The rise of artificial intelligence is disrupting the balance within data centers, particularly when it comes to cooling, which has become a critical issue in terms of performance, costs, and sustainability. Between increasing workload density, new architectures, and energy requirements, AI is no longer content to simply be hosted—it is imposing its own rules and redefining thermal strategies. This roundtable will explore how operators are adapting—and sometimes struggling to keep up—in the face of this technological and environmental pressure. • The explosion in thermal demands driven by AI workloads: graphics processing units (GPUs), density, and extreme hot spots. • Liquid cooling, immersion cooling, natural convection: which technologies are truly gaining traction? • AI for cooling: real-time optimization or a promise that’s still overhyped? • Energy efficiency or the race for power: a balance that remains difficult to achieve. • Sovereignty and data center location: what climate and regulatory constraints affect cooling?