Eaton: AI data centers need aerospace-grade engineering

Eaton’s Peter de Bock says AI workloads are pushing data centers beyond legacy telecom-era designs.

He argues future AI infrastructure needs an aerospace-style systems approach, integrating power and liquid cooling from the start.

That could require redesigning everything from grid connection and power delivery to chip-level cooling.

With racks potentially reaching 600 kW and global AI demand topping 150 GW, infrastructure is becoming a major bottleneck.