Nvidia invests $5 billion in Sutskever's AI startup SSI

Nvidia committed $5 billion to Ilya Sutskever's Safe Superintelligence in one of the chipmaker's largest AI startup investments.

Nvidia Corp. committed $5 billion to Safe Superintelligence Inc., the AI lab founded by OpenAI co-founder Ilya Sutskever, in a deal that gives the chipmaker access to one of the industry's most secretive research operations while pulling a major customer away from Google's competing chip platform.

"We have research that is worthy of scaling up, and having access to a big Nvidia computer will let us do so," Sutskever, co-founder and chief executive of SSI, said in a statement.

The investment, reported earlier by Bloomberg, grants SSI access to Nvidia's next-generation Vera Rubin platform, which the companies said could expand the startup's computing capacity by an order of magnitude. SSI had previously relied on Google's tensor processing units, or TPUs, for its compute needs. The deal came together within weeks, according to people familiar with the matter.

For Nvidia, the deal extends a strategy of investing in high-profile AI startups that become major customers for its hardware — a model that has drawn scrutiny from investors including Michael Burry. Nvidia shares fell as much as 5.3% Monday amid broader weakness in the semiconductor sector, as China's progress in domestic chipmaking weighed on the industry.

The SSI Playbook

SSI was founded in 2024 by Sutskever, who previously served as chief scientist at OpenAI and was instrumental in the development of ChatGPT. The lab operates with a singular focus: building "safe superintelligence" — AI systems capable of outperforming humans across all tasks while remaining under human control. The company has raised roughly $2 billion from venture capital firms including Andreessen Horowitz and Sequoia Capital, reaching a valuation of about $30 billion, despite disclosing almost no research publicly.

Sutskever described the company's research as focused on "aspects of human brain function that have been overlooked," without providing further detail. The partnership with Nvidia represents SSI's most substantive public move since its founding.

Nvidia's Bet on OpenAI Alumni

The SSI investment follows a similar deal in March, when Nvidia invested in Thinking Machines Lab, the startup founded by former OpenAI chief technology officer Mira Murati. That lab released its first AI model, trained on Nvidia hardware, earlier this month.

Both transactions follow the same logic: Nvidia deploys capital to lock in high-profile AI researchers as customers, ensuring they build on its CUDA ecosystem rather than rival platforms from Google, Advanced Micro Devices Inc., or Amazon.com Inc.'s AWS. The strategy creates a self-reinforcing cycle — Nvidia's investment dollars help startups scale, and those startups' growing compute needs flow back to Nvidia as hardware revenue.

The Sutskever Arc

Sutskever, born in the former Soviet Union and raised in Israel, was a mathematics prodigy who studied under Nobel laureate Geoffrey Hinton at the University of Toronto. In 2012, he co-authored a landmark paper on computer vision that demonstrated scaling deep neural networks with massive data and compute could produce measurable gains in intelligence — a finding that ignited the modern AI race.

He joined Google and later co-founded OpenAI in 2015, where his conviction that scaling data and compute would lead to artificial general intelligence — or AGI — directly shaped the development of ChatGPT. His relationship with OpenAI Chief Executive Sam Altman ended dramatically when Sutskever voted to remove Altman from the board, then left the company months after Altman's return.

Since departing OpenAI, Sutskever has publicly questioned whether simply scaling existing architectures is sufficient to achieve true breakthroughs, suggesting SSI may pursue a different technical path.

What's at Stake

The deal reshapes the competitive dynamics of AI infrastructure. By winning SSI as a customer from Google, Nvidia not only secures a high-profile user for its Vera Rubin platform but also signals to other AI labs that its hardware ecosystem offers advantages over custom chips like Google's TPU. For SSI, a tenfold increase in compute capacity could dramatically accelerate its research timeline.

Nvidia trades at roughly 35 times forward earnings, reflecting investor expectations that AI infrastructure spending will continue to grow. The company's data center revenue reached $47.5 billion in its most recent fiscal year, driven by demand from both cloud providers and AI startups. The SSI investment, while large, represents less than 1% of Nvidia's roughly $3 trillion market capitalization.

This article is for informational purposes only and does not constitute investment advice.