Industry NewsAI Business & Ecosystem

Torus Secures $500 Million to Scale Modular AI Infrastructure Foundry

By Ash Kate
Torus Secures $500 Million to Scale Modular AI Infrastructure Foundry

Article content

Torus has secured $500 million in growth capital led by Magnetar as it launches Foundry, a modular AI infrastructure platform designed to meet the rapidly growing demand for AI inference computing.

The funding announcement, made on November 11, 2026, also includes participation from ICONIQ, Origin Ventures, Ridgeline Partners and Trinity Capital. The capital will support the expansion of Torus's manufacturing operations and accelerate deployments of its energy and computing infrastructure across the United States.

Headquartered in Salt Lake City, Utah, Torus is combining its experience in energy storage and power management with modular computing infrastructure. Its Foundry platform integrates power, storage, cooling, networking and compute into prefabricated systems designed to bring AI computing capacity online at existing commercial and industrial sites.

The approach targets a growing infrastructure challenge for the AI industry: securing sufficient electricity to support expanding computing workloads.

Bringing AI Compute to Available Power

Foundry is designed around a different approach to conventional data centre development. Rather than building large campuses and waiting for new grid connections, Torus aims to deploy computing infrastructure at locations where electrical capacity is already available.

Each prefabricated system combines the infrastructure required to operate AI workloads, including GPUs, cooling, networking and power management. Individual installations are designed to deliver between 1 and 20 megawatts of AI inference capacity.

Torus says deployments can move from contract to operational AI workloads in approximately 12 to 16 weeks. Site preparation generally takes about a week, while installation and commissioning can be completed in one to two days.

The platform is paired with Torus Reserve, the company's energy storage and management system. Reserve combines battery and inertial-based storage technologies with power management capabilities to help manage the fluctuating electricity demands associated with AI computing.

By using existing grid connections and managing power demand, Torus aims to make additional computing capacity available without depending entirely on new transmission infrastructure or conventional data centre construction.

Addressing the Growing Demand for AI Inference

The launch comes as the AI infrastructure market moves beyond model training towards the large-scale deployment and operation of AI applications.

Inference refers to the process of running trained AI models to generate responses, predictions and other outputs. As organisations integrate AI into products, enterprise applications and business workflows, demand for computing capacity to support these workloads continues to increase.

Torus cites McKinsey projections that demand for data centre capacity for AI inference could grow by 35% annually through 2030, compared with 22% for training.

This shift is creating demand for infrastructure that can be deployed closer to users and enterprise operations, with greater flexibility around power availability and computing capacity.

Foundry is designed to address this requirement through distributed, modular deployments rather than relying exclusively on large hyperscale campuses.

Early Deployments and Expansion Plans

Torus says Foundry already has thousands of GPUs running live workloads for enterprise customers. Its initial customers include Hydra Host and another AI cloud provider.

The company is targeting more than 47,000 GPUs deployed through Foundry installations by the end of 2027.

Manufacturing is centred at GigaOne, Torus's 545,000-square-foot facility in Salt Lake City. The company says its Foundry production line is online and is expected to support up to 500 megawatts of IT load capacity, with plans to expand towards 1 gigawatt.

The facility is also part of a broader manufacturing strategy that brings design, engineering and production together under one operation.

Funding to Scale Energy and Computing Infrastructure

The $500 million in growth capital will support the expansion of GigaOne, continued development of Torus Reserve and the deployment of Foundry systems across the United States.

Magnetar led the financing, with ICONIQ, Origin Ventures, Ridgeline Partners and Trinity Capital also participating.

For Magnetar, the investment reflects the importance of addressing power availability as AI infrastructure expands. The firm sees opportunities in technologies that make better use of existing electricity capacity while supporting the growing requirements of AI computing.

Torus is positioning its combined energy and computing model as an alternative way to expand infrastructure without relying solely on the development timelines associated with traditional data centres.

Integrating Energy Efficiency into AI Infrastructure

Alongside deployment speed and power management, Torus is highlighting the resource efficiency of its Foundry systems.

The platform uses a closed-loop cooling design that the company says consumes 95% less water than evaporative cooling systems. Its energy storage capabilities are also intended to manage demand peaks and provide backup power during grid outages.

These features address some of the infrastructure considerations facing AI data centre operators, particularly electricity availability, cooling requirements and resource consumption.

With Foundry, Torus is extending its energy technology business into AI computing infrastructure. The $500 million funding provides capital to expand that model as demand for inference capacity continues to grow.


About Torus

Torus is a US-based manufacturer of distributed energy and computing infrastructure headquartered in Salt Lake City, Utah. The company develops energy storage, power management and modular AI infrastructure solutions for utilities, data centres, manufacturers and commercial and industrial facilities. Its product portfolio includes Torus Reserve for energy storage and management and Foundry for modular AI inference computing.

Source: Torus official announcement