovation gap requires treating AI and physical experimentation as equal partners, and building a new model of collaborative development that connects materials innovators and industrial partners far earlier in the discovery process.

12

Sustained ~252–291 GiB/s of HBM memory throughput on H100s under decode load — essentially hitting the hardware roofline. This matters because HBM throughput, not FLOPs, is the real bottleneck in LLM inference. By keeping memory nearly fully saturated, I’ve unlocked far higher efficiency and throughput than standard engines. The result: 0.9–1.36M tokens/sec with ~0.1 ms first-token latency  #AI #LLM #GPURouter #H100 #Inference #CostEfficiency #Innovation #DeepLearning #AIInfrastructure #HighPerformanceComputing NVIDIA AMD OpenAI Google Shilpa Kolhatkar Keith Strier a16z speedrun

5

New York City-based Betaworks has closed its third fund at $66M to invest in early-stage AI startups. Betaworks, a staple of the NYC tech ecosystem since 2008, will focus its new fund on investing in pre-seed and seed-stage AI companies. The firm has been investing in AI since 2016 and has backed companies like Hugging Face. Fund III is larger than the firm's previous funds ($46M Fund II in 2020 and $48M Fund I in 2016) and saw many of its previous limited partners return to invest, a strong signal in a challenging fundraising environment for VCs. The fund will seek to make at least 25 pre-seed to seed investments, as well as 50 investments into startups as part of the Betaworks Camps program. Congratulations to the Betaworks team. TechCrunch: https://lnkd.in/d--hfMpa #VentureCapital #AI #EarlyStage #SeedFunding #NYCTech #Fundraising

2

1 Comment

I’ve been working on the theory of why small, shippable factories are the perfect inversion to the centralized prefab factory model. It all collapsed into one line of math (credit to Gilles Retsin): Factory Efficiency = Usage × (Throughput × Product Value ) / ( CapEx + Deployment OpEx) The idea is that the closer you can drive Usage to 100 % while keeping the denominator tiny, the faster the flywheel spins.  A shippable microfactory tackles this by primarily focusing on a high Usage factor. Because the microfactory can be packed up afterwards and reused, its CapEx gets amortized over multiple projects, effectively raising its overall Usage across the year. If one project alone doesn’t fully occupy the factory, it can simply roll over to another job to maintain high Usage. Its CapEx is often an order of magnitude lower than a centralized factory (< $1M vs. $10-50M min), so the breakeven throughput is more achievable on a small pipeline of work. The real breakthrough here is portability. There is near-zero stranded capital and almost no idle time. To flesh this theory out, I wrote 3000 words on the topic for Brad Hargreaves and Thesis Driven. The article breaks down: - why “fixed factories, shipped goods” is being inverted to shipped factories, fixed goods - what a sub‑$1 M robotic cell does to CapEx per home versus a £45 M off‑site plant - early production/cost data on a shippable microfactory from Mollie Claypool, Gilles Retsin, Sam Baker and the Automated Architecture (AUAR) team - the hybrid reality of robots tackling the heavy structural components and human crews assembling faster and cheaper - technical constraints to pull this model off and future considerations on the optimal business model (own vs. rent) Major shoutout to Gilles Retsin for helping me pull this together. https://lnkd.in/eRTnbZPa

54

2 Comments

Lessons from YC: Culture Isn’t a Slide Deck YC didn’t just change how we built products. It changed how we worked together. Before YC, “culture” was something we’d figure out later. We thought it meant perks, team dinners, or a list of values on the wall. But inside YC, culture meant one thing: how fast your team can move without breaking. That’s when it clicked. Culture isn’t what you say. It’s what you tolerate. If someone cuts corners and you let it slide, that’s your culture. If someone owns their mistake and you celebrate it, that’s your culture. We started treating culture like product — build, test, iterate. That's when we realized small habits create the real DNA: how you communicate under pressure, how you make decisions, how you handle bad days. YC pushed us to grow up as a team, not just as founders. Because a product might attract customers but culture is what keeps them.

