Johnson Controls has unveiled its second data centre reference design guide that offers a globally repeatable blueprint for cooling gigawatt-scale data centres.
The company says that its updated AI Factory Reference Design Guides focuses on air-cooled chillers and offers a plan to support zero water cooling, eliminating cooling towers and saving more than 12 million gallons of water per day.
It also enables high-efficiency air-cooled designs that can return up to 50MW to the AI factory and improve annual energy consumption by 32 percent.
The plan allows operators to achieve upto 20MW peak power savings by quantifying and mitigating heat island effects of air-cooled chiller plants. It also envisions a roadmap to 30 percent Coefficient of Performance (COP) improvement and 27 percent fewer chillers from raising the chilled water temperature to support warm-water Technology Cooling System loops.
As AI brings new scale and complexity to data centre design, the company says that operators are facing mounting challenges including the power needed for cooling systems, rising cooling loop temperatures, efficiency losses from heat islands and limited water availability.
Building on its previous water-cooled chiller guide released in February, this latest blueprint maps the full data centre thermal chain – with additional guides covering absorption chillers and direct-to-chip liquid cooling to follow.
The guide outlines how designers and operators can maximise energy and water efficiency and minimise noise impact on surrounding communities while remaining adaptable to different climates, workloads and growth paths.
It supports the scalable design of data centres of all sizes, up to a 1-GW AI factory, utilising air-cooled chillers.
It outlines a thermal cooling architecture that integrates high-efficiency air-cooled YORK centrifugal chillers (including its YDAM and YVAM chillers), fan coil walls (FCWs) and coolant distribution units (CDUs) to manage both air- and liquid-cooled IT loads.
Additionally, the guide provides sizing references for 220MW compute clusters, including recommended design temperatures and operating conditions across each stage of the thermal chain.
“At gigawatt scale, AI factories require a fundamentally different way of thinking about infrastructure,” Johnson Controls Global Data Centre Solutions president Austin Domenici said.
“The future requires designing integrated systems that can scale predictably, perform efficiently and adapt as technology evolves.”
Image: Johnson Controls Media Centre
Read more: Johnson Controls to acquire thermal management specialist Alloy Enterprises
