US data center developers, facing grid access wait times of up to seven years, are turning to aeroderivative turbines and diesel generators to power the AI boom
Supply chain shortages drive developers to use smaller and less efficient power sources to fuel AI power demand
Context & Ripple Effects
Data-center construction was already constrained by shortages of parts, sites and power, with cooling-system lead times lengthening sharply as companies raced to add AI capacity. The reported grid-access delays make electricity availability a near-term deployment constraint rather than a background operating cost.
Developers had also begun building on-site power plants to work around overloaded grids. The turn to aeroderivative turbines and diesel generators extends that workaround to projects unable to wait for conventional interconnection.
First-order effects
- US data-center developers can advance AI facilities using on-site generation, but must accept smaller, less efficient power sources while grid connections remain unavailable.
- Demand shifts immediately toward aeroderivative turbines, diesel generators and the equipment needed to deploy them, while developers take on more responsibility for power provision.
Second-order effects
- The workaround collides with an existing gas-turbine supply crunch, potentially making generation equipment another schedule risk for data-center projects.
- On-site power becomes a competitive execution advantage for developers able to secure equipment and permitting, while projects reliant on grid timelines face less certainty over when capacity can go live.
Third-order effects
- If this pattern persists, AI-compute expansion will be planned increasingly as a utility-infrastructure project, with grid access and generation procurement shaping where and when new capacity is viable.
- The bottleneck may move from grid interconnection alone to a broader chain of generation equipment, fuel and supporting electrical infrastructure; the balance between temporary generation and durable grid investment remains uncertain.
The trend: AI infrastructure is becoming inseparable from power-system buildout as developers use self-supplied generation to bridge grid and equipment bottlenecks.