30

2 Comments

Voice is a super interesting modality right now - maybe the first modality we're seeing move to open source models across a number of scale ups / enterprises. Reliability concerns, high costs, and open source model performance are pushing engineers to do their own fine tuning vs. relying on third-party vendors of proprietary models. Many of these orgs have already been collecting their own first-party data and now with third-party vendors like Extrian, David AI, etc they can train really high quality models. RL has been insanely hyped, but it's been unclear how long it will take scale ups and enterprises to actually lean in. Voice AI might be hitting that inflection point faster than expected.

39

3 Comments

GPUs have gotten faster, and models have gotten larger. But storage is still stuck in the past. Modern AI workloads need instant access to huge datasets. Instead, teams waste time on slow retrieval, idle GPUs, and costly pipelines. Archil is changing this. It delivers object-storage scalability at block-storage speeds. The platform allows for seamless, high-performance access to massive datasets. By eliminating cold starts and accelerating throughput 30X, Archil enables teams to fully utilize their infrastructure and move faster on training, analytics, and AI deployment. Archil’s founder Hunter Leath brings rare technical depth and operating experience from building Amazon EFS and optimizing Netflix’s cloud performance. This is a team building the foundational layer that AI workloads desperately need. I’m thrilled to lead Archil’s seed at Felicis with Nancy Wang and support their mission to reinvent cloud storage for the AI era. https://lnkd.in/gWueXhzz

65

2 Comments

Investors (esp deep tech investors) generally think about things/bet on "layers": the application, the infrastructure, the model. I think that's becoming the wrong way to look at it? The real opportunity is in the feedback loop *between* the layers. A new social app creates demand for new hardware, which in turn demands a new kind of intelligence - I'm calling it "The Agency Loop". My latest newsletter explores this thesis through three signals: the "weaponized transgression" of Cluey, the AI-driven materials science in Nature, and the architectural critiques of Kenneth Stanley. Bet the loop, not the layer :) You can read the full analysis here: https://lnkd.in/ggdVGEda #vibecap #vibecapital #vc #deeptech #AI #cluely

4

1 Comment

Systems first. AI next. The order matters, a lot. Over the past few weeks, Evan Allen and I have been digging into agentic workflows - reading papers, testing tools, and mapping patterns across different implementations. One insight keeps standing out. Agentic workflows aren’t inherently “AI.” They’re decomposition systems. The core idea is simple: take a complex task, break it into smaller, well-defined steps, execute those steps independently, then reassemble the output.  Who performs each step is secondary. 👉 Some steps are best handled by machines: speed, recall, pattern matching. 👉 Others still benefit from humans: judgment, accountability. 💡 Once you see it this way, the false binary between “automated” and "manual” disappears. For human-in-the-loop systems - like what we're building with Nexus - it shifts the goal from replacing people to assigning work to whoever (or whatever) does it best, step by step. AI becomes a delegate, not an endpoint. That also forces more discipline. You have to be explicit about what the task actually is, what good looks like, and where judgment truly matters. Sloppy workflows don’t magically improve when you add AI - they just fail faster. I suspect the most durable systems won’t be the most automated ones, but the most intentional about where humans stay in the loop. Where have you seen the balance work well, or break down?

14

2 Comments

A Founders Guide to The Long Build™ - Physical AI Thesis Part III This one is for the builders. In this final post, my partner Will Porteous and I along with insights from builders in the trenches and investors who play the long game share our guide for founders building intelligence in the real world. A few core takeaways: • Sim before steel: Runway is precious, model the system behavior • Capital intensity is not the enemy — misaligned capital is • Build complete solutions before you talk platform • Hardware opens the door, software wins the room • Trust is earned through resilience, not hype

