Nvidia (NVDA) is no longer willing to only sell the gear behind the artificial intelligence revolution.The chipmaker has taken a stake in Safe Superintelligence, the secretive AI lab set up by ex-OpenAI chief scientist Ilya Sutskever, in a long-term strategic relationship disclosed July 27.Nvidia did not disclose the size, valuation, or terms of the investment.Safe Superintelligence, commonly known as SSI, will have access to Nvidia’s next-generation Vera Rubin systems, which will let the lab ramp up its computing power by an order of magnitude. The firms also intend to work together on existing and future computing platforms from Nvidia.For investors at Nvidia, the deal is more than just another startup investment.It gives Nvidia a strong frontier-model developer focused on Vera Rubin, technical feedback that can influence future products, and perhaps one more long-term customer for its entire computing platform.The trade-off is that Nvidia is now becoming a key source of finance to the same AI industry that’s buying its infrastructure.That approach can expedite adoption, but it also makes it more important for shareholders to separate independent customer demand from growth fueled in part by Nvidia’s own expenditures.“Ilya has pioneered fundamental breakthroughs at the foundation of modern AI, beginning with AlexNet,” Nvidia CEO Jensen Huang said. “We are excited to see what new breakthroughs SSI will discover powered by our Vera Rubin platform.”Nvidia is building demand around Vera RubinSSI was formed in 2024 with a single declared goal: to build safe superintelligence. The company, which operates from Palo Alto, Calif., and Tel Aviv, has revealed little publicly about its research, product roadmap, or commercial aspirations.Nvidia said it entered the alliance after gaining rare access to SSI’s highly held work.That access could be crucial because SSI is not just receiving hardware. Its researchers will work with Nvidia on the technical development of current and future computing platforms, potentially offering the company an early glimpse into the requirements of the next generation of AI models.Immediately in front is Vera Rubin.A Vera Rubin NVL72 rack includes 72 Rubin GPUs, 36 Vera central processing units, and Nvidia networking and data-processing components. Nvidia believes Rubin can train some mixture-of-experts models with four times fewer CPUs and perform AI inference at up to one-tenth the cost per token compared to Blackwell. Those are Nvidia performance promises, not independent measurements.Rubin-based products are scheduled to be available via partners in the second half of 2026, with major cloud providers such as Amazon Web Services, Google Cloud, Microsoft and Oracle preparing deployments.Related: Nvidia stock is doing something it hasn’t done in yearsSSI CEO Ilya Sutskever said the laboratory has research “worthy of scaling up” and that access to a large Nvidia computer would enable the next stage of that work.That endorsement is another high-profile affirmation of Rubin, as the platform becomes a big income producer.Nvidia’s startup investments are becoming a stock storyNvidia can easily afford the SSI contract.Revenue in the first quarter of fiscal 2027 was $81.6 billion, an increase of 85% from the prior-year period. Data Center sales rose 92% to $75.2 billion with operating profitability of $53.5 billion. Nvidia forecast revenue of around $91 billion for the second quarter.What Nvidia investors should watchThe financial size and ownership terms of Nvidia’s SSI investment.When SSI begins deploying Vera Rubin infrastructure.Whether the partnership leads to identifiable product revenue.How SSI’s research influences Nvidia’s future platforms.Nvidia’s growing exposure to private AI-company valuations.And it is getting increasingly impossible for shareholders to see past its growing investment portfolio.More AI:Workers just sent AI companies an ultimatumPalantir CEO has a blunt verdict on OpenAI and AnthropicElon Musk pulls no punches with AI rivals as Grok 4.5 debutsNvidia reported $42.3 billion in nonmarketable equity securities as of April 26, mainly investments in privately held companies.Nvidia reported net additions of $17.9 billion, leaving a balance of $22.3 billion at the beginning of the quarter. Those numbers are before the announcement of the SSI investment.
Nvidia’s SSI bet tests how far its AI flywheel can runPHILIP FONG / Getty Images
The SSI deal brings Nvidia opportunity and added riskThe case for a bullish view is obvious.Nvidia can foster deeper ties with the labs that are projected to be the biggest consumers of its most powerful computer systems by investing early. SSI gets funds and hardware, and Nvidia gets a partner that can test Rubin against state-of-the-art research workloads.That might help Nvidia maintain its technological dominance and build demand outside the big cloud companies that dominate Data Center sales.The issue is that financing for startups becomes an increasingly crucial aspect of maintaining AI infrastructure growth.Nvidia has said smaller firms often have trouble getting access to cash to finance huge computer initiatives. It also stated constraints of funding, data-center capacity or energy could slow client deployments and reduce the scale of AI adoption.The SSI cooperation appears designed to overcome at least some of these limitations.Nvidia is offering investment funding and access to computer infrastructure, but the businesses did not indicate how SSI would pay for the systems or if it has committed to a set volume of purchases.The lost information restricts the immediate benefit of the purchase to Nvidia’s profit outlook.The accord is still strategically significant. It reveals how Nvidia is leveraging its balance sheet, hardware, and engineering department in concert to influence what platforms upstart AI companies build on.That may provide Nvidia’s stockholders with a huge edge if the labs that receive its finances eventually build products, consumers, and cash flows of their own.The investing thesis is thus more than whether SSI reaches safe superintelligence.Nvidia is banking that owning part of the AI ecosystem would make its chip business stronger. Now investors have to see how much capital that plan consumes and how much real, profitable demand it creates in the end.Related: Nvidia just locked down deal that changes AI race