One week until Nova Future Summit! September 28 to 30 in Napa. 500 tech and telecom leaders. We're co-hosting with @GSMA. Joining us: @btaylor, @zatlyn, @ssamat, @Alfred_Lin, @clarashih, @gaurav_garg_ and more. novasummit.com/ #NovaFutureSummit26

Everyone's arguing about how long a GPU lasts. @Wing_VC Partner @chriszeoli dug into the fleet data and says that's the wrong question. 🖥️ The chips are fine. Google just retired one after nine years and it was still running 📉 It's the rent that falls. An H100 went for $8 an hour in 2023 and about $2.75 now 💵 That means the money comes in fast, with half of what a GPU will ever earn showing up in its first two years 📊 The accounting doesn't see it that way. It spreads the cost evenly over six years, so right now margins look better than the cash says The chips aren't dying early. The money just shows up early, and that catches up the day spending slows. Do you think the market is pricing this in? Full breakdown 👇 wing.vc/content/how-long-doe…

Tahoe-100M became a seminal work. And we are proud that many are building on top. Read the full paper in today's special edition of Cell.

New issue of @CellCellPress out today on AI and biology, featuring several papers from @arcinstitute: State cell perturbation model, scBaseCount AI agent curated single-cell dataset, Virtual Cell Challenge 2026 commentary, and Tahoe-100M from our friends at Tahoe Bio and featured in Arc Virtual Cell Atlas. Congrats to the teams. Lots more exciting AI/bio work cooking at Arc Institute! cell.com/cell/current

Four papers from the Wing community landed in the same issue of @CellPressNews this week. The issue is focused on AI in biology, and it features: - Tahoe-100M from @tahoe_ai's @nalidoust, @genophoria, @iamjohnnyyu, and others - State from @genophoria and others - scBaseCount from @genophoria and others - New synthetic biology research on CAR T cells from @Satpathology's lab at Stanford We can't think of another venture portfolio that has shown up this way in a single issue of Cell, and we're really proud of these teams. Congrats to everyone behind this work! Read the full issue here: cell.com/cell/issue?pii=S009…

Ten companies can each stop AI cold. @Wing_VC Partner @chriszeoli went through their earnings calls and ran one test: if this company stopped shipping tomorrow, could hyperscalers replace it within 24 months? For all ten, the answer was no. Either they own something nobody else can make: 🔬 ASML. The only EUV machine on earth ⚡ NVIDIA. Still no match for NVLink or CUDA, so it keeps 80 cents of every accelerator dollar 🏭 TSMC. Packaging so tight it caps its own customers 🧩 Broadcom. Switching custom chip vendors is an 18 month redesign Or the substitute exists and is just as sold out: 🧠 SK hynix. Half the HBM market, and Micron is sold out too 💾 SanDisk and Kioxia. 27% of the world's NAND, on allocation past 2027 🔌 Eaton and Forgent. Switchgear runs two years, transformers longer 🏗️ Blackstone. 15 gigawatts of powered land, and every rival is in the same queue Full breakdown here: wing.vc/content/the-10-compa…

Kick off your week with the new one from @Wing_VC Partner @chriszeoli. Chris digs into CoreWeave to answer a question most people building in AI never think about: who is paying for all this compute, and on what terms. The answer is debt. Chris traces all eight CoreWeave facilities and finds lenders are not underwriting the hardware. They are underwriting the customer contract attached to it. Land a hyperscaler and you borrow at SOFR+225. Don't and you're quoted +800. So the cost of compute depends on your customer list, which means the financing layer is quietly deciding who gets to compete in AI infrastructure at all. Nobody can say what a three year old GPU sells for. Nvidia itself will guarantee the building and the power but not its own chips. That gets tested next year when the first contracts come up for renewal. Do you think customers will sign again? Read it: wing.vc/content/how-the-gpu-…

Thrilled to be an investor!

The @wafer_ai team has moved incredibly fast to build dramatically faster, more efficient AI inference. Emilio is a one-of-a-kind talent, and we’re thrilled to be working with him and the team. Proud to be part of the journey as Wafer announces its $40M Series A!

We’re excited to announce that we’ve raised a $40M Series A! co-led by @MarathonMP and @chemistry, with participation from @Wing_VC, @AMD Ventures, @outsetcap, @fiftyyears, and @ycombinator, and our existing investors doubling down on @wafer_ai. we are also joined by an incredible list of angels, including @JeffDean (CEO, @DiscoLoopAI), @rauchg (CEO, @vercel), @andyfang (CTO, @DoorDash), @kvogt (CEO, Bot), @akothari (COO, @NotionHQ), @eastdakota (CEO, @Cloudflare), @deepgramscott (CEO, @DeepgramAI), and more! Most inference optimization today is manual, service-heavy, and done one-time before deployment. Wafer’s vision is AI that optimizes AI. Wafer continuously learns from your workload’s traffic patterns and performance constraints, then finds the optimal deployment across model, engine, kernels, and hardware. This capital helps us accelerate automating more of the inference optimization loop, giving every deployment the leverage of an expert inference-performance team that continuously finds ways to improve performance per dollar. Thank you to the customers who trusted us with their workloads, the partners who built alongside us, the investors who believed in our mission, and the Wafer team members who have worked tirelessly to turn this vision into reality. 🫶🧇

