Nvidia Partners With Safe Superintelligence on Vera Rubin
Nvidia and Safe Superintelligence Inc. announced a long-term strategic partnership on July 27, expanding SSI's compute capacity by an order of magnitude through Nvidia Vera Rubin systems. The Financial Times reports a $5 billion Nvidia bet on the AI lab. SSI, founded in 2024, is pursuing what its announcement describes as one goal and one product: a safe superintelligence.
Nvidia and Safe Superintelligence Inc. (SSI), led by Ilya Sutskever and Daniel Levy, announced a long-term strategic partnership on July 27. Nvidia's announcement states that access to its Vera Rubin systems will expand SSI's compute capacity by an order of magnitude.
The Financial Times reports that Nvidia has made a $5 billion bet on SSI. Nvidia's published partnership announcement, as retrieved, describes the compute relationship but does not specify an investment amount or transaction terms.
SSI was founded in 2024. According to Nvidia's announcement, the lab has "one goal and one product: a safe superintelligence." Nvidia identifies Andreessen Horowitz, DST Global, Greenoaks, and Sequoia Capital among SSI's investors. The announcement also credits Sutskever with contributions to AlexNet, AlphaGo, sequence-to-sequence learning, GPT models, and research leading to OpenAI's o1 reasoning models.
Compute access and Vera Rubin
The partnership gives SSI access to Nvidia's Vera Rubin systems. The companies did not publish system counts, deployment dates, training workloads, model specifications, or performance targets in the retrieved announcement.
For ML infrastructure teams, the reported order-of-magnitude compute expansion illustrates the scale at which frontier-model labs increasingly procure capacity. Comparable arrangements can affect access to advanced accelerators, systems integration expertise, networking, power, and data-center capacity, not merely GPU availability.
Nvidia's open-model activity
Gizmodo separately reported that Nvidia announced the Open Secure AI Alliance, a coalition intended to develop and share technologies for safeguarding software and AI agents. Gizmodo listed Palantir, Microsoft, OpenClaw, Salesforce, and SpaceX as inaugural members, and described the initiative as supporting the proliferation of open-weight AI models.
The SSI agreement and the Open Secure AI Alliance concern different parts of the AI ecosystem: compute access for a frontier-model lab on one side, and security tooling for open-weight model deployment on the other. Public reporting has placed the announcements in the continuing debate over open and closed AI systems, though the available Nvidia partnership release provides no additional detail on SSI's research roadmap or planned models.
Key Points
- 1Nvidia's partnership gives SSI access to Vera Rubin systems, increasing the lab's compute capacity by an order of magnitude.
- 2The Financial Times reports a $5 billion Nvidia bet, making the arrangement notable beyond a conventional hardware supply agreement.
- 3Frontier-lab compute deals commonly hinge on systems, networking, power, and deployment integration alongside accelerator availability.
Scoring Rationale
The reported partnership gives SSI access to Nvidia's Vera Rubin systems and substantially expands the lab's stated compute capacity. The Financial Times' reported $5 billion investment figure increases the business and infrastructure significance, although technical deployment details remain undisclosed.
Sources
Primary source and supporting public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
