Enabling the Next Generation of AI Data Centers
DatacenterDynamics, Friday, September 18th, 2026
AI data centers need an integrated approach across power, cooling, permitting, and long-term operations.
Worley's Gireesh Nair argues that AI workloads demand far more power at greater density than the steady CPU workloads traditional data centers were designed around, with data center electricity consumption growing 12 percent annually over five years.
Site selection becomes an exercise in trade-offs across land cost, grid access, water, fiber, permitting timelines, and social license.
Many developers pay a premium for off-grid or hybrid architectures with gas generation, solar, batteries, and microgrid controls to reach market faster. Liquid cooling, heat reuse, and thermal energy storage add demand-response capability.
A Texas example includes 5GW gas generation, 1.25GW solar, and 20 buildings at roughly 250MW each.