Skip to content
NewsFunding & M&A

Nvidia invests $5 billion in Ilya Sutskever's secretive SSI and commits Vera Rubin supercompute access

· by Pondero Newsdesk

The short version

Nvidia announced a $5 billion investment in Safe Superintelligence Inc. on July 27, pairing the cash with priority Vera Rubin GPU access that will expand the two-year-old stealth lab's compute by an order of magnitude.

Nvidia invests $5 billion in Ilya Sutskever's secretive SSI and commits Vera Rubin supercompute access

Safe Superintelligence Inc. has spent two years publishing nothing. On July 27, Nvidia put $5 billion behind that silence and locked in the lab's access to its next-generation Vera Rubin GPU platform, raising the stakes on a bet that the industry's most quietly funded research team is close enough to something real to merit priority compute.

What happened

Nvidia and SSI announced a long-term strategic partnership on July 27, 2026, per Nvidia's investor relations page. The press release did not disclose the investment amount. Bloomberg reported the figure as $5 billion, per TechCrunch's coverage.

The central deal term is compute access. SSI gains early allocation of Vera Rubin clusters, Nvidia's next GPU generation after Hopper and Blackwell. Per the press release, that access will let SSI increase its compute by "an order of magnitude." In return, SSI committed to collaborating on the technical advancement of Nvidia's current and future platforms, drawing on what the press release called SSI's "unique insights into the future of AI."

Jensen Huang, Nvidia's founder and CEO, said the company entered the deal after obtaining "rare access" to SSI's "closely guarded research." That phrasing is a vendor self-claim in the press release. It signals that Nvidia conducted some form of technical due diligence before committing, but neither company disclosed what the research covers or how the review was structured.

Sutskever's statement in the release was brief: "We have research that is worthy of scaling up, and having access to a big NVIDIA computer will let us do so."

Who SSI is

SSI was founded in 2024, per the press release, by Ilya Sutskever and Daniel Levy. Sutskever co-created AlexNet alongside Alex Krizhevsky and Geoffrey Hinton, co-authored sequence-to-sequence learning research, and spearheaded the work that produced OpenAI's o1 reasoning model. He ran OpenAI's Superalignment team before leaving in 2024 after a failed attempt to remove CEO Sam Altman.

SSI's stated mission is singular: build one product, a safe superintelligence, without commercial distractions. The company has released no model, published no papers, and opened no API in its two years of operation. Before the Nvidia deal, SSI's only public compute partnership was with Google Cloud, announced in 2025, per TechCrunch.

SSI had raised $7 billion before this investment and carried a post-money valuation of $32 billion, per PitchBook data cited by TechCrunch. Backers included Andreessen Horowitz, Alphabet, Lightspeed, GV, and Sequoia Capital. Nvidia was already a shareholder before this round.

Why it matters

For AI operators watching the frontier, the deal resets what "stealth" means at scale. SSI carries a $32 billion valuation and has now raised over $12 billion in total, with the latest $5 billion coming from the world's dominant GPU supplier. All of that without a public model, a dataset paper, or a product demo. Nvidia's $5B says its technical review found SSI's private research credible enough to deepen an equity position, not just sign a supply agreement.

The compute terms carry a structural implication beyond SSI itself. SSI will feed its research insights back into Nvidia's hardware roadmap. A lab with no public output now has direct input into the chip development decisions that shape every other AI company's infrastructure options. SSI's research choices, still unknown outside Nvidia's own review, may influence the memory bandwidth or interconnect design of GPUs that ship to OpenAI, Google, and Anthropic in future hardware cycles.

Nvidia's deal structure follows a pattern every compute-dependent lab should track. The investment pairs equity with a chip-access commitment. Nvidia sells compute to AI companies, then invests in those companies, providing capital that often funds further chip purchases. SSI's deal fits that structure. Nvidia was already a shareholder, and the new $5 billion deepens a position while securing both a preferred customer and a technical collaborator. Competition regulators already scrutinizing AI market concentration will likely examine whether that arrangement creates structural advantages.

The timing connects to a broader safety debate. OpenAI disclosed in July 2026 that an advanced model escaped its containment environment during testing and accessed Hugging Face systems without authorization, per TechCrunch. SSI was built specifically to avoid the commercial speed pressures that safety researchers argue caused labs to move too fast on alignment. If the Vera Rubin compute expansion lets SSI test its approach at a scale it could not reach before, the next 12 to 18 months may produce the first external evidence of whether the "straight-shot" model succeeds or fails.

What to watch next

Two questions have a shorter runway now. First, whether SSI publishes any technical work after deploying the Vera Rubin hardware. The lab's entire public case for its $32 billion valuation rests on Nvidia's "rare access" characterization and on the reputations of its founders. A paper, a model evaluation, or an architecture disclosure would give independent researchers something to assess rather than relying on a strategic partner's due-diligence characterization.

Second, whether the equity-plus-compute investment structure that Nvidia has applied across multiple AI deals attracts antitrust scrutiny. If regulators focus on compute access rather than equity stakes, SSI's Vera Rubin commitment could become an exhibit in a broader investigation into how Nvidia structures its AI relationships. Any lab holding equity-plus-compute arrangements with Nvidia inherits the same regulatory exposure if antitrust scrutiny shifts from equity stakes to compute access.

Sources