Just a year after launching, former OpenAI CTO Mira Murati’s AI startup Thinking Machines Lab has secured a massive Nvidia compute partnership to deploy gigawatt-scale AI systems.
Nvidia is making a major public bet on Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati.
Earlier this week, Nvidia announced a strategic investment in the company as part of a multi-year partnership. Under the agreement, Thinking Machines plans to deploy at least one gigawatt of Nvidia’s next-generation Vera Rubin systems.
To most people, “a gigawatt” sounds like a term from the energy sector. In AI, though, it’s shorthand for massive scale — the kind of data-center capacity and chip supply that typically only the largest labs can secure.
That’s why the announcement has drawn so much attention.
Part of how Thinking Machines Lab secured such a large compute commitment comes down to a mix of credibility, capital, and timing.
The one-year-old startup, now valued at over $12 billion, has raised more than $2 billion from investors including Andreessen Horowitz and Accel, giving it the financial backing to deploy infrastructure at scale.
At the same time, Nvidia has been looking to partner closely with the next generation of AI labs that will run its newest chips.
Murati herself is a big part of the story. At OpenAI, she helped lead major technical efforts during the period when AI evolved from a research focus into a global product category. Now she’s attempting to build a new lab from the ground up – under a level of scrutiny few founders face.
What Mira Murati’s AI startup is building
The company’s pitch has been described as less about creating “another chatbot” and more about building AI systems that organizations can rely on.
Thinking Machines has said it aims to develop models that produce reproducible results and can be customized for different applications – language that resonates with businesses looking for consistency rather than unpredictable outputs.
So far, the company has been selective about what it shares publicly. It has launched Tinker, an API and tooling product, but has revealed little about a flagship frontier model or its broader product roadmap.
The team itself has also seen changes. Earlier this year, TechCrunch reported that co-founder and CTO Barret Zoph had departed, with Murati appointing Soumith Chintala as the new CTO.
Other reports have suggested additional co-founders have moved back to OpenAI, underscoring how fluid talent movement can be in the AI industry.
For Nvidia, the investment fits into a wider strategy. In this case, strengthening relationships with the labs most likely to adopt and showcase its next generation of hardware, while expanding the ecosystem of “AI factories” built on its architecture.
For Mira Murati and Thinking Machines, the partnership is effectively a signal that it intends to compete at the very top of the AI market.
The challenge now is turning that headline-level compute commitment under the new Mira Murati Nvidia deal into a reliable engine for delivering models and tools that customers can depend on.
