Nvidia has placed one of its biggest bets yet on the next wave of frontier AI research, investing $5 billion in Ilya Sutskever’s Safe Superintelligence (SSI) and bringing the secretive startup into its hardware ecosystem.
The deal gives SSI access to Nvidia’s Vera Rubin platform and marks a major shift away from Google TPUs, which the company had relied on during its early development. For Nvidia, the partnership strengthens its ties with one of the industry’s most closely watched AI labs. For SSI, it opens the door to far more computing capacity as it pursues its long-term goal of building safe superintelligence.
“Nvidia Corp. has committed to invest $5 billion in Ilya Sutskever’s artificial intelligence startup Safe Superintelligence Inc.,” Bloomberg reported, citing people familiar with the matter. The deal also marks one of the chipmaker’s largest funding deals of the AI boom.
“In addition to the funding, Safe Superintelligence will receive access to Nvidia’s next-generation Vera Rubin platform, according to a joint statement on Monday. Financial terms of the deal were not disclosed. The people described the funding commitment on condition of anonymity, as the information is not public,” Bloomberg added.
The agreement came together within weeks, according to people briefed on the discussions.
In April 2025, Safe Superintelligence raised $2 billion at a $32 billion valuation, highlighting investors’ willingness to back elite AI research teams long before they generate revenue.
Nvidia tightens its grip on frontier AI
The investment extends Nvidia’s strategy of backing influential AI startups long before they become major customers.
SSI joins a growing list of frontier AI companies that have attracted Nvidia’s support. The chipmaker has invested in several high-profile ventures led by former OpenAI researchers, including Mira Murati’s Thinking Machines Lab. Those relationships help Nvidia secure future demand for its GPUs as AI models become larger and more expensive to train.
The SSI partnership carries extra weight because the startup had previously depended heavily on Google’s tensor processing units, or TPUs. Moving into Nvidia’s ecosystem gives SSI access to the software tools, networking infrastructure, and hardware platform that many leading AI developers already use.
The move sends another signal that Nvidia continues to dominate the market for training frontier AI models. Google’s TPUs remain a strong alternative, yet many independent AI labs still choose Nvidia’s broader developer ecosystem and supply network.
A $30 billion company with no product
Few AI startups have attracted as much attention with so little public visibility.
Founded by former OpenAI chief scientist Ilya Sutskever after his departure from OpenAI in 2024, Safe Superintelligence has revealed almost nothing about its technology, roadmap, or commercial plans. Its public website remains little more than a mission statement.
That secrecy has done little to slow investor enthusiasm.
SSI has raised roughly $2 billion and reached an estimated valuation of about $30 billion, despite having no public product, no beta release, and no reported revenue.
When launching the company, Sutskever described its mission in simple terms:
“We will pursue safe superintelligence with a singular focus on one goal and one product.”
The startup appears content to stay focused on research rather than commercialization. That stands apart from many AI companies racing to launch products and generate revenue as quickly as possible.
Building quietly across Silicon Valley and Tel Aviv
SSI has assembled a team filled with respected AI researchers and industry veterans.
The leadership group includes former Apple AI executive Daniel Gross and former OpenAI researcher Daniel Levy. The company has established operations in Silicon Valley and Tel Aviv, where it recently signed office space and recruited machine learning researcher Dr. Yair Carmon from Tel Aviv University. Carmon earned his Ph.D. at Stanford and is widely recognized for his academic work in machine learning.
The company’s first funding round drew backing from Sequoia Capital, Andreessen Horowitz, and DST Global, reflecting investor confidence in Sutskever’s track record long before any product reached the market.
Much of SSI’s public activity has centered on recruiting researchers. Details about its research methods remain tightly guarded.
Sutskever has described the company’s approach as “a new mountain to climb,” though he has offered little insight into how the lab plans to reach that goal.
Nvidia’s investment is about tomorrow, not today
The partnership reflects a broader shift taking place across the AI industry.
Training frontier models now requires billions of dollars in computing infrastructure. Winning the next generation of AI customers means forming relationships years before those customers reach full scale.
Nvidia has become far more than a chip supplier. It is steadily becoming a strategic investor, infrastructure partner, and long-term collaborator for many of the companies shaping advanced AI research.
SSI fits squarely into that strategy.
Its move from Google’s TPU ecosystem to Nvidia GPUs strengthens Nvidia’s position at the center of frontier AI development. It also gives SSI access to the hardware platform that many researchers view as the industry standard for training large-scale AI systems.
The company has yet to reveal what it is building. Investors are betting that Sutskever’s next breakthrough will justify the confidence and billions of dollars flowing into one of the most secretive startups in artificial intelligence.
