Crusoe and Thinking Machines Lab have announced a $65 million annual agreement to power inference for the lab's open-model workloads. The partnership, announced on September 23, 2026, places Crusoe's managed inference service at the center of the lab's production operations.
Crusoe Managed Inference is described as being tuned for throughput, price-performance, and reliability, which aligns with the demands of running large-scale open-model inference. The agreement gives Thinking Machines Lab dedicated capacity for its production workloads rather than ad-hoc or spot infrastructure.
With only this single source, there are no independent accounts to compare against. The announcement frames the deal as a strategic commitment from both companies, but does not disclose the contract length beyond the annual value or specific model deployment details.