Post by Eberhard Erb

Computational Fluid Dynamics (CFD) has been the go-to tool for data center thermal design for decades. But for the next generation of liquid-cooled AI data centers, it may no longer be enough. Why? AI workloads are changing the cooling problem: ✅ Heat is more concentrated. ✅ Loads are more transient. ✅ Cooling must respond in real time. ✅ The full thermal path matters, from chip to chiller. ✅ Controls can make or break efficiency. CFD still plays a critical role, especially for component-level analysis like cold plates and heat exchangers. But when teams need to understand how pumps, valves, coolant loops, heat exchangers, controls, and facility systems behave together over time, dynamic system simulation becomes essential. The future is not CFD vs. system simulation. It is knowing where each tool fits, and building a toolchain ready for liquid-cooled AI infrastructure. 👉 Read the full article: https://lnkd.in/gFrahkTX #DataCenters #LiquidCooling #AIInfrastructure #SystemSimulation #ThermalManagement #Modelon #CFD