Safe Superintelligence

AI · AI Safety · Semiconductors

Partnership

Nvidia reportedly pours $5B into Ilya Sutskever's Safe Superintelligence

A research lab with no product, no revenue, and no launch plans just landed in the same GPU queue as OpenAI.

• Nvidia and Safe Superintelligence (SSI) announced a long-term strategic partnership on July 27, 2026, with Nvidia making an equity investment reported at $5 billion and giving SSI access to its next-gen Vera Rubin GPU platform.
• SSI was founded in 2024 by Ilya Sutskever, OpenAI's former chief scientist, and Daniel Levy, pursuing a 'straight shot' approach to building safe superintelligence with no interim products or revenue plans.
• The deal expands SSI's compute by an order of magnitude and marks a shift from the Google TPUs it relied on early on to Nvidia's GPU ecosystem, including software and networking tools.
• SSI previously raised roughly $3 billion at a $32 billion valuation in April 2025 from Andreessen Horowitz, Sequoia Capital, DST Global and Greenoaks, before this new Nvidia investment.
• Nvidia has been seeding equity in frontier AI labs led by ex-OpenAI researchers, including Mira Murati's Thinking Machines Lab, a strategy that locks in future GPU demand years before these labs generate revenue.
• The bet signals that Nvidia sees long-horizon alignment research as strategically valuable even without a commercial roadmap, a notable stance as rivals race to ship products despite emerging safety incidents.