Leviathan is the fastest way to go from land and power to scaled AI factoryLearn More

Background Image

Leviathan Deploys Megawatt-Scale AI, Without Building a Traditional Data Center

Leviathan delivers 1.77 MW of AI infrastructure, operational in 12 weeks.

Request a Site Assessment
Download Technical Summary

  • 1.77 MW

    IT envelope

  • 576

    GB300 NVL72 GPUs / unit

  • 2.2 MW

    Cooling capacity

  • .12 acres

    Footprint per unit

  • 50 MW+

    Cluster scale path

Megawatt AI Infrastructure, Delivered as a System

Image description

Leviathan is Armada's largest Galleon form factor, purpose-built for GPU-dense AI infrastructure at megawatt scale. Each unit integrates compute, networking, liquid cooling, UPS-backed electrical distribution, and monitoring into a single deployable architecture.

It is not a shell, enclosure, or containerized room. It is a complete AI infrastructure system — power, cooling, compute, networking, controls, and software in one deployable architecture — optimized for 5 MW to 20 MW and designed to scale to 50 MW and beyond.

Image description
  • 01

    1.77 MW IT envelope per unit

  • 02

    Up to 576 NVIDIA GB300 GPUs per unit

  • 03

    Future-Proof IT Module: Vera Rubin, Rubin Ultra, Feynman

  • 04

    Liquid-first cooling

  • 05

    UPS-backed electrical distribution

  • 06

    Integrated monitoring & controls

  • 07

    NVIDIA Quantum InfiniBand & Spectrum-X support

  • 08

    Aligned to NVIDIA Reference Architectures

The Fastest Way to Deploy a Turnkey AI Factory

Image description

An AI Factory is not a general-purpose data center. It is a production system designed to convert power, GPUs, data, and software into AI output at scale.

That requires high-density compute, liquid cooling, low-latency networking, orchestration, tenant isolation, observability, billing, and operations designed to keep GPUs productive. Leviathan provides the physical infrastructure. Bridge provides the software layer. Together, they create a complete AI Factory stack.

Image description

AI Factory

Inputs

Start with a site

  • Land
  • Power
  • Connectivity
  • Local requirements
Software layer

Bridge

  • Orchestration
  • Multi-tenancy
  • Utilization
  • Observability
  • Billing
  • Marketplace
Infrastructure

Leviathan

  • GPUs
  • Networking
  • Cooling
  • Power
  • UPS
  • Monitoring
Outputs

AI Output

  • Inference
  • Model deployment
  • Fine-tuning
  • Training

Engineered for
Megawatt-Scale AI Workloads

GPU Configuration

  • GB300 NVL72 GPUs per unit576
  • GPU super cluster, 8-unit configuration4,608
  • GPUs per unit with HGX B300 / B200512
  • Compute racks at 200 kW each08
  • Networking racks at 42.5 kW each04

From Site-Ready to
Operational AI Capacity

Leviathan deploys on prepared land with available utility power and network connectivity. Armada's solution co-review process defines the final configuration, site interface, cluster topology, and deployment plan before hardware ships.

  1. 01

    Site
    assessment

  2. 02

    Solution
    co-review

  3. 03

    Site
    readiness

  4. 04

    Delivery &
    connection

  5. 05

    Operational
    AI capacity

What the site needs to provide:

  • Land sufficient for the Leviathan footprint and service clearance
  • 480V / 60Hz utility service metered to Armada specifications
  • Network connectivity
  • Local regulatory coordination: grounding, metering, permitting

What Leviathan delivers:

  • IT infrastructure module, compute & networking racks
  • UPS-backed electrical plant
  • Mechanical plant and chillers
  • Liquid-first thermal architecture, monitoring & controls

The Complete AI
Factory Stack

Leviathan deploys on prepared land with available utility power and network connectivity. Armada's solution co-review process defines the final configuration, site interface, cluster topology, and deployment plan before hardware ships.

AI Factory stack diagram

Bridge:
Software Layer

  • GPU orchestration
  • Workload scheduling
  • Multi-tenant access
  • Job telemetry
  • Utilization visibility
  • Billing and metering
  • Marketplace integration

Leviathan:
Infrastructure Layer

  • GPU capacity
  • Cooling
  • Power
  • Distribution
  • Networking
  • Monitoring
  • Horizontal scale

Start With Megawatts.
Scale to AI Factory Campuses

Leviathan is designed to scale horizontally. A single unit delivers 1.77 MW of IT capacity. Four units form a Pod. Multi-unit deployments target 5–20 MW and scale to 50 MW and beyond.

1 Leviathan

1 Leviathan · 1.77 MW

Factory-built AI infrastructure in a single deployable unit.

1 Leviathan Pod

4 Leviathans · 5.35 MW

Multi-unit cluster architecture from four integrated units.

5-20 MW Deployment

3 - 12 Leviathans

For operators bringing regional, enterprise, or sovereign AI capacity.

50 MW+ Cluster

29 Leviathans

Designed for AI Factory Campus expansion as demand grows.

Built for Operators with Land, Power, Customers, or Sovereign Demand

Add AI Capacity Without Waiting on a Full Facility Build

For data center operators, Leviathan provides a faster path to GPU-dense expansion. Deploy alongside existing facilities or on adjacent land, reduce integration complexity, and use Bridge to offer managed AI compute services to customers.

Why Leviathan

  • Faster path to GPU-dense expansion
  • Deploy adjacent to existing facilities
  • Bridge for managed compute services
Expand AI Capacity
Image description

Schedule a Technical Exchange

Start With a Site Assessment

Tell us about your land, power, connectivity, and target capacity. Leviathan is designed for customers with available or planned megawatt-scale power capacity. Armada's engineering team will evaluate site fit and recommend a deployment path.

  • Request a Site Assessment
  • Talk to a Deployment Engineer
  • Download the Technical Summary
Download Technical Summary
Image description