Storia
luglio 30, 2026

Nvidia’s $5 billion SSI deal sharpens the safety pitch — and the AI bubble risk

Safe Superintelligence (SSI), the new AI company founded by former OpenAI co-founder Ilya Sutskever, has announced a major partnership with Nvidia that includes a multi-billion dollar investment. The collaboration will provide SSI with access to Nvidia's advanced Vera Rubin GPU platform to significantly scale its research into developing safe artificial superintelligence.

Nvidia is pouring billions into Ilya Sutskever’s Safe Superintelligence at the exact moment the AI industry is arguing over what comes next: safer systems, or a bigger, more fragile boom. The deal looks like both a vote of confidence in AI alignment research and a reminder that the sector’s financial plumbing is starting to look uncomfortably circular.

On one reading, this is a straightforward power move. SSI, the secretive lab founded by the former OpenAI chief scientist, says it is ready to scale, and Nvidia is supplying the chips and capital to do it. TechCrunch reported the partnership will give SSI access to Nvidia’s Vera Rubin platform and boost compute resources “by an order of magnitude,” while Sutskever said, “We have research that is worthy of scaling up.” The Verge, citing the same announcement, said the deal would let SSI “10x our compute in the next 12 months,” after the company concluded, “We reached the point where our research is worth scaling.”

That is the bullish case: more compute in service of safety. SSI has pitched itself as pursuing a “straight shot” toward “safe, aligned artificial superintelligence,” avoiding the product treadmill and short-term revenue pressures that define much of the industry. The Financial Times cast Nvidia’s move even more bluntly as a giant wager on “Ilya Sutskever’s AI breakthrough.”

But the other reading is less flattering. Axios warned the SSI tie-up lands as Nvidia also explores deeper financial entanglements with OpenAI, reviving worries about “circular” AI deals in which the company that sells the picks also bankrolls the miners. In that view, safety is part of the story, but so is market structure: Nvidia is no longer just the arms dealer of the AI race; it is increasingly becoming a financier of it.

Even the rhetoric around openness cuts both ways. In a repost of Jensen Huang’s message, Yann LeCun amplified Nvidia’s claim that “Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.” Supporters see that as the industry’s roadmap. Skeptics may see it as branding for a business model that keeps getting bigger, costlier and harder to untangle.