Mira Murati’s AI startup Thinking Machines Lab has locked in a multiyear strategic partnership with Nvidia that includes a significant equity investment and access to at least one gigawatt of the chipmaker’s next-generation Vera Rubin computing systems. The deal, announced on Monday, positions the former OpenAI chief technology officer as one of the biggest buyers of Nvidia hardware in the emerging AI arms race.
The partnership will give Thinking Machines Lab access to billions of dollars’ worth of Nvidia infrastructure, including the company’s latest Rubin GPUs and Vera CPUs. Deployment is expected to begin in early 2027. Industry analysts estimate that building one gigawatt of GPU-powered compute capacity costs roughly $50 billion, with graphics processors accounting for about two-thirds of that figure.
From OpenAI’s CTO to Frontier AI Competitor
Murati co-founded Thinking Machines Lab in February 2025, less than six months after stepping down as CTO of OpenAI, where she oversaw the development of ChatGPT, the Sora video generation model, and other products that helped define the current wave of generative AI. The startup raised $2 billion in seed funding at a $12 billion valuation from a consortium that included Nvidia, AMD, and ServiceNow.
Thinking Machines has already shipped its first product, a cloud service called Tinker that allows developers to create fine-tuned versions of open-source large language models using a technique called LoRA. The platform supports more than a dozen models, including Meta’s Llama series. Job postings indicate the company is expanding into audio processing, visual reasoning, and custom AI architectures.
Nvidia Deepens Its Startup Investment Strategy
For Nvidia, the deal extends a pattern of making strategic investments in the same companies that buy its chips. The company has backed a growing roster of AI startups, including xAI, CoreWeave, and Mistral, creating a cycle in which Nvidia’s capital flows back to Nvidia through hardware purchases.
“AI is the most powerful knowledge discovery instrument in human history,” Nvidia founder and CEO Jensen Huang said in a statement. “Thinking Machines has brought together a world-class team.”
Murati framed the partnership as a way to accelerate development at a time when access to compute is one of the biggest constraints facing AI startups. “Nvidia’s technology is the foundation on which the entire field is built,” she said. “This partnership accelerates our capacity.”
What This Means for the AI Startup Landscape
The deal highlights a growing divide in the AI industry between companies that can secure massive compute commitments and those that cannot. Training frontier models now requires billions of dollars in infrastructure spending, and partnerships like this one effectively serve as barriers to entry for smaller competitors.
It also underscores how quickly the AI startup ecosystem has shifted since Murati’s departure from OpenAI. In just over a year, Thinking Machines Lab has raised billions, built on Nvidia’s GPU platform, launched a commercial product, and now secured one of the largest compute deals in the industry’s history.
The partnership is expected to make Thinking Machines Lab one of the largest single consumers of Nvidia’s forthcoming Vera Rubin architecture when deployments begin next year.
