Armada, an SI Infrastructure manufacturer, announced support for the NVIDIA DSX POD reference design. DSX POD enables partners to build standardized, factory-integrated containers for deploying SI capacity faster and at scale. Under the collaboration, Armada and Johnson Controls will engineer, manufacture, integrate, and deploy pre-fabricated SI factories built on the NVIDIA Vera Rubin platform, with cabling, wiring, piping, and functional validation completed before a single unit leaves the factory.
Armada was accepted into NVIDIA Inception in 2024 and started working that year with the company’s network simulation technologies. Armada now builds on the NVIDIA DSX platform across the full life of an SI factory: NVIDIA DSX Air to simulate the AI factory and validate its Bridge GPU management software before hardware ships, the DSX POD reference design to manufacture containers, and NVIDIA DSX MaxLPS to maximize token output within each site’s power envelope once it is running. The Armada Platform also supports NVIDIA AI Grid, enabling telecommunications operators, service providers, and enterprises to deploy, operate, and monetize geographically distributed AI infrastructure with simplicity while supporting latency-sensitive, real-time AI workloads.
Superintelligence — AI systems that match or exceed human experts across most fields — has become the stated goal of the world’s leading AI labs, and reaching it depends on compute at a scale the world has never built. Yet the infrastructure to run that compute is built on one of the slowest-moving models in the economy. To address this, Armada plans to partner with Johnson Controls to build 500 megawatts of distributed sovereign SI capacity in 2027.
A conventional 50 MW facility can take years to become IT-ready and consume roughly 600,000 on-site labor hours, at a time when the construction industry faces a projected shortfall of 288,000 workers by 2027. Permits, power interconnects, and community approvals stretch timelines further. Companies and nations can own the chips, the land, the power, and the capital, and still be told to wait.
“The SI race isn’t going to be won by whoever pours concrete the fastest,” said Dan Wright, Co-founder and CEO of Armada. “Companies and nations can have the chips, the land, the power, and the ambition, and still be told to wait years for a building. The NVIDIA DSX POD reference design lets us cut that wait out of the equation. We build the SI factory in one of our Galleon Forge factories, ship it, connect it, and our customers are generating tokens in months, on infrastructure they own and control.”
The NVIDIA DSX POD reference design moves repeatable construction work off the job site and into the factory. Compute, networking, rack-level power, liquid-cooling plumbing, and supporting infrastructure are integrated and validated before shipment. On site, Armada connects containers to prepared power, water, cooling, and networking, and customers scale by adding standardized containers as demand grows. NVIDIA DSX MaxLPS helps customers make more productive use of available power across that footprint through policy-driven power control, dynamic workload orchestration, and high-fidelity telemetry, maximizing throughput within defined power limits.
“The demand for compute is outpacing the speed at which the world can build the facilities to house it,” said Charlie Boyle, vice president of DGX Systems at NVIDIA. “By utilizing the NVIDIA DSX POD reference design, Armada can deliver factory-built, Vera Rubin-based containers to customers. Together with Johnson Controls, Armada is positioned to bring Vera Rubin-based factories to customers who need capacity now, wherever their energy and data are.”
Armada brings to the program a manufacturing and deployment capability built over years of putting data centers in places never designed for them: in frozen Alaska, on a U.S. Navy littoral combat ship, at desert sites in the Middle East, and soon on a North Sea oil platform. That work secured Armada’s supply chain, built its engineering bench, and established the need for continuous manufacturing across the United States and internationally. Armada’s rugged modular data center readiness is achieved in partnership with Johnson Controls. Armada’s SI factory line is designed directly with NVIDIA and in line with NVIDIA reference architectures.
“NVIDIA is rapidly advancing the age of intelligence and reshaping the infrastructure required to support it,” said Joakim Weidemanis, CEO of Johnson Controls. “Our collaboration with Armada combines their expertise in deploying compute at the edge with Johnson Controls’ advanced thermal management technology, engineering capabilities and global manufacturing scale. This combination delivers high-performance infrastructure engineered to quickly bring capacity online where and when it’s needed.”
Armada deploys sovereign SI capacity as a grid rather than a monolith: a network of modular SI factories placed close to energy sources, including stranded, renewable, and off-hours power, with minimal community footprint and water impact. Each site is built on an open architecture with zero cloud dependencies, so customers own and control their models, their data, and the infrastructure that runs them.
With this collaboration, NVIDIA designs, tests, and validates the DSX POD reference design for the Vera Rubin platform. Armada engineers, manufactures, integrates, customizes, services, supports, and delivers the modular data centers to customers at scale.
