US-based energy infrastructure developer Torus has launched Foundry, a modular data centre system engineered for AI inference, while securing $500m in growth capital led by Magnetar.
According to Torus, the prefabricated platform integrates power, cooling, networking, and compute to provide between 1MW and 20MW of capacity per site within 12 to 16 weeks.
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
The financing round included ICONIQ, Origin Ventures, Ridgeline Partners, and Trinity Capital.
Torus will allocate the capital to nationwide deployments in the US and the expansion of its 545,000ft² GigaOne manufacturing plant in Salt Lake City, Utah.
The facility employs nearly 400 staff, with an active production line able to build 500MW of IT load and plans to scale to 1GW.
The launch addresses shifting data centre requirements as AI models transition from training to enterprise inference.
Foundry connects behind existing commercial meters to access available electrical capacity without requiring fresh utility infrastructure.
Torus CEO and founder Nate Walkingshaw said: “The industry has been trying to solve AI’s power problem by building net-new infrastructure, which causes costs to potentially rise and takes years to bring compute online.
“Inference doesn’t have to wait. We’ve spent years learning how to make the grid more affordable, flexible, resilient, and intelligent, and Foundry applies that to compute: capacity in weeks, with power that’s already available. Most providers are bringing power to the chips. Foundry brings the chips to power.”
Each installation pairs with Torus Reserve, utilising flywheel and battery storage to regulate power dispatch and deliver backup electricity.
The modular hardware incorporates waterless closed-loop cooling that consumes 95% less water than evaporative units.
Site preparation takes approximately one week, followed by one to two days for installation and commissioning, with enclosures scaling upwards from the dimensions of a small cargo van.
Foundry is available as a powered shell, a turnkey bare-metal unit with chips, or a direct single-tenant deployment. Launch customers include Hydra Host and an unnamed AI cloud provider.
Headquartered in Salt Lake City, Utah, Torus manufactures distributed power and modular computing infrastructure for data centres, utilities, and commercial and industrial sites.
The company currently operates thousands of graphics processing units and plans to have more than 47,000 graphics processing units (GPUs) running in Foundry units by the end of 2027.
