What I like about $NBIS is that the opportunity goes beyond selling GPU capacity because Aether Soperator and Token Factory help customers get more useful work from the same infrastructure. Eigen AI and Clarifai extend that into inference where better software can lower compute needs and increase what Nebius earns from same installed infrastructure.

We just wrapped Draper Summit 2026 at Half Moon Bay. Every year we host this reunion of people who are changing the world. We heard from some of the top leaders today including: Baiju Bhatt, Founder of Robinhood → Aetherflux (this man does not slow down) Rosie Rios, former Treasurer of the United States. Rob Kaplan, Vice Chairman of Goldman Sachs. Sal Khan, who is revolutionizing how every human on Earth learns. The conversations were so futuristic, you would have to hear it from our entrepreneurs to believe it. HexemBio is rejuvenating all your blood cells. There were stealth companies regenerating limbs via bioelectricity, replacing brain cells, and rethinking human reproduction entirely via artificial wombs. Muse Bio launched the world's first topical rejuvenative serum made from menstrual blood stem cells - promoting scarless healing and glowing skin. Autolane is building the orchestration layer the autonomous world needs to function. Feltsense is building founder agents that source ideas, build product, and go to market without a human in the loop (they replicated the entire YC batch a few weeks ago). Venus Aerospace is developing the most efficient rocket engine ever built. Gemma Robotics sent a robot to autonomously apply professional make up - airbrushed foundation, contour, etc. The results were fantastic. And we live taped Meet the Drapers Silicon Valley! You'll have to tune in to see the winner… Another summit in the books, for the books. The greatest takeaway about what made Draper summit successful: We all were able to spend time in a beautiful location, learning about frontier tech, getting to know each other deeply, and creating real, meaningful friendships while simply having a lot of fun. At Draper, we know that VC runs on relationships that compound over decades. We are here to build trust with you as you build generational companies. Thank you to all our LPs, founders, ecosystem members, co-investors, and of course the relentless Draper team for coming and making this all possible. Let me know in the comments if you should come next year. Hope to see you there.

Wanna get out of here? You have to beat me in a Pokemon battle first! 💪 𝐀𝐞𝐭𝐡𝐞𝐫 𝐅𝐨𝐮𝐧𝐝𝐚𝐭𝐢𝐨𝐧 𝐄𝐦𝐩𝐥𝐨𝐲𝐞𝐞 𝐏𝐨𝐤𝐞𝐦𝐨𝐧 will be 𝐅𝐑𝐄𝐄 for all my beloved 🅿️atrons! Check my cmt for more info!

As uncle @JensenHuang would say, that loser premise makes no sense to me. Capability doesn't advance only as a smooth race for the biggest general model or the most FLOPs. The frontier is jagged, and the cleanest illustration is the strawberry problem. Ask a model that can hold its own with a physics PhD how many R's are in the word "strawberry" and for a long stretch it would answer two, confidently, every time. The same system could work through a graduate-level proof and then insist 9.11 is a bigger number than 9.9. Each specific embarrassment gets patched as it goes viral. The shape underneath survives every patch, that is - enormous strength in some places, sudden collapse in others, and no way to tell which ground you're standing on until you test it. But counting letters and numbers are funny failures. The expensive version of the same problem shows up in security. Hugging Face was mid-incident, working an agentic attack, and needed a model to review logs. The commercial frontier models refused due to "safety" reasons. From inside a guardrail, a defender reading attack traffic and an attacker reading attack traffic look identical, so the safest available behaviour is to decline. The team switched to an open-weights model on their own infrastructure and got the analysis done. The most capable system in the room was the one that couldn't help, at the exact hour it was needed. That refusal is one of many permanent structural feature of general-purpose models, and it sits right on top of a business. Aether AI operates in the gap between those two failures, and @tryaether_ai AI is Australian. So telling Australian founders the game is over at the model layer, dig up some lithium and prepare a robot tax, is exactly the wrong message. The scoreboard was never general-model scale. It's about who owns the jagged edges that customers, regulators, and national security actually pay for. Australia can own plenty of them. Some of us already do.