At Computex Taipei, keynote speaker and CEO of Nvidia Jensen Huang announced that his company will create one cloud data centre specifically designed to catch these scenarios so they can save on compute, projected to be at least 50% of all current generative AI compute.