Safe Superintelligence raises $2bn at $32bn valuation with no product

Safe Superintelligence, the company Ilya Sutskever founded after leaving OpenAI, raised roughly $2 billion in a round that valued it at $32 billion, according to reporting confirmed by multiple outlets. Greenoaks Capital led the round, reportedly committing around $500 million, with Andreessen Horowitz, Lightspeed Venture Partners and DST Global returning from earlier rounds and Alphabet and Nvidia joining as strategic investors. SSI did not comment publicly on the figures.

The valuation was roughly six times what the company had commanded seven months earlier, when it raised $1 billion at a $5 billion valuation in September 2024, itself only months after Sutskever, Daniel Gross and Daniel Levy founded the company in June 2024. SSI had shipped no public product, published no model, and stated no near-term plan to do either — its stated mission, in Sutskever’s words, was to reach “a safe superintelligence” and only then decide what to release.

The round illustrated how much of frontier-AI valuation in this period tracked researcher reputation rather than conventional business fundamentals: investors were pricing SSI primarily on Sutskever’s record as OpenAI’s former chief scientist and co-inventor of several of the field’s central techniques, alongside the credibility of Gross and Levy, rather than on revenue, users or a released product. Alphabet’s participation was notable given Sutskever’s departure from a rival lab; Nvidia’s investment mirrored a broader pattern of the chipmaker taking equity stakes in AI labs that were also large customers for its hardware.