10

1 Comment

🚨 LLM Benchmarks: Proceed with Caution! 🚨 In the race to crown the "best" large language models, we're witnessing a troubling trend: many of the latest LLMs are being trained directly on popular benchmark datasets like GLUE, SuperGLUE, or MMLU. This isn't just optimization—it's essentially fine-tuning models to ace the test they've already seen the answers to. The result? Inflated scores that don't translate to real-world performance. We're overfitting to benchmarks, not advancing true intelligence. It's reminiscent of the Dieselgate scandal, where cars were engineered to pass emissions tests but failed in everyday use. Are we heading toward a "Token Gate" in AI, where models game the system for hype rather than genuine capability? As professionals in AI, tech, and data science, we need to demand better: dynamic, unseen benchmarks; adversarial testing; and transparency in training data. #AI #LLMBenchmarks #MachineLearning #EthicsInAI #TokenGate

2

2 Comments

CRV Security: Request for Startups I never know if this actually works for our friends over at YC but figured we'd try. Here's what we want to fund in 2026! 1. Golden Artifacts: Think Chainguard but more broad. Artifact attestation exists for open source. Almost nothing exists for internal software — especially the vibe-coded tooling now running in production. We want the company building cryptographic proof of secure software delivered from secure artifacts: who built it, how, and whether it was reviewed. If more things are being yeeted into the world via Claude Code (myself included), this feels like an issue. 2. MCP & Agentic Security: Agents are getting real credentials and taking real actions. The security posture of most orgs around this is basically zero. That changes fast. You'd never give an employee hardcoded API keys or write access to your email without supervision/trust. Why give it to agents? 3. AI Governance: Boards are asking CISOs to account for AI risk. CISOs have no good answer other than "Palo has a module" 4. Next-Gen Endpoint: CrowdStrike was built for a world of static binaries and human operators. AI workloads, cloud-native infra, and AI-assisted attackers need a new architecture. The category is ready to be reinvented. 5. Networking in the AI Era: Zero trust was designed for humans. What does network security look like when the entity requesting access is an agent? Nobody's really solved this. 6. Email Security + Next-Gen Phishing: LLMs have made spear phishing infinitely scalable. I've never truly understood why Abnormal and KnowBe4 aren't one company. Maybe this time it's different. 7. Frontier Security Lab: We'd back a credible, well-staffed lab focused entirely on red-teaming models and setting the evidentiary standard the industry needs as LLM built apps become the norm. 8. Dependency Security: That Actually Remediates Malicious and vulnerable dependencies are a top attack vector. The tooling is mostly noise — scanners that don't close the loop. The winner here ships fixes, not just alerts. 9. Critical Infrastructure Cyber: Data centers, satellites, power grids, undersea cables. The physical backbone of the internet is increasingly exposed and wildly under-defended. We have data centers in space, for God's sake. Surely we need better cyber for critical infrastructure? 10. PAM for the Modern Era Legacy: PAM was built for static roles, human users, on-prem directories. Cyberark was founded in 1999.....Agents, ephemeral workloads, and cloud-native infra have broken all of those assumptions. Is anyone rebuilding this from scratch? If you're building in any of these areas — or something we haven't thought of — reach out. james@crv.com

445

48 Comments

DealWire.TECH Deal of the Week - Feb. 20, 2026. Waymo raises $16B to scale robotaxi infrastructure Deal: $16B late-stage round, Valuation: ~$126B Lead investors: Dragoneer, DST Global, Sequoia, Alphabet Waymo, the autonomous driving unit of Alphabet, has raised a massive $16 billion late-stage round, valuing the company at approximately $126 billion. The financing brings together leading global venture and crossover investors, with Alphabet continuing as the strategic anchor. The raise comes as Waymo transitions from years of R&D and pilots into full-scale commercial deployment. The company already operates fully driverless ride-hailing services in multiple U.S. cities and is targeting rapid geographic expansion, with ambitions to reach 1 million weekly rides by the end of 2026. The new capital will accelerate fleet deployment, expand into new urban markets, and build the infrastructure required to support large-scale autonomous mobility networks. Why this matters: • Autonomous driving becomes infrastructure Waymo’s $16B round reframes robotaxi networks as long-term infrastructure assets rather than experimental tech. • Part of the trillion-dollar AI infrastructure cycle Unlike foundation model raises, this capital is flowing into AI deployed in the physical world—vehicles, sensors, and city-scale operations. • Winner-take-most dynamics emerging Only a few companies globally can fund autonomy at scale, making capital access a key determinant of long-term leadership. • Capital intensity becomes the moat Success will depend not just on algorithms but on financing and operating large fleets and urban logistics networks. • Structural advantage for Waymo More than a decade of development, billions of miles of data, and deep integration with Alphabet’s AI and compute stack give Waymo a meaningful edge as competitors like Tesla and Cruise continue to push forward. • Start of the AI transport wave Robotaxis are likely just the first step toward broader AI-driven autonomy across trucking, logistics, maritime, air cargo, and drone fleets. DealWire Takeaway: Waymo’s $16B financing marks a turning point: autonomous driving is shifting from frontier technology to city-scale AI infrastructure. As AI moves into the physical economy, the winners will be those who combine advanced algorithms with capital, data, and operational scale. Waymo’s latest round positions it among the frontrunners in the emerging autonomous transportation stack. #DealWire #Waymo #AutonomousVehicles #AIInfrastructure #Robotaxi #VentureCapital #FutureOfMobility #ArtificialIntelligence

