Exclusive: Google deepens Thinking Machines Lab ties with new multibillion-dollar deal
New Strategic Partnership for AI Infrastructure
Thinking Machines Lab, led by ex-OpenAI executive Mira Murati, has signed a significant multibillion-dollar agreement with Google Cloud. This partnership will allow the AI startup to use Google’s powerful infrastructure, which includes systems based on Nvidia’s state-of-the-art GB300 GPUs, for developing and deploying their AI models. The deal is reportedly valued in the single-digit billion range.
Pushing the Boundaries of AI Research
Google is expanding its cloud services for leading AI companies, bundling AI computing with products like managed databases and Kubernetes engines. Thinking Machines joins other top AI firms like Anthropic, which recently secured major deals with both Google and Amazon for access to large-scale processing power.
Earlier this year, Thinking Machines partnered with Nvidia, but this is the company’s first major agreement with a cloud provider. While the deal with Google is not exclusive, it signals Google’s desire to support cutting-edge research teams as AI development accelerates.
Focus on Advanced AI Model Training
The agreement covers resources for training AI models using reinforcement learning—a technique crucial to many AI breakthroughs at companies like DeepMind and OpenAI. According to Google, the GB300-powered infrastructure offers twice the speed for training and serving models compared to previous GPU generations. This positions Thinking Machines among the first to benefit from Google’s most advanced AI systems.
Thinking Machines Lab’s main product, Tinker, helps automate building custom AI models, and the new partnership enables greater experimentation and scaling. "Google Cloud got us running at record speed with the reliability we demand," said Myle Ott, a founding researcher at Thinking Machines.
About Thinking Machines Lab
Founded in early 2025, Thinking Machines Lab quickly raised a $2 billion seed round at a $12 billion valuation. The startup has remained relatively secretive, but is rapidly positioning itself at the forefront of AI model innovation.
For more details, read the original article by Rebecca Bellan on TechCrunch.
