Thinking Machines Signs $65m Annual Inference Deal with Crusoe
Thinking Machines Lab has signed a $65m annual agreement with Crusoe to run inference for its open models, including GLM 5.2 and 5.3, on a dedicated Nvidia B200 cluster. Crusoe claims this deal pushes their managed inference business past $100m in contracted annual revenue.
Quote
“Crusoe Managed Inference took us from evaluation to production quickly and met the standards we hold our own systems to, giving us the scale and economics to reinvest in the research at the core of what we do,” said Myle Ott, ML infrastructure lead at Thinking Machines Lab.
Collaboration Details
On Wednesday, Thinking Machines Lab and Crusoe announced a $65m annual agreement. Crusoe will handle inference for the lab’s open models on Crusoe Cloud. These include Inkling models and GLM 5.2 and 5.3, along with their fine-tuned variants.
A dedicated deployment of Nvidia HGX B200 systems will power the workloads, connected via Nvidia Quantum-2 InfiniBand networking. This setup constitutes a Tailored Deployment, directly managed and supported by Crusoe with a service-level agreement. This relieves Thinking Machines Lab from managing its own inference stack.
The partnership also explores batch inference for large-scale synthetic data generation.
“Thinking Machines Lab has a very sophisticated engineering team that expects partners to meet their high bar as they grow,” noted Erwan Menard, senior vice president of product management at Crusoe Cloud.
Market Impact
This deal follows a significant month for both companies. Crusoe secured $3.9bn at a $30.9bn valuation last week and signed a $13bn five-year agreement with Jane Street earlier in September. Thinking Machines Lab is reportedly in talks to raise $6bn.
China’s Z.ai, the developer of GLM open models, recently open-sourced its ZCode tool.