Safe Superintelligence partners with Nvidia to expand its AI research

In short: Safe Superintelligence, the AI lab founded by Ilya Sutskever, has signed a long-term partnership with Nvidia to get more computing power for its research.

What happened

Safe Superintelligence, often called SSI, has been working quietly for about two years. The company says it is focused on building AI that is “safe” and “aligned,” meaning it should follow human intentions and rules.

SSI announced a long-term partnership with Nvidia, the company best known for making the chips many AI systems run on. The deal includes an investment amount that was not shared publicly. A source familiar with the deal told TechCrunch Nvidia’s investment is in the “multiple billions.”

SSI will get access to Nvidia’s Vera Rubin GPU platform. A GPU is a type of computer chip, and you can think of it like a very large engine used to train and test AI models faster. Nvidia said this could increase SSI’s available computing power by about 10 times.

SSI said it has reached research milestones and now has work “worthy of scaling up,” according to a statement from Sutskever. Nvidia also said it signed the partnership after getting “rare access” to SSI’s research.

SSI previously partnered with Google Cloud to power parts of its research. The company has raised $7 billion in total and was valued at $32 billion after its last funding, according to PitchBook.

Why it matters

Training advanced AI needs huge amounts of computing power, which is expensive and hard to get. Deals like this can speed up what a research lab can build, and they also raise questions about how quickly powerful AI should be developed and tested for safety.

Source: TechCrunch AI