White Paper · Upstream Coherence Measurement Stratum (UCMS) · October 2026

America's Grid Is Running Out of Runway: Six Pressures Converging on U.S. Utilities

Ronald Brogdon · Upstream Coherence Measurement Stratum (UCMS)
ORCID: 0009-0009-0507-2971

Introduction

We are entering territory the U.S. power grid was never engineered for. Not a decade from now, but now, as AI data centers, extreme weather, and aging infrastructure press on the same system at the same time. The signs are visible in public data: interconnection queues holding more capacity than the country's entire installed generating fleet, wildfire liabilities weighing on investor-owned utilities, and load forecasts that no longer resemble anything in the historical planning record.

This is not a story about one bad storm season or one mismanaged utility. It is a story about structural strain accumulating faster than the system can absorb it, across physical infrastructure, digital control systems, and financial capacity at once.

Six Converging Pressure Points

Looking across the country's largest utilities and every major grid operator, six pressures recur:

1. Catastrophic risk exposure. Liability and losses from wildfires, hurricanes, and extreme weather.
2. Interconnection and capacity backlogs. New generation waiting years behind transmission bottlenecks.
3. Hyperscale and AI load coupling. The fastest, least predictable load growth in grid history.
4. Isolated grid geometry. Island and weakly interconnected systems with little redundancy margin.
5. Capital and regulatory stress. Rate cases, cost recovery, and the retirement of aging thermal fleets.
6. Hydro and renewable volatility. Capacity that swings with weather rather than demand.

No single utility faces all six. But many face several at once, and that overlap is where structural strain concentrates.

What the Public Data Shows

Every input below comes from public sources that anyone can check: NERC reliability assessments, FERC and SEC filings, Lawrence Berkeley National Laboratory (LBNL) queue data, grid-operator reports, and the rating agencies.

Queues are enormous, and most projects never finish. Proposed generation and storage waiting to connect to the U.S. grid peaked near 2,600 GW at the end of 2023 and still stood at about 2,061 GW at the end of 2025.¹ Of the capacity that requested interconnection between 2000 and 2020, only 13% had reached commercial operation by the end of 2025; 75% had withdrawn.¹ The median wait from request to signed agreement now runs well over three years, and reaching operation takes more than five.¹

Data center demand is rising fast. U.S. data centers used about 4.4% of the nation's electricity in 2023. Department of Energy-commissioned research projects that share at 6.7% to 12% by 2028, roughly a doubling or tripling.²

Texas shows the scale. ERCOT's leadership has cited more than 250 GW of forecast demand from large-load projects, against a record peak of about 85.5 GW.³ Not all of it will be built, but planning for even a fraction of it is unprecedented.

This is today's grid: operating near its physical, financial, and regulatory limits at the same time, with little slack left to absorb the next shock, whatever form it takes.

How the Pressures Map Across the Grid

The same six pressures show up across every major region: CAISO, ERCOT, PJM, MISO, SPP, NYISO, ISO-NE, and the Southeast.

• Catastrophic risk concentrates in the West's wildfire corridors and the Gulf and Atlantic storm coasts.
• Hyperscale load is clustering in Northern Virginia, Texas, the Southeast, and parts of the Midwest.
• Isolated geometry is starkest in Hawaii and in weakly connected systems at the grid's edges.
• Queue backlogs are deepest in PJM, MISO, CAISO, and ERCOT.
• Capital and regulatory stress follows utilities carrying large rebuilds, nuclear cost recovery, or contested rate cases.
• Hydro and renewable volatility is most pronounced in the Pacific Northwest and the Mountain West.

This white paper describes pressures, not verdicts. It does not rank or score individual utilities, and naming a region is not a forecast about any company in it.

The Gap Conventional Monitoring Doesn't Measure

Most utility monitoring works by thresholds. A line sags, a transformer fails, a reserve margin drops, a credit outlook changes, and only then does the system register a problem. Those signals are essential. But they describe components and outputs, not how well the relationships holding the whole system together are actually doing.

A grid can keep meeting its performance targets while those relationships change underneath. That gap—between reported performance and the system's structural state—is what the Standard Coherence Fidelity Layer (SCFL) is designed to measure.

SCFL has been applied to real grid data, including ERCOT stress events such as Winter Storm Uri and 738 outage events from Oak Ridge National Laboratory's EAGLE-I dataset. It is a measurement, not a forecast or an alarm. Its results have not yet been independently replicated, and the research record documents open questions alongside findings. The full body of work is public and organized by domain on this site.

The Bottom Line

The utilities facing these pressures are not isolated stories. They are data points in one national pattern: a grid carrying more simultaneous physical, financial, and digital strain than it was built for, and running out of margin to do it quietly.

The country is not approaching a reckoning for its electrical infrastructure. It is already inside one. Seeing the structural state of the system, and not only its threshold alarms, is where better decisions start.