Armada launches Leviathan, a megawatt-scale liquid-cooled compute system deployable at remote energy sites
Jul 28, 2025
Key Points
- Armada launches Leviathan, a megawatt-scale liquid-cooled compute system shipped in containers to remote energy sites, challenging the centralized data center model for AI infrastructure.
- The containerized approach lets companies deploy inference and training workloads wherever spare generation capacity exists, enabled by Starlink connectivity and mature liquid cooling.
- The product expands Armada's Gallion lineup as part of a broader strategy to consolidate AI workload distribution around American-controlled hardware and energy resources.
Summary
Armada launched Leviathan, a megawatt-scale liquid-cooled compute system designed to deploy AI infrastructure at remote energy sites and underutilized power sources across the US. The product expands Armada's Gallion lineup.
Leviathan is a shipping container that can be delivered to distributed locations where land and energy already exist. The play positions American AI infrastructure at the edge rather than concentrated in traditional data centers, bridging computational access to remote regions.
The timing aligns with several converging trends: Starlink's availability for remote connectivity, the maturity of liquid cooling at scale, and the ability to physically ship containerized compute to non-traditional deployment sites. The technical bet is that megawatt-scale inference and training workloads no longer require centralized hyperscale facilities. Instead, they can run wherever spare generation capacity exists.
Armada's founders are expected to detail the product and commercial strategy this week. The company is positioning Leviathan as part of the infrastructure race to consolidate AI workload distribution around American-controlled hardware and energy resources.