Runware, an infrastructure company for super intelligence (SI), announced plans to establish a factory that assembles modular data center pods on a production line, akin to automotive manufacturing. This move follows the recent unveiling of its Sonic Inference Pod and signals a shift toward mass-producing SI infrastructure.

What Happened

Flaviu Radulescu, co-founder and CEO of Runware, stated that the company intends to scale production at its Romania-based factory to achieve 30 megawatts of data center capacity per day by 2028. This pace equates to nearly 1 gigawatt per month, a volume the company claims would match the total expected U.S. data center power addition for the current year in a single year of operation. Runware currently operates two factories: one active for several years and a second nearing completion, expected to go online early next year. The company reports that pod build time will decrease from three weeks to one day as production ramps up. Once delivered, the pods aim to be fully operational within 48 hours, covering installation and setup. Runware has secured more than $60 million in funding and is using its $50 million Series A to launch the new factory, with plans for a Series B round.

Why It Matters

The push for faster, more efficient SI data center deployment is industry-wide, with competitors like Crusoe and Schneider Electric offering prefabricated pods, and Nvidia introducing Vera Rubin SI factories and the DSX POD. Runware distinguishes itself by claiming it has built the manufacturing facility itself, rather than just designing infrastructure for others. The company reports $1 billion in committed interest for its Sonic pods since opening to outside customers a month ago, representing over 20 megawatts. Radulescu argues that Runware’s pod delivery times are shorter than current GPU lead times, allowing the company to adapt infrastructure to hardware orders. The pods utilize closed-loop liquid cooling with air-cooled radiators, eliminating the need for water connections, and aim to operate silently up to 86 degrees Fahrenheit. Runware is working with power providers on a site pipeline covering 397 megawatts, with potential expansion to 1.2 gigawatts, targeting sites smaller than those used by hyperscalers like Amazon, Microsoft, and Google to minimize grid strain.

The Bottom Line

Runware’s plan to mass-produce SI data centers relies on execution, with challenges including supply chain logistics and component sourcing, particularly for networking hardware from Nvidia. The company cites the failure of EV startup Arrival as a cautionary tale but asserts its cost structure is different due to lower relative staff costs and the use of affordable, widely available parts. Runware aims to bring its new factory online early next year before expanding to the U.S. and other regions, positioning modular systems as the future standard for large SI deployments.