Johnson Controls joins Singapore AI data centre testbed

Johnson Controls has joined the Sustainable Tropical Data Centre Testbed 2.0 at the National University of Singapore, a programme that will examine AI-ready data centre systems in tropical conditions.

It will contribute thermal management, smart controls, engineering and operational expertise to the project, hosted by the university's College of Design and Engineering and based on Jurong Island.

The move comes as Singapore expands data centre capacity while trying to limit the impact of rising power, water and land use. The city-state has approved additional capacity for operators, but pressure on infrastructure has sharpened the focus on efficiency in facilities designed for AI workloads.

STDCT 2.0 brings together industry and academic partners to test data centre and AI technologies in a live tropical environment. The work is part of a wider effort with JTC Corporation and other partners to use Jurong Island as a test site for lower-carbon digital infrastructure.

Resource pressure

Cooling is central to that effort. High temperatures and humidity can reduce cooling performance and increase energy demand, especially as AI systems drive higher rack densities and put more strain on power and building systems.

The research will focus on reducing electricity and water use while increasing computing capacity from existing infrastructure. Johnson Controls said more can be done to improve the overall thermal system in data centres rather than treating cooling as a standalone function.

In a modelled design, its integrated thermal management approach could cut non-IT energy consumption by up to 50% compared with conventional cooling methods. The company also pointed to a 1-gigawatt data centre blueprint in which an absorption chiller reference design could use waste heat from on-site power generation to provide cooling.

According to Johnson Controls, that design showed the potential to unlock up to 97 MW of additional AI computing capacity from existing power infrastructure while reducing cooling-related electrical demand by up to 44%. The figures were presented as reference outcomes from modelling rather than operating results from a live facility.

Austin Domenici, President of Data Centre Solutions at Johnson Controls, said the industry needed to rethink how data centre infrastructure is designed as AI demand grows.

"Meeting the demands of advanced computing requires a new way of thinking about infrastructure," said Domenici. "As AI workloads continue to grow, success will depend not only on computing power, but on how efficiently we manage the resources behind it. Johnson Controls brings expertise across the full thermal ecosystem, helping customers improve efficiency, reduce resource consumption and enable greater computing capacity. This collaboration is an opportunity to explore and validate new approaches to infrastructure that can support the future of AI."

Testing platform

The university sees the programme as a way to move beyond testing isolated components. Instead, the testbed is intended to examine how cooling, controls, power and other systems work together under local operating conditions.

Professor Lee Poh Seng, Programme Director of the Sustainable Tropical Data Centre Testbed and Head of the Department of Mechanical Engineering at NUS College of Design and Engineering, said the challenge had changed as AI altered the physical demands on data centres.

"AI is fundamentally changing the design envelope for data centres, with much higher rack densities placing unprecedented demands on cooling, power, controls and resource efficiency," said Lee. "STDCT 2.0 is designed to move beyond individual technologies and evaluate how these systems can be integrated and optimised as one infrastructure platform under real tropical operating conditions. Johnson Controls brings deep expertise in thermal management, controls and mission-critical infrastructure to this collaboration. Together, we aim to translate research into validated, deployable solutions that enable greater computing capacity with constrained power and water resources, while strengthening Singapore's capabilities, talent base and leadership in sustainable AI infrastructure."

Beyond technical testing, the programme also includes research, training and knowledge-sharing. That reflects broader concern in the region that the growth of AI infrastructure will require more engineers and operators with experience managing dense, energy-intensive facilities in difficult climates.

Ali Badreddine, Vice President and General Manager for Southeast Asia Business and Asia Pacific Data Centre Solutions at Johnson Controls, linked the project to a wider regional buildout.

"As AI drives a new wave of infrastructure demand across Asia Pacific, advancing research, developing talent and strengthening industry collaboration will be critical to meeting that growth responsibly," said Badreddine. "STDCT 2.0 brings these elements together to help advance AI-ready infrastructure across Singapore and the broader Asia Pacific region."