Nvidia backs Safe Superintelligence (SSI) tenfold compute expansion plan, providing chip supply and infrastructure support to the AI startup founded by Ilya Sutskever.

Nvidia is investing a reported $5 billion in Safe Superintelligence (SSI) and has entered a long-term partnership with the AI research lab. SSI, led by co-founder and CEO Ilya Sutskever, will gain access to Nvidia's advanced Vera Rubin AI computing platform. Sutskever expects the infrastructure to substantially expand the processing resources available for SSI's research.

SSI's plan calls for a tenfold capacity increase within the 12 months following the partnership announcement. Nvidia's equity investment totals about $5 billion, according to a person briefed on the deal, though neither Nvidia nor SSI has publicly confirmed the amount or disclosed other financial terms. The deployment of usable hardware will serve as the partnership's first measurable milestone.

Nvidia is both an infrastructure supplier and financial partner, extending a financing model it has deployed with AI-cloud providers. The company is financing a long-duration research initiative and supplying infrastructure for expanded experiments—not hardware for a product nearing commercialization. SSI is pursuing safe superintelligence, a hypothetical AI capability that exceeds human performance across numerous tasks, with no interim commercial release planned.

Vera Rubin gives SSI a path to expanded computing capacity while maintaining the independence of its research mission. However, SSI's results will ultimately depend on deploying delivered hardware and using it for sustained, meaningful work.

Previously, SSI relied primarily on Google Cloud for research, using Google's tensor processing units (TPUs)—specialized chips for machine-learning calculations. The shift toward Vera Rubin represents a significant expansion via Nvidia infrastructure, though the disclosed terms do not establish that SSI will abandon its Google systems.

Nvidia and SSI plan to collaborate on computing platforms informed by insights from SSI's work. SSI gains access to advanced infrastructure, while Nvidia may receive technical insights from a frontier research customer. Neither company has publicly defined the scope of that collaboration or how directly such insights will inform Nvidia's platform development.

Sutskever tied the hardware expansion to SSI's research maturity: "We have research that is worthy of scaling up, and having access to a big Nvidia computer will let us do so."

More computing power can support larger or more numerous experiments, but it cannot independently validate SSI's methods or guarantee that the lab's work will produce safe superintelligence. SSI's researchers must deploy and operationalize the installed systems to create sustained capacity and repeatable experiments before the expansion demonstrably changes the lab's research capabilities.

Sutskever and two partners founded SSI in 2024 with a singular focus on safe superintelligence. The lab prioritized this focused research mission over recurring product cycles.

To date, SSI has disclosed minimal information about its research publicly. The lab has raised two financing rounds: $1 billion in 2024 and $2 billion in 2025, with the latter valuing SSI at a reported $32 billion. The first round closed before SSI had released any public AI model or established any product-revenue track record.

SSI defines safe superintelligence as its sole product and roadmap, with safety and capabilities developed in parallel. The lab plans to withhold interim commercial models until reaching this goal. This approach leaves outsiders with few intermediate measures of research progress, though computing capacity gains can be measured sooner than scientific breakthroughs.

SSI must now translate planned Vera Rubin systems into operational capacity. The 12-month period following the announcement will be critical: whether SSI successfully deploys and operates sufficient hardware to achieve its tenfold expansion target will represent the partnership's most concrete near-term test.