NVIDIA is strengthening its position in the global AI computing landscape by betting on one of the most secretive startups in the field.
The chipmaker announced a "significant" investment in Safe Superintelligence (SSI), an AI laboratory founded by former OpenAI chief scientist Ilya Sutskever, alongside a long-term partnership agreement.
Under the terms of the deal, SSI will gain access to a large supply of NVIDIA's flagship GPUs, with the potential to boost its computing power by "an order of magnitude." Previously, the company primarily relied on Alphabet's tensor processing units (TPUs).
This investment not only provides crucial computing support for SSI but also marks another win for NVIDIA in its strategy to secure top-tier clients during the AI boom. For the market, this partnership means SSI's research pace could accelerate significantly, while NVIDIA further extends its reach into the most cutting-edge and closely watched AI research institutions.
NVIDIA Bets on Another OpenAI Alumnus
NVIDIA decided to make this "significant" investment after receiving a rare internal briefing on SSI's research progress, though the exact financial terms were not disclosed.
This is not NVIDIA's first move of this kind. In March, NVIDIA reached a similar investment agreement with Mira Murati, another former OpenAI executive and former chief technology officer. Murati co-founded Thinking Machines Lab last year, which earlier this month released its first AI model trained on NVIDIA hardware.
Both deals follow the same logic: NVIDIA uses strategic investments to lock in star entrepreneurs, ensuring high-profile clients prioritize its chips in the AI arms race while applying pressure on its competitors.
SSI: A $30 Billion Valuation Mystery Lab
SSI was co-founded by Sutskever in 2024 with a highly focused mission. The company has stated publicly it has only one goal: to pursue "safe superintelligence" in a "straight-line sprint" – an AI system capable of performing all human tasks without posing a threat to humanity.
Leveraging Sutskever's industry reputation, SSI quickly raised approximately $2 billion from top venture capital firms like Andreessen Horowitz and Sequoia Capital, achieving a valuation of around $30 billion last year. Despite this, the company remains highly secretive, with its research findings almost never publicly disclosed.
In a statement, Sutskever said the company's research "focuses on neglected aspects of human brain function." The partnership with NVIDIA is one of the most substantive public developments for SSI to date.
Sutskever: From AI Revolution Founder to Independent Explorer
Born in the former Soviet Union and raised in Israel, Sutskever is a mathematical prodigy. During his graduate studies in Canada, he studied under Nobel laureate and AI pioneer Geoffrey Hinton. In 2012, he co-authored a landmark paper in computer vision with Hinton, which demonstrated that scaling deep neural networks with massive data and computing power could lead to measurable gains in intelligence, effectively igniting the modern AI race.
Sutskever subsequently joined Google and became a co-founder of OpenAI in 2015. During his time at OpenAI, his research insights directly contributed to the creation of ChatGPT, and he gained fame for his belief that "scaling up training data and compute power will lead to artificial general intelligence (AGI)."
However, his relationship with OpenAI CEO Sam Altman ended dramatically. He voted to fire Altman and left the company months after Altman was reinstated. In rare public comments after leaving OpenAI, Sutskever began to question whether simply scaling up the current paradigm would be enough to achieve a true breakthrough.
The Logic of the Chip Giant Behind the Computing Power Battle
The deeper context for this partnership is the increasingly competitive landscape of the AI chip market. SSI's shift from primarily using Google TPUs to NVIDIA GPUs is a symbolic win for NVIDIA. It represents not just a return on investment but also the acquisition of a highly influential client in the battle for computing power against competitors like Google.
For SSI, an increase in computing resources "by an order of magnitude" means a qualitative leap in its research capabilities, potentially accelerating its pursuit of safe superintelligence. In the current AI race, the ability to secure computing power often directly determines the speed and depth of research progress.
Comments