[Runware Plans to Mass-Produce Modular AI Data Centers on an Assembly Line, Aiming for 30 Megawatts Per Day by 2028] According to a TechCrunch report on October 30th, AI infrastructure company Runware plans to mass-produce its Sonic Inference Pod modular data centers on an assembly line, much like manufacturing cars. Co-founder and CEO Flaviu Radulescu stated that the Romanian factory's capacity will increase to 30 megawatts per day by 2028, equivalent to nearly 1 gigawatt per month. At this rate, Runware's annual output would roughly equate to the estimated new data center power capacity added in the US this year, approximately 10 gigawatts. The company already has two factories, with the second expected to be completed within months and operational by early next year. Runware claims that the manufacturing time for a single module will be reduced from three weeks to one day, and installation and deployment can be completed within 48 hours of arrival at the customer's site. Customers can rent computing power or bring their own GPUs and pay monthly for modules and operations, with electricity and network costs extra. The company stated that since launching its external sales, it has received $1 billion in written commitments, covering capacity exceeding 20 megawatts. Runware has raised over $60 million in total funding and is currently using its $50 million Series A funding round in December of last year to advance its new factory, with plans to raise a Series B round. TechCrunch points out that these are still just plans, and whether mass production can be achieved depends on funding, supply chain, and manufacturing execution. Electric vehicle company Arrival was once valued at over $13 billion but later struggled due to its inability to achieve low-cost mass production. Furthermore, some of Runware's pumps and fans are sourced from Germany, and its network equipment still relies on Nvidia.