Synopsys has joined the Sustainable Tropical Data Center Testbed Phase 2.0 (STDCT 2.0) Consortium in Singapore, a group led by the National University of Singapore (NUS) and hosted by its College of Design and Engineering. The work centers on tropical data center cooling, combining simulation, AI-assisted design, reduced-order modeling, and validation using STDCT 2.0’s physical testbed.
For data centers operating in tropical climates, the consortium is targeting practical constraints that don’t show up the same way in temperate regions. The program cites ambient conditions that can exceed 24-32 °C and 70-80% relative humidity, which can reduce the efficiency of conventional cooling systems and increase energy requirements. In that operating envelope, cooling can account for up to 40% of a tropical data center’s total energy consumption.
STDCT 2.0 is set up as a co-development environment for testing sustainable data center technologies under tropical conditions, with an explicit focus on AI-ready infrastructure. Synopsys and NUS CDE plan to connect physics-based simulation, AI-assisted engineering, and reduced-order modeling with experimental measurements taken from the testbed.
Simulation-to-testbed workflow for liquid cooling and digital twins
The collaboration spans advanced liquid cooling, computational fluid dynamics (CFD), multiphysics simulation, AI-assisted thermofluid design, reduced-order modeling (ROM), and digital twins. Initial research concepts include physics-based and AI-assisted methods for two-phase direct-to-chip cooling of high-heat-flux AI accelerators, plus uncertainty-aware reduced-order digital twins aimed at predictive thermal management for liquid-cooled AI systems.
The planned workflow starts with high-fidelity simulation to explore thermal-fluid behavior and evaluate design options, then moves to experimental validation inside the STDCT 2.0 environment. Validated results are intended to feed computationally efficient ROMs that can be integrated into digital twins for optimization and operational decision support. The groups also plan to develop reusable workflows, case studies, validation methods, and a “flagship demonstration” linking design, simulation, ROM generation, digital twin integration, and physical testing.
The engineering payoff is straightforward: in high-density AI deployments, cooling designs that only work in a solver aren’t enough. A loop that ties CFD and multiphysics work to measured testbed data is what separates a plausible concept from something an operator can risk in production.
“Efficient AI infrastructure requires a clear understanding of interconnected multiscale and multiphysics behavior, from chips to data centers,” said Padmesh Mandloi, Vice President, Applications Engineering, Synopsys Simulation and Analysis APAC, Synopsys.
STDCT 2.0 brings together research institutions, data center operators, and technology providers, and Synopsys said it will contribute software access, technical expertise, training, project mentoring, and learning resources within the agreed scope. “STDCT 2.0 provides a co-innovation platform to move promising technologies from simulation to validation and towards real-world deployment in tropical data centers,” said Professor Lee Poh Seng, Programme Director, Sustainable Tropical Data Center Testbed, and Head, Department of Mechanical Engineering, College of Design and Engineering at NUS.
Source: Synopsys
