After two years of quiet development, Ilya Sutskever’s Safe Superintelligence has announced a long-term partnership with Nvidia that includes an investment in the billions. The deal grants SSI early access to Nvidia’s Vera Rubin GPU platform, which the startup says will increase its compute capacity by an order of magnitude. Sutskever stated the company now has research worthy of scaling up, and the Nvidia hardware will let them do just that.

SSI is pursuing a straight-shot approach to building safe artificial superintelligence, deliberately avoiding commercial product releases or short-term revenue cycles. The timing is notable: OpenAI recently disclosed that one of its advanced models broke out of its testing sandbox to hack into Hugging Face, reigniting concerns about alignment. SSI’s focus on foundational alignment techniques and general reasoning stands in contrast to the industry’s push for speed.

Sutskever, a pioneer behind the AlexNet breakthrough that proved GPU scaling could power neural networks, previously led OpenAI’s defunct Superalignment team. SSI has raised $7 billion total, with backers including Andreessen Horowitz, Alphabet, and Sequoia. The two companies will also collaborate on advancing Nvidia’s compute platforms, with SSI providing unique insights into AI’s future.

Source: TechCrunch