Torus Raises $500 Million for AI Energy Storage
The Salt Lake City startup will use the Series C funding to build its shipping-container-sized Foundry product.
Updated on Nov. 11, 2026 in Data Centers

Live Poll
Do you think AI data centers are raising electricity bills in your area?
Torus, an energy-storage startup based in Salt Lake City, raised $500 million in a Series C round led by Magnetar, with participation from ICONIQ Capital, Origin Ventures and Trinity Capital. The company plans to use the funding to build Foundry, a unit that pairs battery storage with AI computing power.
Why it matters
U.S. data-center developers are scrambling to secure sufficient electricity for power-hungry AI workloads. Torus's approach places AI chips at sites with existing power supply rather than moving power to data-center sites.
Foundry units are shipping-container-sized and range from 1 to 20 megawatts in size, and can be installed within four months. The source does not specify the battery capacity or AI compute specifications of each unit.
The players
Torus
Torus is an energy-storage startup based in Salt Lake City that sells modular battery systems and is developing the Foundry product.
Magnetar
Magnetar is an investment firm that led Torus's $500 million Series C round.
Hydra Host
Hydra Host is an AI infrastructure firm and a Torus customer.
The details
Foundry containers can be placed on-site at businesses or plants whose local grid has unused electrical headroom. Hydra Host, an AI infrastructure firm, is listed as a Torus customer, and Torus has previously sold modular battery systems that help businesses control energy consumption and cut electricity costs.
Timeline
Last week: Democrats who ran on tougher data-center curbs won governorships in Texas and Ohio.
Within four months: Foundry units can be installed on-site.
Infrastructure Shift
The Texas and Ohio governorship results from last week, won by Democrats who ran on tougher data-center curbs, signal closer scrutiny of large campuses. Torus's model of placing AI computing at sites with existing grid headroom responds to that pressure by shortening schedules.
Because Foundry units can be installed within four months at sites with spare grid capacity, businesses with unused electrical headroom may gain a faster path to AI computing. Residents' electricity bills are not addressed in the source.
The takeaway
Placing AI computing where power already exists could ease grid strain while speeding deployment. Watch whether regulators in states with data-center curbs respond with new siting rules.
Further reading
For more on how power demand is reshaping AI infrastructure, explore Data Centers coverage.
Live Poll
Do you think AI data centers are raising electricity bills in your area?










