Runware launches a portable, modular data center pod

AI infrastructure company Runware announced on Tuesday the launch of its own modular data center, called Sonic Inference Pod. Designed as a single transportable unit, the Pod is meant to offer a more flexible kind of compute that can sit alongside hyperscalers’ massive data center projects. Runware says the Pod can offer inference at a higher quality but lower cost than other serverless inference platforms and GPU clouds, and that its modular design makes it easy to add capacity by creating new pods rather than expanding a fixed facility.

“We believe distributed compute, positioned closer to end users for faster inference, is what will win in the long term,” said Flaviu Radulescu, Runware’s co-founder and CEO. He said the pods can scale and add capacity fast, deploy anywhere there is power, and adapt quickly to new hardware releases. The pods also skip water-based cooling for a closed-loop system that can be built in days, compared with the months or years traditional data centers take.

Runware currently has 10 pods deployed across the U.S., Europe, and Asia-Pacific, Radulescu said, already providing inference to a few companies including Higgsfield AI and Wix, with 160 sites available to power additional pods. The company raised a $50 million Series A in December to build infrastructure for companies generating images, and sees the pod expansion as part of its broader mission of providing inference rather than a single product.

Radulescu doesn’t see AI labs’ own data center buildouts — including OpenAI and SpaceX, with OpenAI reportedly close to a $500 billion deal to build a data center in Ohio — as a threat to Sonic Inference Pods, pointing to the pods’ networked flexibility as a differentiator: if one pod goes offline, traffic shifts to another on the same network, while customers wanting dedicated hardware can get whole pods to themselves. “Demand for inference is growing faster than facilities can be built,” Radulescu said. “What we want is to power the world’s intelligence, to be the backbone every AI model runs on with capacity that keeps up with demand instead of throttling it.”

On the resource concerns that have made data center construction controversial — including rising utility costs reported by communities near existing sites — Radulescu said Runware’s approach avoids transmission losses and water use in cooling, and draws on power that already exists rather than requiring new grid capacity, which he said means less new grid and water use for the same compute.

Why it matters

Runware is betting that modular, distributed pods can be deployed faster and cheaper than traditional data centers just as inference demand is outpacing hyperscaler buildouts, positioning the startup as a lower-friction alternative for AI labs racing to secure compute capacity.