🏭Runware Plans Factory Pumping Out 30 Megawatts of Data Centers Daily by 2028

Your GPU capacity could be live in 48 hours

Runware wants to build AI data centers the way Toyota builds cars. The company is setting up a factory to assemble its Sonic Inference Pods on a production line, targeting 30 megawatts of capacity per day by 2028. That works out to nearly 1 gigawatt a month. If you're deploying inference at scale, this changes the math on capacity. Runware says it can bring a pod online within 48 hours, covering installation and setup, not build time. You can supply your own site and GPUs and pay a monthly fee for the pod and operations, or rent compute from them directly. Power and internet are billed separately, so budget for those. The pods are a standard design but can be configured for specific GPUs, and spare parts come with them if components fail. The numbers are big. Runware says it will produce roughly as much in a year as the entire U.S. was expected to add in data center power this year, about 10 gigawatts. It has two factories: one already used for a few years, and a second expected to finish within months and go online early next year. Its site pipeline covers 397 megawatts with potential to reach 1.2 gigawatts. The company has also reported $1 billion in committed interest since offering the pods to outside customers a month ago. The catch is execution. Runware is still on a $50 million Series A and plans a Series B later. Watch whether that factory hits its ramp schedule.