Poseidon Cooling

COOLING.
BUILT FOR SCALE.

Liquid cooling infrastructure for AI data centers — Spider Cooler and CDU engineered for precise thermal control in high-density compute environments.

Poseidon Spider Cooler

BÉTHUNE — FIGURES FOR THE PROJECT

240 MW IT total capacity

4 data halls — 2,000 m² each, 60 MW IT each

25,000 m² building area — former BRIDGESTONE industrial site

MEGABAYS coolers — 340 MW

HV transformers — 6 × 80 MW

HV electrical substation — 90 kV / 20 kV

Temporary gas generators — 15 × 2 MW

Poseidon Spider Cooler for AI data centers

Engineered for AI-Scale Cooling

Poseidon Cooling is engineered for the thermal demands of modern AI data centers — combining Spider Cooler and CDU in a modular liquid-cooling approach for high-density compute.

Poseidon cooling deployment

Faster Deployment

Poseidon simplifies high-density liquid cooling rollout, helping operators scale AI capacity with greater speed and thermal confidence.

  • Modular infrastructure for scalable deployment

  • Designed for high-density AI workloads

  • Streamlined installation and integration

Spider Cooler

Direct thermal capture at the GPU and accelerator layer — the core Poseidon product for high-density AI racks.

Poseidon Spider Cooler for high-density AI racks

Spider Cooler — Rack interface

Direct thermal capture at the GPU and accelerator layer. The core Poseidon product for high-density AI racks.

Spider Cooler — Rack interface

Coolant loop — Controlled flow

Stable supply and return paths between the CDU and Spider Cooler, sized for continuous AI heat loads.

Coolant loop — Controlled flow

Monitoring & controls

Sensors and controls keep temperatures within target bands and surface service alerts before thermal risk builds.

Monitoring & controls

Cooling Distribution Unit

Bridges facility chilled water and the Spider Cooler loop — flow, temperature, and redundancy in one module.

Poseidon Cooling Distribution Unit (CDU)

CDU — Distribution heart

Cooling Distribution Unit bridges facility chilled water and the rack loop — flow, temperature, and redundancy in one module.

CDU — Distribution heart

Facility chilled water

Interfaces with the site chilled-water plant. Prepares the secondary loop that feeds Spider Cooler rows at hall scale.

Facility chilled water

Scalable hall layout

Repeatable Spider Cooler + CDU building blocks for dense rows — with chillers available in the broader portfolio.

Scalable hall layout

Key specifications

Core figures for the Spider Cooler and CDU stack — built for AI-density heat loads, reliability, and rapid hall deployment.

Spider CoolerCore product
CDUDistribution
AI / GPUTarget workload
LiquidCooling type
ModularArchitecture
Chilled waterFacility link
Chiller readyPortfolio
24/7Support

Request a free technical study

We’re here to help. Get a tailored assessment of your AI data center cooling infrastructure, free of charge.

Follow us: