08/04/2026 | Press release | Distributed by Public on 08/04/2026 17:12
Runware just put a data center in a 20-foot shipping container.Demand for AI inference is growing faster than data centers can be built and it bites harder when a fixed site takes months or years to build. Runware's new Sonic Inference Pod compresses deployment from months to days. Inside the container are more than 1,200 GPUs, liquid-cooled and powered by roughly 1 MW of infrastructure. The pods operate as part of one network, routing inference requests to available capacity. If one pod goes offline, the rest keep serving traffic.10 pods are live across the US, Europe and Asia-Pacific, with Wix and Higgsfield AI among the customers. The target Runware set for this year is 10,000 GPU nodes online, one container at a time.
Read more in TechCrunch