Runware plans assembly-line factory to build container-sized AI inference data centers

Runware, an AI inference company that unveiled a containerized data center in August, now wants to produce those units on an assembly line. TechCrunch reported on October 30 that the startup is preparing a factory to turn out its Sonic Inference Pods at a pace closer to car manufacturing than to conventional data center construction.

What the factory is meant to do

According to TechCrunch, Runware has used one factory for a few years to build pods and started a second one earlier this year. That new facility is expected to be finished within months and to come online in early 2027, with output ramping up gradually from next year. The company's stated goal is 30 megawatts of data center capacity per day by 2028, from a Romania-based factory, before expanding to the United States and other regions. To run the line, Runware says it has hired a plant manager from the automotive industry.

Building a pod once took three weeks; the company says this will shrink to one day as production scales. Separately, it aims to have pods running within 48 hours of delivery, a figure that covers installation and setup only. Runware says the Series A money is funding the factory and that it plans to raise a Series B. It has raised more than $60 million in total, including a $50 million Series A in December 2025.

The product and the demand claims

Each pod packs 1 megawatt of compute into a 20-foot shipping container with closed-loop liquid cooling, according to Runware's August announcement; its blog put capacity at roughly 1,200 GPUs. Customers can rent compute from Runware or bring their own GPUs and pay a monthly fee, with power and connectivity billed separately.

Runware told TechCrunch it has received $1 billion in committed interest since offering pods to outside customers about a month ago, representing more than 20 megawatts. It also says it is negotiating a site pipeline of 397 megawatts with power providers that could grow to 1.2 gigawatts.

Caveats

These are plans, not delivered capacity. TechCrunch notes that a factory at this scale needs staff, inventory, supply chains and capital, and points to Arrival, an electric-vehicle startup once valued above $13 billion that struggled to mass-produce cost-effectively. Runware still depends on Nvidia for networking and imports some parts, such as pumps and fans, from Germany. Rivals selling prefabricated units include Crusoe and Schneider Electric.