AI data center site readiness
Cliff reviews power access, utility rules, air permits, local approvals, and public records to identify what could delay, shrink, or block a proposed data center project.
De-rate analysis · Apr 2026
ERCOT West · 280 MW load
SB 6 curtailable · Cluster D · COD Q3 2027
Nameplate
280 MW
Effective capacity
198 MW
−82 MW · 29% de-rate
Binding constraint
Recommendation
Negotiate a higher backup runtime cap before signing the interconnection agreement, or re-spec to firm interconnection in Cluster B. Either path closes the 144-hr gap.
What people ask us
Site readiness for data centers cuts across power access, utility rules, air permits, local approvals, and public records. Here's how we'd answer the questions developers, counsel, and consultants actually bring us.
The binding constraint is usually federal RICE NESHAP runtime hours, not generator nameplate. Sites that did the interconnection math without the air-permit math are routinely surprised by the modeled summer de-rate — 20-30% is typical when backup gen is capped at 100 nonemergency hours per year and curtailment events stack up in the same window.
When the firm interconnection queue pushes energization out by years and the site has enough air-permit headroom to run backup generation during curtailment events. The math flips quickly when NESHAP runtime hours run out before curtailment events do, or when the site sits in a nonattainment area that limits dispatch.
There's no clean public ranking — what you can pull is each county's recent zoning approvals, special-use permit history, and comprehensive plan amendments. Northern Virginia counties tightened after community pushback; Texas varies county-by-county; parts of Pennsylvania and Ohio are still permissive but moving.
State DEQ workload is the usual bottleneck, not federal review. PSD review adds material time when triggered. Permit modifications are much faster than new permits when the site can be engineered to fit existing conditions, so the timeline question often turns into a design question.
The April-July 2026 deadlines are real and being enforced. PCLR (Permanent Controllable Load Resource) status requires telemetry and dispatchability commitments most developers haven't priced into their model. Projects that miss the docket windows get pushed to later batches, which compounds with queue movement.
Utility distribution capacity studies and TSP/EHV interconnection queues are the obvious ones. The underused signals are county comprehensive plan amendments, water utility long-range forecasts, and EPA air monitoring data flagging counties drifting toward nonattainment. Each one closes a door before the IA conversation starts.
The trap, in three steps
A faster interconnection path can energize earlier, but it may still be curtailable during peaks.
Backup generation only closes the gap if the federal rules and site permit let those hours count.
When those hour buckets run out before the curtailment events do, nameplate capacity quietly turns into a de-rate.
Air permit hours vs. expected curtailment
ERCOT West · 280 MW load · summer 2026 dispatch model
Why now
Power is the bottleneck. Faster interconnection products only matter if you can model the real operating envelope, and that answer is still moving in live dockets and guidance memos.
Live regulatory windows
Three ISO-level dockets plus state-level rules layered on top. Each on a different clock.
Run it yourself
The calculator returns an effective MW number, the binding rule, and a $/MW-yr net value as you type. The full report unlocks with your email.