AI’s rapid growth is pushing today’s data centre model toward hard limits in power, cooling and infrastructure. This session explores why those constraints are becoming unavoidable, which solutions can realistically extend Earth-based compute, and when more radical approaches may be required. The panel examines what must change to support AI at scale in the years ahead.

• Why are AI data centres hitting hard limits in power, cooling and grid capacity?

• Which solutions can realistically extend Earth-based compute and where do they fall short?

• When do off-world and alternative compute models begin to make strategic and economic sense?