EnerNex: large loads, new rules — SkimNews
Get the Energy newsletter
Daily energy & climate — solar, EVs, oil, the policy fights and tech bets shaping the transition. Free.
- AI training campuses are proposing loads of 1,000 MW or more — a new category of grid customer whose size and electrical behavior exceed the static load assumptions conventional interconnection studies were designed to evaluate.
- AI compute workloads can shift power consumption by hundreds of megawatts within seconds as training jobs start, pause, checkpoint and conclude, a dynamic that traditional power-flow and positive-sequence dynamic studies were not built to capture.
- Utilities across North America are turning to electromagnetic transient (EMT) studies to assess fault ride-through, ramp-rate limits, low-frequency and sub-synchronous oscillations, and power-quality issues like flicker and harmonics for large AI facilities.
- EMT model quality is now a critical bottleneck: equipment vendors often supply models with incomplete parameter sets, undocumented controls, or inconsistencies between EMT and positive-sequence representations, and one deficient model can stall multiple projects in clustered interconnection queues.
- Validated EMT models frequently do not exist for UPS systems, cooling drives and distribution technologies that define data center electrical behavior, pushing utilities to establish ride-through, ramp-rate, flicker, harmonic and reactive-power requirements directly at the point of interconnection.
- EnerNex (CESI Group) is delivering advanced interconnection studies and large-load performance assessments to help utilities and developers connect AI infrastructure reliably across North America.
Why it matters: For utilities, developers and EPC firms, the gating factor for interconnecting 1,000+ MW AI campuses is no longer engineering capability but vendor model completeness — a single deficient EMT model can delay multiple clustered projects at once, and validated models often don't exist for the UPS, cooling and distribution gear that defines data center behavior.
Ask SkimNews