6

YellowDog.ai just set a 10x benchmark uplift in scale computing, delivering 40,000 tasks per second (TPS) and managing 100,000 compute nodes in the cloud. What's even more interesting, that's 2x IBM Symphony and opens an intriguing pathway for these until-know closed/captive systems.

10

I spoke to Daphné Leprince-Ringuet from Sifted about the work Synthesia is doing on security and trust. You can find her article linked in the comments. I made two arguments that often get lost when people talk about these issues. First, that application layer companies approach this problem space differently than a research lab or a foundational model developer. Our focus on shipping products rather than training models means we’ve built our security systems, processes and policies in a way that’s different from what model developers do. As a result, we spend our time thinking about the practical misuse of AI video in the context of employee training, sales enablement or customer support, rather than fundamental risks such as preventing the kind of CBRN disasters that keep model developers awake at night. Secondly, Synthesia sells enterprise products and services so our approach to AI security is rooted in slightly different principles compared to companies developing consumer AI solutions. For us, building security follows well-established governance principles described by global standards such as ISO 42001, ISO 27001 or ISO 27701, rather than the rules set by a consumer app store. Despite these differences, I still believe values matter, whether you're building an AI app or training an AI model, and whether you're targeting businesses or consumers. And when it comes to values, Europe has an advantage and a reputation when it comes to building trust.

35

4 Comments

Just downloaded Tesla Robotaxi in SF, a real threat to Uber, Lyft, and Grab? Most experts measure Robotaxi by today’s metrics: cost of operations, active users, global adaptability. By that logic, scaling looks impossible. But Tesla flips the script. Unlike Waymo’s costly retrofits, a Tesla only needs a software update to become a Robotaxi. With 1.7M+ Hardware 4.0 vehicles already on the road, the scale potential is unmatched. Cars shift from being liabilities to income-generating assets overnight. This isn’t just product innovation, it’s a redefinition of value. At ElasticDash we’re doing the same for QA. Today, QA costs are labor-driven, with humans building and maintaining tests. Our AI agents change that: they build and adapt end-to-end tests automatically as your app evolves. Just as Tesla is rewriting mobility economics, we’re rewriting QA economics. ElasticDash is on the way.

37

21 Comments

Just watched Daniel Han, founder of unsloth.ai, in the “LLMs for Everyone” conversation—and came away inspired. Onslaught is reshaping how we fine-tune large language models, driving progress through: Speed – cutting fine-tuning time dramatically Cost-efficiency – reducing compute needs and lowering barriers to adoption Scalability – making customization feasible across diverse teams and domains This approach unlocks the ability for organizations to deploy domain-specific LLMs with precision—without the usual trade-offs of high cost or long lead times. From my perspective in remote sensing and GIS, the implications are huge: tailored LLMs could accelerate spatial data interpretation, streamline automated feature extraction, and strengthen geospatial decision-support systems. https://lnkd.in/dBCSkpfi

36