Nvidia invests $5 billion in OpenAI founder Ilya Sutskever’s secret AI lab

Nvidia is investing about $5 billion in Safe Superintelligence (SSI), the AI venture founded by OpenAI co-founder Ilya Sutskever. The cash comes as part of a multi-year strategic partnership that also gives SSI access to Nvidia’s latest Vera Rubin AI systems. While neither Nvidia nor SSI disclosed the figure, sources told Reuters and the Financial Times it’s roughly $5 billion.

The deal stands out because Safe Superintelligence hasn’t launched a commercial product and shares almost nothing publicly about its technology. Even so, it now ranks among the best-funded AI startups. The move underscores Nvidia’s shift from pure chipmaker to strategic backer of the global AI industry.

Officially, Nvidia and SSI call it a “long-term strategic collaboration.” SSI will tap the new Vera Rubin platforms, which are expected to boost the company’s compute capacity roughly tenfold within about a year. The two will also co-develop future AI computing systems.

For Nvidia, this is about more than selling GPUs. By putting in its own capital, the company gets a front-row seat to the development of a potential next generation of AI models.

It’s a playbook Nvidia has leaned into in recent years. Where it once focused almost solely on hardware, it now builds stakes across AI startups that, in turn, run on Nvidia infrastructure.

Ilya Sutskever is one of the most influential AI researchers of his generation. He co-founded OpenAI, served for years as its Chief Scientist, and played a key role in developing GPT models and ChatGPT.

After leaving OpenAI in 2024, he launched Safe Superintelligence. The company has only a few dozen employees and operates largely in stealth.

SSI says it has pursued a new research direction over the past two years aimed at building powerful AI systems that remain safe and controllable. The company has not disclosed what technical breakthroughs, if any, it has achieved.

That secrecy is precisely what makes this investment striking.

Typically, investors want a working product or rising revenue before writing multi-billion-dollar checks. SSI is the exception.

It has no public AI model, no consumer product, and has published little technical documentation. Yet investors now value it in the tens of billions of dollars.

That confidence rests almost entirely on the team behind the company and the belief that foundational AI research could prove far more valuable than today’s generative AI apps.

Nvidia’s move shows some players are willing to commit vast sums long before commercial use cases are visible.

The investment fits a much broader trend.

Nvidia is minting record profits from AI processors and increasingly plows that cash back into funding AI companies. The result: an ecosystem where Nvidia is supplier, investor, and technology partner all at once.

That brings clear advantages.

AI startups get easier access to capital and cutting-edge hardware. Nvidia, meanwhile, secures long-term chip demand and stakes an early claim in companies that could deliver the next big AI breakthrough.

Step by step, Nvidia is shifting from hardware vendor to central power broker across the AI stack.

That strategy also raises questions.

Analysts note Nvidia is increasingly financing companies that then buy large volumes of Nvidia hardware. Those mutual dependencies could make the market more fragile if AI investment cools.

Reuters Breakingviews warned Monday that Nvidia is taking on growing credit and investment risk while many AI firms have yet to prove profitability.

Nvidia’s expanding influence could also further squeeze competition. Nearly every major AI developer depends on Nvidia hardware, while Nvidia itself is becoming a shareholder in more players across the same market.

The SSI deal signals that AI’s next phase is about more than smarter models.

Access to chips, data centers, power, and financing is just as critical—and Nvidia now shapes nearly all of it.

If Sutskever does crack a new class of AI, Nvidia will be there from day one as investor, hardware partner, and technology supplier.

Whether SSI has achieved a true breakthrough remains unclear. It hasn’t published research or released a model. For now, this is a multibillion-dollar bet on Ilya Sutskever’s reputation—and the conviction that the next big AI leap will, once again, run on Nvidia hardware.