Behind every leap in AI capability lies an equally immense leap in infrastructure. Training and running large-scale models isn’t just about algorithms, it’s about the physical systems that make them possible. From energy-hungry data centres and global supply chains to cooling systems and site selection, scaling intelligence requires an extraordinary amount of real-world engineering. In this keynote, we’ll pull back the curtain on what it takes to power AI at scale the unseen costs, trade-offs, and innovations shaping how intelligence is built and sustained. Key takeaways: → What drives the true cost of AI infrastructure - beyond GPUs and electricity. → How design, energy strategy, and operational execution define AI’s limits and potential. → Why the next era of AI advancement depends as much on power grids and data centres as it does on model architecture. This is a rare inside look at the hidden machinery of intelligence and the physical world that keeps it running.