AI startup Thinking Machines clinches capital and a major chip supply deal from Nvidia

AI startup Thinking Machines Lab said on Tuesday it has struck a multi-year ​partnership with Nvidia that will see it receive a significant investment ‌and procure at least one gigawatt of ​the chipmaker’s next-generation processors.

Financial terms of the deal ⁠were not disclosed.

Under the agreement, Thinking Machines, 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 power will primarily be ‌used to train the startup’s artificial intelligence models.

Industry executives ‌have ⁠said 1 gigawatt of computing power, enough to ⁠power roughly 750,000 U.S. homes, can cost around $50 billion.

The deal will help Thinking Machines compete with larger rivals in building powerful AI ​systems, and underscores the industry’s ‌eagerness to scale computing capacity.

Thinking Machines quickly became one of Silicon Valley’s most closely watched AI startups after raising about $2 billion in a seed funding round ‌led by Andreessen Horowitz that valued the company ​at $12 billion. Nvidia was also an investor in the round.

The startup has recently been seeking ⁠to raise more in a new funding round that could value it at tens of billions of dollars, sources ‌told Reuters earlier.

The company has recently seen several departures, including co-founder and former Chief Technology Officer Barret Zoph and co-founder Luke Metz, who both returned to their former employer OpenAI amid fierce competition for AI talent.

The partnership also highlights Nvidia’s growing role as ‌a financier of the startups that rely on its AI chips.

It ​has made a recent $30 billion investment in OpenAI and invested $10 billion in Anthropic, while also ⁠supplying the graphics processing units (GPUs) used to train and run ⁠their models, a dynamic that some industry analysts say creates a circular flow of capital and computing ‌resources. That in turn has given rise to comparisons with the late 1990s tech bubble.