Crusoe and Perplexity sign multi-year AI cloud partnership

Crusoe and Perplexity have signed a multi-year partnership spanning Perplexity’s model training and production inference on Crusoe Cloud, as well as Crusoe’s adoption of Perplexity software.

Crusoe said Perplexity will train frontier models on dedicated clusters powered by NVIDIA GB300 NVL72 systems and NVIDIA InfiniBand networking. For production workloads, Perplexity will use Crusoe’s Managed Inference service, taking the arrangement from model development into the computing required to answer live user requests.

The software commitment runs in the other direction. Crusoe said it will adopt Perplexity Enterprise Pro and Max for its 1,800 employees, with plans to use the products for web and internal knowledge search, multi-step research, data analysis and access to frontier AI models.

Perplexity co-founder and CEO Aravind Srinivas said, “Running AI at Perplexity’s scale means every millisecond of latency is felt by users.”

Crusoe’s announcement did not disclose the contract value, GPU count, allocated compute or power capacity, expected workload volume, or a deployment timetable.