Back
Nvidia invests in Ilya Sutskever’s AI lab as part of long-term partnership

Nvidia invests in Ilya Sutskever’s AI lab as part of long-term partnership

The multiyear deal gives Safe Superintelligence access to Nvidia’s Vera Rubin systems and is set to increase available compute tenfold over the next 12 months, with the equity investment valued by people briefed at about $5 billion.

Fact Check
The WSJ primary source directly confirms Nvidia invested in Ilya Sutskever's Safe Superintelligence lab as a long-term partnership with undisclosed financial terms, announced Monday. BlockBeats corroborates the order-of-magnitude boost to compute via Nvidia GPUs (SSI previously relied on Google TPUs). Reuters provides consistent background on the prior Nvidia investment relationship. Every element of the claim — the investor (Nvidia), the recipient (SSI/Sutskever), the long-term partnership framing, the order-of-magnitude compute increase, and undisclosed terms — is supported by the sources.
    Reference: 123
Summary

Nvidia has agreed to invest about $5 billion in Safe Superintelligence, the AI startup led by OpenAI co-founder Ilya Sutskever, as part of a long-term strategic partnership that gives the company access to Nvidia’s Vera Rubin computing systems. Nvidia and SSI described the investment only as substantial and did not publicly disclose its value, but people briefed on the agreement said the equity investment was approximately $5 billion. The companies said the infrastructure will lift SSI’s available computing capacity tenfold over the next 12 months. SSI, which launched in 2024 and has not released a product or published details of its research, was valued at $32 billion in a previous $2 billion funding round. The deal adds to Nvidia’s strategy of investing in AI developers that are also major users of its hardware, underscoring how access to advanced chips and computing power remains central to frontier AI research.

Terms & Concepts
  • Vera Rubin computing systems: Nvidia hardware systems designed to provide large-scale computing power for advanced AI training and research.
  • compute: The processing capacity available to train and run artificial intelligence systems.