Professor Poh Seng Lee is a mechanical engineer whose research focuses on advanced thermal management, sustainable cooling and energy-efficient infrastructure for high-performance computing and artificial intelligence. His work spans the fundamentals of heat transfer and fluid mechanics through to component, system and data-centre-scale innovation, with particular emphasis on translating thermal-science advances into deployable solutions for high-density computing.

A major theme of his research is next-generation liquid cooling for AI and high-performance computing systems. His group investigates single- and two-phase direct-to-chip cooling, microchannel and advanced cold-plate architectures, flow boiling, hotspot mitigation, additive-manufactured thermal structures, thermal-fluid optimisation, coolant distribution and system-level integration. The research addresses the rapidly increasing heat fluxes and rack power densities associated with accelerated computing, while seeking to improve energy efficiency, thermal reliability, compute performance and resource utilisation.

Beyond individual cooling technologies, Professor Lee adopts a whole-system approach to sustainable data centres, integrating IT thermal management with facility cooling, heat rejection, power infrastructure, controls, water use and operating conditions. Particular attention is given to tropical environments, where high ambient temperature and humidity create distinctive challenges for data-centre efficiency and resilience. His research increasingly examines the co-design of computing, power and cooling infrastructure for future AI factories and high-density digital infrastructure.

Professor Lee also works on energy-efficient air conditioning, thermal-energy systems, waste-heat utilisation and sustainable energy technologies, drawing upon heat-transfer science, computational modelling, experimentation, optimisation and intelligent control.

A defining feature of his research is the translation of fundamental engineering knowledge into full-scale demonstration, industry deployment, standards and policy-relevant evidence. He leads multidisciplinary collaborations involving academia, government agencies, data-centre operators and technology companies, including large-scale tropical data-centre testbeds that enable technologies to be evaluated under realistic operating conditions.

His overarching research objective is to develop the thermal and energy infrastructure required for sustainable computing at scale—bridging fundamental thermal sciences, engineering innovation and real-world implementation to enable increasingly powerful digital infrastructure within energy, water, carbon and reliability constraints.