Crusoe and Thinking Machines Lab Signed Inference Deal
The partnership includes a $65 million annual commitment to run open models on Nvidia hardware.
Updated on Sept. 23, 2026 in Artificial Intelligence

Live Poll
Is now a good time to invest in companies heavily focused on artificial intelligence infrastructure?
Crusoe has secured a $65 million annual inference agreement with Thinking Machines Lab to power open AI models. This deal pushes Crusoe's contracted annual recurring revenue for its inference business to $100 million.
Why it matters
Thinking Machines Lab entered this partnership to gain the scale and economic efficiency required to reinvest in ongoing AI research. For Crusoe, the deal reinforces its position as a specialized provider for high-demand, tailored AI deployments.
The agreement utilizes dedicated Nvidia HGX B200 systems integrated with Nvidia Quantum-2 InfiniBand networking. These workloads are managed by Crusoe as a Tailored Deployment, supporting open models such as Inkling, GLM 5.2, and GLM 5.3.
The players
Crusoe
This technology infrastructure firm provides specialized cloud services and data center solutions.
Thinking Machines Lab
This research organization focuses on the development and training of open AI models.
Nvidia
This technology company designs the graphics processing units and networking hardware used for AI workloads.
The details
Thinking Machines Lab will leverage Crusoe Cloud to deploy its latest open models on optimized hardware clusters. The two companies are also actively evaluating future capabilities, specifically the use of batch inference to accelerate synthetic data generation.
Timeline
September 23, 2026: The deal between Crusoe and Thinking Machines Lab was announced.
September 2026: Crusoe raised $3.9 billion and finalized a $13 billion deal with Jane Street.
The Tech Race
This deal underscores the aggressive pursuit of high-performance compute clusters by AI developers in an increasingly resource-constrained market. The adoption of the Nvidia HGX B200 architecture marks a clear shift toward specialized hardware deployments designed specifically to optimize inference efficiency.
This partnership improves the accessibility of high-powered computing resources needed to train and run complex open-source AI models. Developers using these systems can expect faster inference times and more efficient resource scaling for their machine learning projects.
The takeaway
The growth of Crusoe's inference business highlights the rising market demand for specialized AI infrastructure services. As labs focus on scaling, partnerships that bundle hardware with management services are becoming standard in the AI ecosystem.
Further reading
For broader trends in infrastructure supporting modern AI, visit the Artificial Intelligence section.
Live Poll
Is now a good time to invest in companies heavily focused on artificial intelligence infrastructure?






