Nvidia partners with Ilya Sutskever's AI startup to build superintelligence
Safe Superintelligence taps Nvidia’s latest compute platform in a long-term deal aimed at scaling safety‑focused AI research.
Building powerful AI requires more than better models. It also depends on access to massive computing power. Nvidia and Safe Superintelligence (SSI), the AI startup founded by former OpenAI chief scientist Ilya Sutskever, have announced a long-term partnership to expand the company's research capabilities.
The collaboration, announced on July 27, 2026, will give SSI access to Nvidia's next-generation AI computing platform as it works towards building safe artificial superintelligence.
A major boost in computing power
As part of the partnership, SSI will use Nvidia's new Vera Rubin GPU platform. The companies say the upgrade will increase SSI's computing capacity by roughly ten times. More computing power allows researchers to train larger AI models, run more experiments and test new ideas more quickly.
Although neither company disclosed the financial details, TechCrunch reported that Nvidia's investment is worth several billion dollars, while Bloomberg estimated it at around $5 billion.
Why safety remains the focus
SSI was launched with a single goal: building artificial superintelligence that is safe and aligned with human values. Unlike many AI companies that are focused on releasing commercial products, SSI says its priority is long-term research.
The company believes safety should be built into AI systems from the beginning rather than added later. Moreover, this partnership comes at a time when AI safety is receiving increased attention across the industry, following recent discussions about how advanced AI systems should be tested and monitored.
Nvidia will help shape future AI hardware
The collaboration goes beyond providing GPUs. Nvidia said it will work closely with SSI to understand the computing challenges involved in training future AI systems. That feedback could influence the design of future AI chips and large-scale computing clusters.
SSI also continues to use Google's TPU infrastructure for some of its research, suggesting the company plans to use different hardware platforms depending on the task.
What's next?
According to TechCrunch, SSI has raised around $7 billion so far and is valued at approximately $32 billion. Its investors include Andreessen Horowitz, Alphabet, GV, Lightspeed and Sequoia. The partnership with Nvidia signals that SSI is entering a new phase of growth as it scales its research efforts.
The next challenge will be turning that additional computing power into meaningful progress. Industry watchers will be looking to see whether SSI can develop more capable AI systems while maintaining the high safety standards it has promised from the start.