Forecasts are useful. Earnings calls are useful. But nothing cuts through like a practitioner, working in the field, on real projects, stating what they see every day.

On May 30, 2026, a power industry professional posted this on X:

Field Report — Power Industry Practitioner, May 30, 2026 "I work in the power industry. I can tell you that 85% of the data center projects I'm involved in for the major players are either delayed or cancelled because of a lack of grid interconnect agreement, power gen equipment, or permitting."

"Only the DCs that are grid-independent — coming with their own power plant — are getting actually built out. Or the few that are getting built in city centers on top of an existing building and grid interconnect they can leverage."

85%. Not a projection. Not a model. A practitioner's direct observation from working on major player data center projects in 2026.

WHAT THIS CONFIRMS

Every number GridReadiness has been tracking points to the same conclusion — but a field practitioner's direct observation cuts through in a way that no analyst report can. Here is what the 85% figure confirms:

The Three Blockers — Confirmed by Field Reality Grid interconnect agreement: queue lengths of 7–10 years in Northern Virginia, 5–7 years in Texas
Power gen equipment: HV transformer lead times of 48–60 months at major US OEMs
Permitting: interconnection studies, environmental review, utility coordination — years of process
Combined effect: 85% of major player projects delayed or cancelled in 2026

THE RESIN BUSHING DETAIL — HOW DEEP THE BOTTLENECK GOES

A second practitioner in the same thread added a detail that illustrates how far the constraint extends into the supply chain:

"Resin bushings for HV transformers were back ordered for 36 months before the Strait of Hormuz. There will be a lot of robbing Peter to pay Paul."

Resin bushings are the insulating components that connect the transformer's internal windings to the external electrical connections. They are not the transformer itself — they are a component that goes into the transformer. And they were already on 36-month backorder before additional supply chain disruptions.

The bottleneck is not just the transformer. It is the components that make the transformer. It is the steel (GOES, Cleveland-Cliffs monopoly) that makes the core. It is the copper windings. It is the resin bushings. Every layer of the power equipment supply chain is simultaneously stressed.

THE BROWNFIELD INSIGHT — WHAT THE PRACTITIONER CONFIRMED

The practitioner's statement about what IS getting built is as important as what isn't:

"Only the DCs that are grid-independent — coming with their own power plant — or the few getting built on top of an existing building and grid interconnect they can leverage."

This is the brownfield thesis stated plainly. The projects that are proceeding are the ones that already have grid access. The ones that are failing are the ones starting from scratch.

GridReadiness has been documenting exactly this dynamic in France since launch: former aluminum smelters, EDF thermal plants, steel mills, and chemical complexes that already have HV connections are the only sites that can deliver AI data center capacity on a competitive timeline. They are not waiting for interconnect agreements. They are not waiting for transformer deliveries. They are not waiting for permitting that takes years.

THE FRENCH ANSWER TO AN 85% FAILURE RATE

Brownfield France vs US Greenfield — Timeline Comparison US greenfield grid interconnect: 7–10 years (Virginia) · 5–7 years (Texas)
US HV transformer lead time: 48–60 months
Combined US greenfield timeline: 8–12 years to first power
France brownfield (existing HV connection): 18–36 months RTE process
EU second-tier transformer (Efacec, Pauwels): 20–32 months
Combined France brownfield timeline: 24–42 months to first power
Difference: 5–10 years faster

The 85% failure rate is not a France problem. France has brownfield sites with existing HV connections across seven industrial categories — former aluminum smelters, EDF thermal plants, steel mills, chemical complexes, military bases, paper mills and automotive plants. GridReadiness has identified 40+ of these sites beyond the government's publicly announced list of 63.

The practitioners confirm the problem. The brownfield database confirms where the solution exists.