Runware is building an automotive-style factory in Romania to mass-produce prefabricated AI data center modules, hoping to ramp production to nearly 1 gigawatt per month by 2028 and drastically cut deployment times.
- Runware plans 30 MW production per day by 2028
- Modular pods designed for 48-hour onsite setup
- Factory uses automotive-style assembly line techniques
What happened
Runware unveiled plans to build a factory that mass-produces modular AI data center pods on an assembly line, much like automotive manufacturing. The company already operates one factory and is finalizing a second near-completion, with ambitions to increase production capacity to 30 megawatts per day by 2028. These pods arrive prefabricated and can be configured based on customer needs, including latest GPU technologies, allowing for a quick site installation within 48 hours.
This approach contrasts with traditional data center builds by dramatically shortening delivery and installation times. Runware has positioned itself distinctively in the market by actually constructing the factory, unlike competitors who primarily design prefab systems. The company has raised over $60 million and intends to use its current and forthcoming facilities to supply large-scale AI compute capacity more rapidly than most existing options.
Why it matters
The rapid growth of AI workloads demands unprecedented scale and speed in data center deployment. Runware’s factory model aims to meet this demand by standardizing and automating data center production, reducing cost and lead times. If successful, it could reshape how infrastructure is provisioned for AI, which currently faces bottlenecks like GPU availability and custom build delays.
Learning from challenges faced by startups in adjacent sectors such as electric vehicle manufacturing, Runware is leveraging cost-effective staffing, in-house hardware design, and experienced management from the automotive industry. Its modular pods offer flexibility by adapting to different GPU generations and customer specifications, potentially positioning the startup as a pivotal supplier in the fast-evolving AI compute ecosystem.
What to watch next
Key indicators of Runware’s progress will include the operational launch of its second factory early next year and the ramp-up in daily production capacity. Observers should track whether the company can maintain quality and cost discipline as it scales toward gigawatt-level output monthly. The ability to secure stable supply chains for critical components like pumps and fans sourced from Europe will also be crucial to sustain uninterrupted manufacturing.
Additionally, market adoption of Runware’s prefabricated pods—via direct deployment or compute rental services—will reveal demand dynamics in the AI infrastructure sector. Competitors such as Nvidia with their AI factory projects will also influence how standards evolve for modular, fast-deploy data centers. Runware’s execution in factory efficiency, customer integration, and adapting to GPU supply constraints will ultimately determine its competitive viability.