Safe Superintelligence Inc., the AI lab founded by former OpenAI chief scientist Ilya Sutskever, has announced a partnership with Nvidia aimed at increasing its compute capacity tenfold within a year. However, no official announcements have corroborated such claims, and Nvidia’s role appears to be primarily financial, providing equity investment rather than hardware support.
SSI raised a funding round in April 2025 that valued the company at approximately $32 billion. Both Nvidia and Alphabet/Google participated as investors in that round.
SSI has also been working closely with Google Cloud, which has been supplying its Tensor Processing Units (TPUs) for research purposes. Reports have indicated that SSI actually prefers TPUs over Nvidia’s GPUs for certain workloads.
Who is SSI, and why does this matter
SSI was founded by Ilya Sutskever in 2024, shortly after his departure from OpenAI. His new venture has positioned itself as a pure research operation with a single stated goal: building safe superintelligence. No chatbots, no API products, no consumer apps.
The $32 billion valuation for a company with no shipping product reflects how the market is pricing AI research talent. Strategic capital from the likes of Nvidia and Alphabet isn’t just financial — these companies want proximity to whichever lab cracks the next major capability threshold.
What this means for investors
Nvidia isn’t just selling chips. It’s making equity investments in the labs that consume those chips, creating a vertically integrated ecosystem where hardware supplier and customer are financially intertwined.
SSI’s willingness to use both Nvidia GPUs and Google TPUs suggests the AI hardware market hasn’t consolidated around a single winner. Reports concerning SSI’s compute capacity continue to lack definitive details, indicating a cautious approach to projections about its operational capabilities in the near term.