NVIDIA invests $5B in Scale Seismic Intelligence (SSI), an AI infrastructure/compute software company

On July 27 (local time), Nvidia announced that it agreed to invest approximately $5 billion in Safe Superintelligence Inc. (SSI), an AI startup founded by Ilya Sutskever, former Chief Scientist at OpenAI.

Both parties announced this long-term strategic partnership but did not disclose specific financial terms. According to sources with knowledge of the deal, SSI will also gain access to Nvidia's next-generation Vera Rubin computing platform, with computational capacity expected to increase tenfold within the next 12 months.

This should have been heartening news, yet market reaction proved surprising: on the day of the announcement, Nvidia stock plummeted nearly 5%, and its market cap was surpassed by Apple.

What makes the situation even more puzzling is the investment target itself. SSI, a company whose valuation has been driven to hundreds of billions of dollars, has yet to release a single product or publicly disclose any research findings, and its team comprises only dozens of people.

Its founder Ilya carries his own baggage. He was drawn into co-founding OpenAI at Elon Musk's behest, was a central figure in the board drama that ousted Altman in 2023, and only founded SSI after departing OpenAI in 2024.

SSI is an AI company founded by Ilya Sutskever, OpenAI's former Chief Scientist, in 2024. Unlike OpenAI, Anthropic, and similar companies, SSI has not prioritized rapidly launching models or capturing user markets since its inception. Instead, it has consistently locked its sights on a longer-term objective: developing safe and controllable superintelligence.

According to sources familiar with the matter, the strategic partnership between SSI and Nvidia took only a few weeks from initial contact to finalization. According to both companies' statements, SSI will receive substantial flagship GPU resources from Nvidia and be among the first to use Nvidia's next-generation Vera Rubin computing platform, which will help SSI increase its available computational capacity to ten times its current level within the next 12 months.

In a statement, Ilya said the company's research has reached a stage worthy of scaling, and gaining access to Nvidia's large-scale computing infrastructure would enable them to take that step.

Nvidia founder Jensen Huang responded by noting that Ilya's early work on AlexNet and similar projects had made pioneering contributions to modern AI's foundational layer, and he looked forward to seeing SSI's progress with the new platform.

Notably, Nvidia revealed that this investment was finalized only after gaining deep understanding of SSI's closely guarded research progress—the closest the outside world has yet come to glimpsing what SSI is actually researching. However, as of now, neither party has disclosed a single detail.

In fact, this is not Nvidia's first investment in SSI. According to reports, SSI completed its previous funding round last year at approximately a $32 billion valuation, with Nvidia already being a shareholder at that time. This investment represents an increase in an existing relationship.

Prior to this collaboration, SSI's computing power primarily relied on TPUs from Google, an Alphabet subsidiary. The addition of Nvidia's GPU resources now also represents an adjustment in its supplier structure.

As SSI received Nvidia's backing, it also operated against a backdrop of intensifying competition in AI infrastructure. Throughout this year, chipmakers including AMD and Nvidia have increasingly used investment to bind top AI companies to themselves, and the relationship between compute suppliers and model companies is becoming increasingly intimate.

On July 22, AMD, Nvidia's competitor, just announced plans to invest up to $5 billion in Anthropic. Tracing further back, in March this year, Nvidia had already reached a similar large-scale arrangement with Thinking Machines Lab, led by Mira Murati.

According to reports, Nvidia is currently in talks on AI infrastructure-related deals totaling over $750 billion, which besides this SSI investment includes a potential $250 billion guarantee for an OpenAI data center project and a collaboration with South Korea's SK Group worth over $500 billion.

This naturally raises concerns that such cycles—where "chipmakers invest in AI companies, and AI companies in turn procure computing power from chipmakers"—are blurring the boundaries between suppliers and customers.

The credit market provided early warning signals. On the day the news was announced, Nvidia's five-year debt default swap protection costs rose by some 0.14 basis points at one point, marking the largest single-day increase for the contract since active trading began in November 2025. In stock trading that day, Nvidia closed down 4.99%, while AMD fell 5.17%.

Attention ultimately returned to SSI: what kind of value is hidden within this company that has maintained extremely high research secrecy yet managed to achieve a valuation in the hundreds of billions of dollars in a short time?

