# Nvidia takes a stake in Safe Superintelligence and gets access to Ilya Sutskever's research in return > Source: > Published: 2026-07-27 13:53:18+00:00 *Nvidia's new Safe Superintelligence deal is not just another chip sale. It gives Jensen Huang's company an equity stake, a Vera Rubin customer, and a rare look inside Ilya Sutskever's research lab.* Start with the news. Nvidia and Safe Superintelligence said on July 27, 2026 that they had signed a long-term partnership that gives SSI access to Nvidia's next-generation Vera Rubin systems and expands the lab's compute by an order of magnitude. Nvidia also made an investment in the company. The terms were not disclosed. That is the loop. Nvidia invests in a frontier AI lab. The lab gets more money and more access to Nvidia hardware. The work done on that hardware then gives Nvidia insight into what future chips need to handle. Bloomberg has been tracking these circular AI arrangements since January 2026, and its point is blunt enough: chipmakers, cloud companies and AI labs are tying themselves together with equity, compute contracts and supply deals that make the whole market harder to read from the outside. This SSI deal fits that pattern, but it isn't only a financing story. Nvidia's own announcement says the companies will collaborate on current and future compute platforms by using SSI's insights into the future of AI. The Wall Street Journal also reported that Nvidia moved after gaining unusual access to SSI's closely guarded research. That is rare. Most AI labs want the chips and keep the research wall high. SSI is letting the world's most important AI hardware company closer to the work. ## The Research Access Is The Deal SSI was founded in June 2024 by Sutskever, Daniel Gross and Daniel Levy. Sutskever's credibility is not ornamental here. He co-authored the 2012 AlexNet paper with Alex Krizhevsky and Geoffrey Hinton, worked at Google Brain, contributed to AlphaGo and later co-founded OpenAI, where he helped lead the research culture behind the GPT line. When someone with that record says he's building a lab with one product, safe superintelligence, investors listen. The lab still has no public product, no revenue and no normal roadmap. According to reporting from the Financial Times, cited by The Information, SSI raised $2 billion at a $32 billion valuation in April 2025, with Greenoaks leading and Andreessen Horowitz and Lightspeed among the participants. TechCrunch reported earlier that SSI's first $1 billion round in 2024 included Andreessen Horowitz, Sequoia, DST Global and SV Angel. You don't need to pretend that is conventional startup math. It isn't. Frankly, that is the bet. Investors are not paying for current sales. They are paying for Sutskever, a small technical team, and the possibility that a lab without product pressure can find something OpenAI, Anthropic, xAI and Google DeepMind miss while serving customers, launching features and raising ever larger rounds. Vera Rubin makes the arrangement more concrete. Nvidia unveiled Vera, the CPU that powers Vera Rubin systems, on May 31, 2026, saying the chip was built for agentic AI workloads and runs up to 1.8 times faster than x86 processors on certain workloads. A Vera Rubin allocation is not a line item you can casually replace next quarter. If SSI is getting enough of it to expand compute by an order of magnitude, Nvidia is giving the lab something scarcer than capital. The point is access. Money alone doesn't solve allocation when every frontier lab wants the same racks, the same networking and the same power contracts. Nvidia now has a customer whose research can influence future hardware design, and SSI has a supplier with every reason to keep it near the front of the queue. ## Google Is Still In The Room SSI is not giving Nvidia the whole stack. TechCrunch reported in April 2025 that SSI was using Google Cloud TPUs for its research, with Google Cloud saying the chips would help the startup work toward safe superintelligent AI. That makes the new Nvidia deal more interesting, not less. SSI is now tied to Google's in-house accelerator ecosystem and Nvidia's Vera Rubin platform at the same time. That is uncomfortable for everyone else. OpenAI, Anthropic and xAI can spend heavily, and they do. TechCrunch reported in March 2026 that Jensen Huang said Nvidia's investments in OpenAI and Anthropic were likely to be its last in those companies because both were moving toward IPOs. SSI gives Nvidia a different opening: a private lab, no product cycle, a famous founder and a research relationship written into the public announcement. The $32 billion valuation was already hard to defend by ordinary measures. A research lab with no product and no benchmark trail sits outside the way you usually price a startup. But this market has stopped pricing frontier AI companies on ordinary measures. The real question is whether SSI's quiet research direction produces a capability jump or simply joins the long list of expensive AI promises built on scarce compute and patience. Nvidia declined to disclose the size of its stake. The Journal reported that SSI did not respond to requests for comment. For now, the only hard facts are enough: Nvidia has put money into Sutskever's lab, Vera Rubin capacity is part of the arrangement, and the research relationship runs both ways. That is still unresolved. It is also exactly why the deal matters. **Also read:** [Brian Armstrong tells crypto founders who pivoted to AI they got it backwards](https://startupfortune.com/brian-armstrong-tells-crypto-founders-who-pivoted-to-ai-they-got-it-backwards/) • [JPMorgan eliminates its standalone AI chief role and the move says everything about where enterprise AI is headed](https://startupfortune.com/jpmorgan-eliminates-its-standalone-ai-chief-role-and-the-move-says-everything-about-where-enterprise-ai-is-headed/) • [Nvidia's $750 billion AI financing web is drawing scrutiny over circular money flows](https://startupfortune.com/nvidias-750-billion-ai-financing-web-is-drawing-scrutiny-over-circular-money-flows/)