Former OpenAI CTO Mira Murati's Thinking Machines Lab has secured a multibillion-dollar chip supply agreement with Nvidia, deploying at least one gigawatt of next-generation Vera Rubin processors. The San Francisco-based startup announced the multiyear partnership Tuesday, marking one of Nvidia's largest customer investments in the AI infrastructure space.

The deal includes a significant additional investment from Nvidia, which previously backed Thinking Machines' $2 billion seed round that valued the company at $10 billion. This arrangement follows Nvidia's pattern of investing in major AI chip consumers, though such circular financing arrangements have raised industry concerns about the sustainability of these funding models.

Thinking Machines, founded by Murati a year ago, develops Tinker, an enterprise platform that lets businesses customize large language models without managing complex AI infrastructure. The partnership will help Thinking Machines design training systems for Nvidia architectures and expand access to frontier AI models. This deal comes amid recent executive departures from the one-year-old startup, including three co-founders who returned to OpenAI.