Nvidia has announced a landmark partnership with Thinking Machines Lab, the AI research company founded by former OpenAI Chief Technology Officer Mira Murati. The deal includes a significant investment from Nvidia and a commitment to provide an unprecedented amount of compute power.
A Gigawatt-Scale Commitment
The multi-year strategic partnership, announced on March 10, 2026, will see Thinking Machines Lab deploy at least one gigawatt of Nvidia’s next-generation Vera Rubin GPU systems. To put this in perspective, a gigawatt of compute capacity is equivalent to the power consumption of a small city — representing one of the largest AI infrastructure commitments ever made.
The Vera Rubin systems, which were released earlier this year, will begin deployment in early 2027. The size of the financial deal has not been disclosed, but it represents a substantial vote of confidence in Thinking Machines Lab’s approach to building advanced AI systems.
Thinking Machines Lab’s Vision
Founded by Murati after her departure from OpenAI in 2024, Thinking Machines Lab has positioned itself as a different kind of AI company. The startup aims to create highly customizable AI systems that users can adapt to their specific needs — a vision that emphasizes user control and flexibility over the closed approaches taken by some competitors.
The partnership with Nvidia gives Thinking Machines Lab access to the most advanced AI training infrastructure available, potentially allowing it to compete with well-funded rivals like OpenAI, Anthropic, and Google DeepMind.
What This Means for the AI Industry
The deal signals several important trends:
Hardware is still king: Despite growing interest in alternative computing approaches, Nvidia continues to dominate the AI infrastructure landscape. Major AI players are locked in a race to secure GPU supply, and partnerships like this ensure access to cutting-edge systems.
Customization is the next frontier: Thinking Machines Lab’s focus on user-adaptable AI suggests that the industry may be moving toward more customizable models rather than one-size-fits-all solutions.
The startup ecosystem is maturing: For a relatively new company to secure such a substantial partnership speaks to the maturation of the AI startup ecosystem and the continued willingness of incumbents to invest in innovative newcomers.
The announcement comes just ahead of Nvidia’s annual GTC developer conference, where CEO Jensen Huang is expected to further elaborate on the company’s AI infrastructure strategy.