Thinking Machines Lab said it has signed a multi-year partnership with Nvidia that includes both a material investment by the chipmaker and a commitment to supply at least one gigawatt of Nvidia's forthcoming processors. The financial specifics of the arrangement were not disclosed.
Under the agreement, Thinking Machines - which was founded last year by former OpenAI Chief Technology Officer Mira Murati - will deploy Nvidia's upcoming Vera Rubin systems starting early next year. The computing capacity provided through those systems will be used primarily to train the startup's artificial intelligence models.
Industry executives quoted in prior reporting have estimated that one gigawatt of computing power - an amount of capacity large enough to power roughly 750,000 U.S. homes - can cost around $50 billion. That scale of compute is widely viewed as a competitive asset for firms aiming to build larger AI systems, and the deal reinforces the broader industry's push to rapidly expand computing resources.
Thinking Machines has been among the most closely watched AI newcomers in Silicon Valley since completing a roughly $2 billion seed funding round led by Andreessen Horowitz that valued the company at $12 billion. Nvidia was also an investor in that seed round.
The startup has been seeking additional capital in a new funding round that, according to prior reports, could value the company at tens of billions of dollars. At the same time, Thinking Machines has experienced several departures: co-founder and former Chief Technology Officer Barret Zoph and co-founder Luke Metz have both left the company and returned to their former employer OpenAI.
The partnership with Thinking Machines also highlights Nvidia's increasing role as a financier for firms that rely on its AI chips. Nvidia has made a recent $30 billion investment in OpenAI and invested $10 billion in Anthropic, while also supplying the graphics processing units used to train and run their models. Some industry analysts say this combination of capital and chip supply creates a circular flow of funding and computing resources, a dynamic that has produced comparisons to the late 1990s tech bubble.
Separately, a platform called ProPicks AI evaluates Nvidia alongside thousands of other companies each month using more than 100 financial metrics. The platform uses AI to screen for stocks it considers to offer favorable risk-reward profiles based on current data, and has cited past winners including Super Micro Computer (+185%) and AppLovin (+157%). The platform states it flags which stocks are featured in its strategies, including any current coverage of Nvidia.
What this means
- The agreement secures large-scale compute capacity for Thinking Machines to train advanced AI models.
- Nvidia continues to expand its role as both supplier and investor within the AI startup ecosystem.
- The move supports Thinking Machines' efforts to compete with larger firms while it seeks further funding.