• Runware's 30 MW-per-day target for 2028 is a stated goal from its CEO, not output. No independent check of it was found.
• The reported $1 billion in committed interest covers over 20 MW of deliverable capacity, and its terms are undisclosed.
• A 48-hour deployment window covers installation and set-up only, not the build.
• The factory is financed by a $50 million Series A, and the company says it plans a Series B.
• The 2024 failure of EV maker Arrival is the comparison point for the execution risk.

Runware says it can build AI data centers the way carmakers build cars, and it has put numbers on the claim. According to TechCrunch's October 30, 2026 report, the company aims to scale a Romania-based factory to 30 megawatts of capacity per day by 2028. Those figures come from CEO Flaviu Radulescu. Nothing in the reporting shows output at that scale, so the numbers are targets to be measured against delivered pods.

What Runware is actually selling

Runware's Sonic Inference Pod is a modular data center built in a factory and shipped ready to install, per TechCrunch. Runware's August 4, 2026 announcement describes a 1 MW unit in a 20-foot shipping container, with up to 1,200 liquid-cooled GPUs and a build time of about three weeks. Customers can supply their own site and GPUs and pay a monthly fee, or rent compute from Runware. Power and internet are charged separately.

The scale targets are stated, not shown

Radulescu told TechCrunch that Runware targets 30 megawatts of capacity per day by 2028, or nearly 1 gigawatt per month. The company says a pod that once took three weeks to build will soon take one day. TechCrunch notes that this pace would roughly match the 10 gigawatts of U.S. data center power expected to be added this year. That comparison is the reporter's, and the production ramp is the company's own projection.

Demand figures come without terms

Runware claims $1 billion in committed interest since it began offering pods to outside customers a month before the article ran. TechCrunch quotes Radulescu saying that this represents over 20 megawatts the company is working to deliver. It also reports a 397-megawatt site pipeline. Our arithmetic puts $1 billion against 20 megawatts at roughly $50 million per megawatt, which suggests the figure covers more than the pod hardware. Contract terms are undisclosed, and 'interest' is not a signed sale.

Funding is thin against the ambition

TechCrunch reports that Runware has raised more than $60 million and is finishing the factory with a $50 million Series A raised last December. The company says it plans a Series B. A second factory, under construction since earlier this year, is expected to go online early next year, followed by expansion to the U.S. and other regions. Capital need at the stated scale is not disclosed.

Where the execution risk sits

TechCrunch points to Arrival, the EV startup once valued above $13 billion that struggled to mass-produce vehicles cost-effectively. Radulescu said Runware's cost structure is different and that it has hired an experienced automotive plant manager. The pods still depend on Nvidia for networking, and some parts, such as pumps and fans, come from Germany. Crusoe and Schneider Electric also sell prefabricated pods, so speed is the claimed differentiator.

Verdict: a credible concept with an unproven ramp

The concept is plausible, and Runware already has one factory that has built pods for a few years, per TechCrunch. But the evidence for the headline numbers is the company's own. Until the second factory runs and delivered megawatts are verified by customers or filings, treat 30 megawatts per day as a target and the $1 billion as unconfirmed interest. The next test is the early-next-year factory start and the first customer deliveries.