STULZ on Liquid Cooling and AI: Wait, Watch or Act?

Data center operators worldwide face a strategic dilemma: upgrade AI infrastructure now with current GPUs, or wait for the next generation? STULZ argues that hesitation is itself a risk. Liquid cooling has already become the de facto minimum standard for AI workloads, and the expertise required to operate it cannot be bought overnight — it must be built through real projects.

What was announced

In a new expert analysis, Jonathan Aaron J"urgens, Team Lead Liquid Cooling Global Product at STULZ, examines why data center operators should act on AI-ready cooling today rather than wait for the next wave of GPUs. He points to structural GPU shortages, tightening EU regulation, and regional differences in AI infrastructure strategy as reasons to start pilot projects now.

Key facts from the analysis:

• PUE 1.2 mandate: The German Energy Efficiency Act (EnEfG) requires new data centers to achieve a PUE of 1.2 or lower from 2026.
• Heat recovery: The EU Energy Efficiency Directive requires heat recovery analyses for new facilities above 1 MW.
• Rack density target: STULZ recommends designing AI infrastructure for 125"140 kW per rack, with headroom for one to two future generations.
• Field experience: Over 100 global liquid cooling projects delivered since the launch of the CyberCool CDU.

Why this matters

For CIOs, CISOs, IT directors and procurement leaders, the message is clear: AI-ready infrastructure is no longer a future project. Waiting for the next GPU generation exposes organizations to allocation shortages, since hyperscalers absorb the first waves of supply. Regulatory pressure in Europe effectively mandates liquid cooling for high-density AI racks, and geopolitical risk in the Asian supply chain adds urgency to early planning.

Operators who begin now build hands-on knowledge of CDU maintenance, thermal behavior under AI load, and pay-as-you-grow scaling models. That know-how transfers directly to the next hardware generation and becomes a durable competitive advantage.

Those who wait today are not only waiting for better hardware, but also for supply chains that could quickly grind to a halt in the event of geopolitical shifts

Technical details

• Rack capacity design: 125"140 kW baseline, scalable by one to two GPU generations.
• Modular CDU architecture: predefined interfaces for future hardware and pre-dimensioned hydraulic paths activated as utilization grows.
• Digital twin: established during planning to unlock later optimization of temperature profiles and energy efficiency.
• Co-engineering approach: replaces traditional bidding with joint design between operator and technology partner.
• Sustainability alignment: heat recovery readiness and PUE targets aligned with EnEfG and EU Energy Efficiency Directive.
• Regional balance: combines Asian speed of adaptation, U.S. scaling models, and European regulatory standards.

Softprom and STULZ

Softprom is the official distributor of STULZ. Our team helps data center operators design, source and deploy AI-ready liquid cooling infrastructure aligned with EnEfG and EU efficiency requirements.

This content was prepared as part of the Softprom DistriFlow project " an automated system for monitoring and adapting vendor news. Original source: original article.