Runware wants to make modular AI data centers on a factory production line, with the pace and repeatability associated with car manufacturing. The company says its pods can be installed and set up within 48 hours, and it plans to increase production at its Romania-based factory as demand for AI infrastructure grows.
A production line for data center pods
Runware’s Sonic Inference Pod is a modular data center built in a factory and shipped ready to install. The company already has one factory that has been building pods for a few years. It began building a second factory earlier this year and expects to finish it within months, with operations starting early next year.
Runware plans to gradually raise production starting next year. By 2028, it aims to reach 30 megawatts of data center capacity per day, or nearly 1 gigawatt per month. At that pace, the company says its annual output would be roughly as much capacity as the entire U.S. was expected to add in data center power this year: about 10 gigawatts.
Prefabricated data center pods are not unique to Runware. Companies including Crusoe and Schneider Electric also sell them. Runware’s distinction, according to co-founder and CEO Flaviu Radulescu, is its effort to build a factory that assembles the units on a production line.
Short setup times, with options for customers
Customers can provide the site where a pod will be assembled, or Runware can find a location for them. The company’s 48-hour target refers to installation and setup, not the time required to build a pod. Runware says one pod that previously took three weeks to build will soon take one day as production ramps up.
Customers can either rent compute from Runware or provide their own GPUs and pay a monthly fee for the pods and operations. Power and internet are charged separately. Runware also says it will provide spare parts if components fail.
The pods have a standard design but can be configured to customer requirements and to accommodate the latest GPU chips. Runware says its delivery times are shorter than current GPU lead times, giving it time to adapt infrastructure while customers wait for hardware. Its in-house hardware team can also design, build, and test components for new GPU generations.
Scaling brings manufacturing and supply challenges
The factory plan remains a test of execution. It will require staff, inventory, supply chain logistics, and funding. The source points to electric vehicle startup Arrival as a cautionary example: once valued at over $13 billion, it struggled to mass-produce vehicles cost-effectively compared with traditional plants.
Runware has raised more than $60 million to date. It says it plans to raise a Series B, but is finishing and launching its new factory with the $50 million Series A it raised last December. Radulescu said the company’s cost structure differs from Arrival’s, citing the cost of factory staff relative to the value of deploying AI compute, the use of widely available and affordable parts, and its own system designs. Runware has also brought in a plant manager from the automotive industry to work with its chief of hardware.
Some components still depend on outside suppliers. Runware says it designs and builds its own cooling and other pod systems using standard components, but some parts, including pumps and fans, come from Germany. It also depends on Nvidia for networking.
Energy, sites, and the next phase
Runware says it has received $1 billion in committed interest for its Sonic pods since it began offering them to outside customers a month ago. Radulescu said that interest represents over 20 megawatts that the company is working to deliver.
Data centers face scrutiny over their energy use and effects on grids and other resources. Runware says its pods need power and an internet connection, and use closed-loop liquid cooling with air-cooled radiators to release heat, so they do not require a water connection. The company says it is working with power providers on a site pipeline covering 397 megawatts, with potential to expand to 1.2 gigawatts.
Radulescu said the sites typically already have power and are much smaller than those used by Amazon, Microsoft, and Google. He also said the pods run “completely silent up to temperatures of 86 degrees Fahrenheit” and that Runware uses renewable energy whenever possible, with the aim of making renewables its main energy source.
After bringing the new factory online early next year, Runware plans to expand to the U.S. and other regions worldwide. Its larger claim is that factory-built modular systems could let customers plan data center deployments from the beginning and have them working in months instead of years. Whether that promise holds will depend on how quickly the company can bring its factories, supply chain, and customer sites together.