Torus Launches Modular AI Infrastructure, Foundry, to Meet Surging AI Inference Demand and Secures $500M in Funding

Foundry brings 1–20 megawatts of inference capacity per site in 12–16 weeks from contract to compute, using power the grid already has

New funding will support expanded manufacturing capacity and accelerate Foundry deployments

SALT LAKE CITY, UT, November 11, 2026 – Torus, a U.S. manufacturer of distributed power and compute infrastructure, today launched Foundry, modular AI infrastructure designed to meet the fast-growing demand for inference compute. Torus Foundry packages power, storage, cooling, networking, and compute into a prefabricated system that delivers new capacity to existing sites in weeks, instead of the years it takes to build a traditional hyperscale campus. Additionally, Torus announced $500 million in growth capital led by Magnetar.

AI demand is shifting from training models to running inference models for enterprises. Data center demand for AI inference is projected to grow 35 percent a year through 2030, compared with 22 percent for training, according to McKinsey.

Securing the power to meet AI demand is the biggest bottleneck today, and it is one Torus has spent years solving. That bottleneck is constrained power and capacity, not building new generation. Torus’s technology unlocks unused grid capacity that sits between substations and existing commercial and industrial buildings by using existing interconnections. Every Foundry is paired with Torus Reserve, the company’s energy storage and management product. Reserve provides grid-enhancing technologies, including inertial-based firm power, grid flexibility with load curtailment capabilities, and ramp-rate support, all delivered through load-shaping recipes: predefined controls that govern how power is dispatched. These capabilities are important for AI compute infrastructure in shaping loads to manage demand, supporting grid stability, and ramping power in milliseconds. The result is distributed edge compute for the enterprise that takes advantage of existing grid interconnections and is online in 12–16 weeks from contract to token.

“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,” said Nate Walkingshaw, CEO and Founder of Torus. “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.”

The result reflects a simple premise behind Torus's work: much of America's energy challenge is a capacity problem before it is a generation problem. The grid has unused capacity at many sites today, and Torus systems make that capacity usable.

Inference compute works differently. It runs continuously, serves users in real time, and benefits from being distributed closer to where demand is. This difference is pulling more compute toward on-premises and edge sites, and inference providers need capacity they can deploy, in more places, sooner. Foundry is built for that: distributed, smaller sites, deployed faster, on power that is already available. Enterprise customers want security, IP protection, and control over where their workloads run. Foundry’s design gives them that, in ways centralized capacity can be harder to match.

Foundry at-a-glance

• From contract to running AI workloads in 12–16 weeks. Foundry is prefabricated and modular, at 1–20 MW per site, with no building construction required. The smallest system is roughly the size of a small cargo van.
• No new grid infrastructure. Foundry is installed behind the meter, using available headroom, or available electricity, at individual transformers. It adds steady, high-load-factor demand without requiring infrastructure upgrades that often lead to rate increases, or affecting grid reliability.
• Full stack, one provider. Fiber to site, chiller to coolant distribution unit, grid to rack: Torus vertically integrates the entire stack, from Torus Reserve for load shaping and redundancy, to Torus Foundry for inference compute, to Torus Protect for integrated networking and security. Every deployment is turnkey, with ongoing operations and maintenance provided by Torus.
• Fastest path to token. Foundry units are prefabricated at GigaOne, Torus’s Salt Lake City manufacturing facility. Prefabrication accelerates the time to deploy by eliminating construction time at the deployment site. Site prep takes about a week, and installation and commissioning take one to two days.
• More compute on the same interconnection. Integrated inertial flywheel and battery storage and management cover demand peaks so more compute fits on an existing interconnection. During grid outages, it provides backup power.
• Made in Utah. Foundry is designed, engineered, and manufactured in Utah by Torus's team of nearly 400 employees at GigaOne, with key components built in-house.

Multiple ways to deploy Foundry

Torus is launching Foundry with its first customers, Hydra Host and another AI cloud provider. Foundry is available in the following three models:

• Strategic: Powered shell includes power and enclosure, without GPUs.
• Partner: Turnkey bare metal unit includes power, enclosure, and GPUs.
• Direct: Sovereign inference capacity provided directly to enterprise customers. Single tenant hardware is allocated exclusively to a company’s workload.

Torus Foundry has thousands of GPUs deployed today, with plans to have more than 47,000 GPUs deployed in Foundry units by the end of 2027.

Capital to accelerate growth

The round was led by Magnetar, with participation from ICONIQ, Origin Ventures, Ridgeline Partners, and Trinity Capital. The capital will be used to fund the continued expansion of GigaOne, and Torus Reserve and Torus Foundry deployments across the United States.

“One of the biggest challenges in scaling AI infrastructure is access to power, and the solution isn't simply generating more of it, but making better use of what's already available,” said Dave Snyderman, Managing Partner and Global Head of Alternative Credit & Fixed Income at Magnetar. “Torus has spent years building the energy infrastructure and expertise to address this challenge. Foundry builds on that foundation by bringing compute to available power, helping unlock capacity within the existing grid. As demand for AI inference accelerates, we believe solutions like Foundry will become increasingly important, and we're excited to continue supporting Nate and his team as they scale.”

US-made manufacturing

Foundry is designed, engineered, and manufactured at GigaOne, the company’s Salt Lake City-based 545,000 square-foot manufacturing facility. GigaOne opened its first production line in early 2026, and has opened a new production line every 21 days in 2026. The Foundry production line is online and will be capable of producing 500 MW of IT load, with plans to expand to 1GW in the near future.

Sustainable by design

Torus built Foundry to be compute infrastructure that is friendly to communities and the environment. Its closed-loop cooling system is waterless, using 95 percent less water than evaporative coolers. Because it runs on existing grid capacity, it avoids the new transmission and generation buildout that large data centers often require. Each system also donates GPU time to medical research through Folding@home.

About Torus

Torus designs, engineers, and manufactures distributed mesh energy and modular AI infrastructure that delivers reliable power, reduces costs, improves resilience, and supports surging demand fueled by electrification and AI-driven load growth. Its full-stack technology integrates battery and inertial-based storage, modular AI infrastructure, AI-powered management, cybersecurity, and long-term operations and maintenance for utilities, data centers, manufacturers, and commercial and industrial facilities. Headquartered in Salt Lake City, Torus has earned a TIME Best Invention award, a 2025 R&D 100 Award for innovation, a 2026 Energy Storage award from Reuters, the Next Best Thing in Tech 2026 from Fast Company, and the 2026 Battery Manufacturer of the Year from The Battery Show. Learn more at torus.co.

Contacts

Media Contact: press@torus.co

View the original press release on Business Wire:
https://www.businesswire.com/news/home/20261111907147/en/Torus-Launches-Modular-AI-Infrastructure-Foundry-to-Meet-Surging-AI-Inference-Demand-and-Secures-%24500M-in-Funding