As AI chips and memory technologies race towards the realm of 3D stacking, thermal management has emerged as a critical bottleneck, constraining the expansion of artificial intelligence infrastructure. Professor Kim Jung-ho from the Korea Advanced Institute of Science and Technology cautions that the concurrent implementation of 3D stacking technology in system semiconductors and memory will intensify thermal dissipation challenges, potentially stifling further advancements in AI computing power. He proposes that leveraging AI-driven design automation and digital twin technologies holds promise in effectively curbing power consumption and heat generation within AI infrastructure.