Applying our deep thermal-fluid and CFD expertise to a new frontier — cooling high-density AI computing. From direct liquid cooling to immersion systems.
Direct-to-chip liquid cooling for GPU clusters. Cold-plate design, manifold distribution, leak detection and monitoring — engineered with the same CFD rigor we apply to coating ovens.
Full server submersion in dielectric fluid for maximum thermal density — supporting 100+ kW per rack for AI training farms, up to 40% energy saving versus air cooling.
Close-coupled, row-based precision cooling with hot/cold-aisle containment for data halls and edge rooms — ASHRAE-compliant.
Container or skid-mounted modular data centers with integrated power, cooling and fire suppression — leveraging our steel fabrication strengths for fast deployment.
| Technology | Max Rack Density | PUE | Best For |
|---|---|---|---|
| Air Cooling | 20-30 kW | 1.4-1.8 | Standard servers |
| Liquid Cooling (CDU) | 60-100 kW | 1.1-1.2 | GPU clusters |
| Immersion Cooling | 100-200 kW | 1.03-1.08 | AI training farms |
| Rear-Door HX | 30-60 kW | 1.2-1.35 | Retrofit upgrade |
Nearly 30 coating lines delivered. Let us help you plan, build or upgrade yours.
Scan with WeChat