📊 New from Wing VC Partner @chriszeoli: The AI Memory Stack Everyone thinks the AI race is about chips, but it is really about memory. The chips got fast and the memory that feeds them never kept up, so the expensive part of an AI computer is no longer the brain, it is the thing handing the brain information. Three things worth knowing: 🔻 The gap is enormous. Over 20 years, server compute grew about 3x every two years while memory bandwidth grew only 1.6x. That is roughly a 600x gap, and it now decides how AI infrastructure gets built. 💰 The money has moved. The fast memory sitting next to the chip is about half the manufacturing cost of an NVIDIA B200, and packaging gets you to two thirds. The shortage cascaded down too, pushing ordinary computer memory up 93 to 98% in a single quarter. 🪜 There is no one memory stock. Memory, packaging, connectivity, and the software deciding what data lives where all have different winners, and they are not the same companies. 👉 Read it: wing.vc/content/the-ai-memor…

Today, @Wing_VC Founding Partner @gaurav_garg_ joined @Cohesity's sales kickoff for a board fireside chat. Gaurav first invested in the company when it was only three people. The early bet was simple: rethink the enterprise data stack and build for massive scale. Cohesity now helps organizations secure, manage, and recover some of their most critical data. As enterprises put AI into production, that foundation will matter even more. The first wave of consumer AI drew heavily on public data. Enterprise AI will depend on the proprietary data, workflows, and policies that make each business distinct. Companies will need to know that data is protected, recoverable, and trustworthy before they can put it to work with confidence. Proud to have backed Cohesity from the beginning and excited for what comes next.

Congrats to @ronusedh and the @intology team on their @tbpn interview!

Thrilled to back this incredible team!

We're thrilled to share that @OuterBio has launched! The hardest problem in AI for biology isn't the models, it's the data. Human skin dies in the lab within a week, long before the biology that matters begins. Outer Bio's platform, Yuna, keeps living human skin alive and measurable for four weeks or more, and uses that data to train an AI engine that discovers new skincare bioactives. Read more about our investment from @Wing_VC Partner, @SaraChoi_ wing.vc/content/reengineerin…

Meet @outerbio and YUNA. We are building the future of skin health. outerbio.com Welcome from Michael, Me and the entire team at Outer Bio.

Skin is the largest organ in the human body, and a major site of immune surveillance, aging, and disease. Developing scalable and data-driven methods for drug screening in complex human skin tissues will be critical for drug discovery in inflammation, tissue repair, cancer response, and many other aspects of disease-relevant skin biology. Wonderful to work with Michael, @ladygaga, and the entire @OuterBio team! @Wing_VC

Michael Polansky — better known publicly as Lady Gaga's partner and a former top deputy to Sean Parker — has quietly spent years building an AI-driven startup that keeps living human skin tissue alive for weeks outside the body to discover new skincare spr.ly/6019B1mqHB

Structured data is the largest data modality on earth, and yet no one has built the foundation model for it. Each of those numbers is a decision waiting to happen, from a machine heading for failure to inventory about to run dry. That’s why we backed @Synthefy and led the company’s $6.5M seed round. Language had its moment. Structured data is next. Congrats to @somi_synthefy and the team. I wrote more about why we invested and the opportunity ahead. wing.vc/content/investing-in…

I’m super excited to announce Synthefy’s $6.5M seed round, led by Wing Venture Capital and joined by Haystack, Samsung Next, Canonical, Lightscape, and angels and operators from OpenAI, Microsoft, and Meta. Setting the price of every Uber ride in real time. Spotting the early signs of disease from thousands of patient measurements. Blocking a fraudulent payment before money moves. Keeping the power grid stable as supply and demand change every second. These decisions are not waiting for a better chatbot. They require Machine Intelligence, systems that can learn from massive volumes of structured data and make predictions beyond human scale. At Synthefy, our vision is to advance Machine Intelligence, enabling enterprises and users to make predictions on their troves of structured data that enable them to make critical business decisions. Nori is our first foundation model in a class of AI that will transform prediction. It already unlocks brand new ways of looking at prediction. But this is only the beginning of our journey towards zero-shot prediction that outperforms any bespoke ML. Our mission is to give customers predictive answers from their structured data that they cannot get from today’s tooling, 100× faster and at one-tenth the cost of bespoke ML. We aim to make Nori the universal prediction engine that enterprise software and agents call whenever they need to forecast, score risk, or make a decision from structured data. I’m deeply grateful to my co-founders, the Synthefy team, our early users, and investors. Special thanks to @ai for backing us from day -10, when I was still learning what PMF meant, and to Gaurav (Wing) for seeing the ambition and helping us take it further.