The Pulse
Torus raises $500 million and launches modular AI compute system
Torus has raised $500 million in growth capital led by Magnetar and launched Foundry, a modular system for AI inference. The company plans to use the funding to expand manufacturing and deploy Foundry and its energy-storage product across the United States.

Torus has raised $500 million in growth capital and launched Foundry, a prefabricated system for running AI inference at existing sites. The company said Foundry units can deliver 1–20 megawatts per site and take 12–16 weeks from contract to running AI workloads.
Funding for production and deployments
Magnetar led the growth round, with participation from ICONIQ, Origin Ventures, Ridgeline Partners and Trinity Capital. The money is intended to fund continued expansion of GigaOne, the company’s manufacturing facility, and deployments of Foundry and its Reserve energy-storage product across the United States.
Using power at existing sites
Foundry combines power, storage, cooling, networking and compute in a modular system. Torus says the system is installed behind the meter and uses available electricity at individual transformers, without requiring new grid infrastructure.
Every deployment is paired with Torus Reserve, an energy-storage and management product. Reserve uses predefined controls to govern how power is dispatched, including controls for load curtailment and changes in the rate at which power is supplied.
According to Torus, integrated flywheel and battery storage covers demand peaks so more computing capacity can fit on an existing grid interconnection. It also provides backup power during grid outages.
Torus Protect handles integrated networking and security, while the company provides ongoing operations and maintenance. The smallest system is roughly the size of a small cargo van. Torus estimates that site preparation takes about a week, followed by one to two days for installation and commissioning.
Customers and product options
Foundry is launching with Hydra Host and another AI cloud provider as its first customers. The company says Foundry has thousands of GPUs deployed today. It plans to have more than 47,000 GPUs deployed in Foundry units by the end of 2027.
The product is available in three models. The strategic model provides power and an enclosure without GPUs, while the partner model includes power, an enclosure and GPUs. The direct model offers inference capacity to enterprise customers on hardware allocated exclusively to one company’s workload.
Manufacturing capacity
Foundry units are made at GigaOne in Salt Lake City. The facility spans 545,000 square feet, and key components are built in-house.
The company says GigaOne opened its first production line in early 2026 and has opened a new line every 21 days in 2026. The Foundry line is online and will be capable of producing 500 megawatts of IT load, with an expansion to 1 gigawatt planned in the near future.
Torus also says its closed-loop cooling system is waterless and uses 95 percent less water than evaporative coolers. Each system also donates GPU time to medical research through Folding@home.