TL;DR
Nvidia is investing $5 billion in Safe Superintelligence Inc. (SSI), the AI safety startup founded by former OpenAI chief scientist Ilya Sutskever. The investment values SSI at over $50 billion and includes a strategic partnership for GPU infrastructure. The deal deepens Nvidia’s bet on AI safety and positions SSI as one of the best-funded labs pursuing superintelligence.
The Deal
The $5 billion investment is split between primary funding and secondary share purchases. It makes Nvidia one of SSI’s largest investors alongside Andreessen Horowitz and Sequoia Capital. The deal includes a strategic partnership where Nvidia will provide SSI with priority access to its latest GPU clusters for training runs.
SSI was founded in 2024 by Ilya Sutskever, Daniel Gross, and Daniel Levy after Sutskever departed OpenAI. The company’s stated mission is to develop “safe superintelligence” — AI systems that surpass human capabilities while remaining aligned with human values.
The company has raised approximately $15 billion total, making it one of the most well-funded AI startups in history despite having no commercial products.
Why Nvidia Cares
The investment reflects Nvidia’s strategy of backing the most promising AI labs regardless of their commercial viability. CEO Jensen Huang has repeatedly stated that the company’s revenue is driven by AI training and inference demand, which grows as labs push toward more capable models.
Sutskever’s reputation as one of the most accomplished AI researchers gives the investment credibility. His work on scaling transformers at OpenAI directly contributed to the GPT series, and his expertise in alignment research is considered unmatched in the industry.
The partnership also gives Nvidia influence over one of the few labs pursuing superintelligence with a safety-first approach — a hedge against the possibility that AGI arrives before alignment solutions are ready.
SSI’s Approach
Unlike competitors that focus on commercial applications, SSI has no plans to release products or APIs. The company operates with a small team of researchers working on fundamental safety research. Sutskever has described the approach as “building superintelligence the right way.”
The company’s headquarters remain in Palo Alto, with a research lab in Tel Aviv. SSI has not disclosed what model it is currently training, though leaks suggest the company is working on a large-scale language model with novel safety architectures.
Market Implications
The investment signals that Nvidia is doubling down on its position as the infrastructure layer of the AI industry. By backing both commercial and safety-focused labs, Nvidia ensures demand for its chips regardless of which approach to AI development proves most successful.
For the AI safety community, the deal provides validation and resources for one of the few labs prioritizing alignment over commercial deployment. Whether that approach produces meaningful safety breakthroughs remains to be seen.