SSI was founded in June 2024 by Ilya along with Daniel Gross, who previously led AI projects at Apple, and Daniel Levy, his colleague from OpenAI. The team is based in California and Tel Aviv, Israel.

Since its founding, SSI has not launched any product and has explicitly stated that it has no intention to sell AI services externally in the near term. The team has consistently numbered only dozens of people.

This approach—concentrating the team, capital, and compute on a long-term objective while trying to avoid being disrupted by competitive market dynamics—stands out as distinctly unconventional in an industry where everyone is rushing to release models and capture users.

In September 2024, SSI completed its Series A funding round of $1 billion, with investors including Andreessen Horowitz, Sequoia Capital, DST Global, and SV Angel, valuing the company at $5 billion.

In March 2025, SSI reportedly completed another funding round of approximately $2 billion. In June 2025, co-founder Daniel Gross left the company, with Ilya formally becoming CEO and Daniel Levy taking on the role of President.

Prior to Nvidia's investment, according to PitchBook data, SSI had accumulated approximately $7 billion in total funding, with a valuation of $32 billion, and its investor roster included institutions such as Alphabet and GV.

Ilya was born in the Soviet Union and grew up in Israel, and is recognized as a mathematical prodigy. He studied at the University of Toronto in Canada under Geoffrey Hinton, one of the "Big Three" of deep learning.

In 2012, Ilya and his advisor Hinton, along with another graduate student, jointly published the AlexNet paper, demonstrating that scaling up deep neural networks combined with massive data and computational resources could significantly enhance system capability. This work is widely credited with opening the curtain on the modern AI era.

Subsequently, Ilya joined Google Brain as a research scientist, and in 2015 was persuaded by Elon Musk to become one of OpenAI's co-founders, serving as Chief Scientist. He later directed the development of a series of critical models from GPT-1 to GPT-4 and ChatGPT.

This is exactly the kind of core figure who made significant contributions to OpenAI's success, yet in 2024 chose to leave the company and seek out his own path in AI.

OpenAI was formally established in December 2015, founded by Musk, Sam Altman, and other Silicon Valley heavyweights as a nonprofit organization. Ilya joined the founding team at Musk's behest.

Over the nearly ten years that followed, Ilya, drawing on his intuition and judgment about deep learning, came to be widely regarded as one of the key figures behind OpenAI's success.

The turning point came on November 18, 2023. On that day, without advance notice, OpenAI's board voted to suddenly oust CEO Sam Altman, and Ilya was one of the core participants.

The situation quickly spiraled: employees moved to follow Sam Altman out en masse, Microsoft prepared to step in, and the company teetered on the brink of dissolution.

Seeing things spiral out of control, Ilya reversed course a few days later that weekend, publicly apologized for his role in the board action, and signed the employee petition demanding Sam Altman's return.

Ultimately, Sam Altman and co-founder Greg Brockman returned to the company, and the turmoil settled.

But for Ilya, the conflict seemed far from truly resolved. On May 15, 2024, he announced on social media his "decision to leave OpenAI," leaving only one pregnant phrase: a wish that people would "build AGI that is both safe and beneficial."

A month later, he co-founded SSI with Daniel Gross and Daniel Levy.

This past chapter gained new details this May. On May 11, in the courtroom proceedings of Musk's lawsuit against OpenAI, Ilya took the stand in rare public testimony, for the first time confirming that he holds approximately $7 billion in OpenAI equity and revealing that he had spent a year compiling a 52-page memo that he submitted to the board.

Ilya acknowledged in court that what he regretted was not collecting evidence or voting to oust, but "participating in board action itself": "I have a strong sense of belonging to OpenAI. I don't want it destroyed."

From joining OpenAI in 2015, through the 2023 board drama that shook the global AI community, to founding SSI in 2024 and now securing Nvidia's $5 billion bet, Ilya took a decade to forge a path that is entirely his own.

A company established just two years ago, still maintaining a low-key research posture, was able to attract substantial investment from the world's highest-market-cap chipmaker—not on the basis of a business story, but on the credibility Ilya has accumulated in this industry, and the technical path he represents, which diverges sharply from the mainstream large model companies.