Google Cloud announced a new agreement to provide expanded AI infrastructure capabilities to Thinking Machines Lab, according to a company statement. The agreement will increase Thinking Machines' use of Google Cloud services for research, platform development, and model training.

Under the agreement, Thinking Machines will use A4X Max virtual machines with NVIDIA GB300 GPUs through Google Cloud's AI Hypercomputer platform. The company will be among the first Google Cloud customers to utilize NVIDIA GB300 NVL72 technology.

Thinking Machines reported a 2X increase in training and serving speed with A4X Max VMs compared to previous generation GPUs during early testing. The company uses Google Cloud services including Google Kubernetes Engine, Spanner, Cluster Director, Cloud Storage, and Anywhere Cache for its frontier models and fine-tuning product called Tinker.

"By leveraging A4X Max and the AI Hypercomputer integrated stack, Google Cloud got us running at record speed with the reliability we demand," said Myle Ott, Founding Researcher at Thinking Machines Lab.

Mark Lohmeyer, VP and GM of AI and Computing Infrastructure at Google Cloud, stated that the partnership with NVIDIA will help Thinking Machines "accelerate even further, using Google Cloud's AI Hypercomputer which brings together purpose-built hardware, open software and flexible consumption models."

Ian Buck, Vice President and General Manager of Hyperscale and HPC at NVIDIA, noted that "NVIDIA GB300 NVL72 provides the performance leap and interconnect bandwidth needed to reduce bottlenecks and improve goodput."

Thinking Machines began its partnership with Google Cloud in 2025.