Google Deepens Ties With Mira Murati's Thinking Machines in New Multi-Billion-Dollar Deal
Google Cloud locks in another fast-rising frontier lab with a single-digit-billion-dollar agreement covering Nvidia GB300 GPUs and reinforcement-learning workloads.
Google has signed a new multi-billion-dollar agreement with Thinking Machines Lab, the frontier AI startup founded by former OpenAI CTO Mira Murati. The deal, reported on April 22, expands Thinking Machines' access to Google Cloud's latest AI infrastructure and bundles it with Google's broader storage, Kubernetes and database services.
The contract is in the single-digit billions of dollars and is structured as a non-exclusive arrangement, leaving Thinking Machines free to work with other clouds. At its core, it gives the lab room on Google Cloud systems powered by Nvidia's new GB300 chips, which Google says deliver roughly a 2x improvement in training and serving speed over the previous generation — a meaningful edge for reinforcement-learning workloads.
Founded in February 2025, Thinking Machines has moved quickly to establish itself among the new wave of frontier labs. Its first product, Tinker, launched in October 2024 and automates the creation of custom AI models. The company has emphasized reinforcement learning as a core research bet, an approach that is compute-hungry in ways that make cloud relationships strategically important.
"Google Cloud got us running at record speed with the reliability we demand," Myle Ott, founding researcher at Thinking Machines, said of the partnership.
The agreement is the latest in a flurry of cloud-meets-lab tie-ups. Anthropic recently secured multi-gigawatt commitments from both Google and Broadcom on the TPU side, plus a five-gigawatt deal with Amazon for Claude training and deployment. By locking in Thinking Machines early, Google further hedges its frontier exposure beyond its in-house Gemini program — and signals that the era of clouds quietly hosting AI labs has given way to one of openly bidding for them.
