Runware, the AI inference startup founded by Romanian developers Flaviu Radulescu and Ioana Hreninciuc, has launched Sonic Inference Pods, a network of modular data centers built for AI inference. Each Pod packs roughly 1,200 GPUs and one megawatt of compute into a standard 20-foot shipping container, arriving fully assembled and tested, needing only ground, power and a network connection, and installing in about a day. The launch occurs amid an infrastructure squeeze: of the roughly 16 gigawatts of US data center capacity announced for 2026, only about 5 gigawatts are currently under construction, and Sightline Climate expects 30 to 50 percent of those projects to be delayed or canceled. Power is the binding constraint; in Northern Virginia, grid connections can take five to seven years, and Dominion Energy has told large customers new connections could stretch as long as seven years. SemiAnalysis argues delay estimates rest on a flawed denominator, but even skeptics agree interconnection timelines, transformer shortages and community opposition are slowing traditional buildout. Each Pod houses around 1,200 GPUs in nodes of two to eight, with Nvidia's RTX PRO 6000 as the workhorse and B200 and B300 chips handling larger workloads, plus local NVMe storage to eliminate cold starts. The Pods use closed-loop liquid cooling recirculating the same 1.5 cubic meters of water continuously, contrasting with an industry drawing over 560 billion liters of water