American Electric Power, a utility serving more than 5.6 million customers across 11 states, has warned that it may abandon interconnection agreements with two regional power grids, citing years-long delays in getting new generation resources connected to the system. The warning, which AEP executives have raised in public forums and regulatory discussions in recent months, lands at a moment when the nation’s grid operators are buckling under a surge of new electricity demand driven largely by AI-focused data centers.
AEP has not filed formal withdrawal notices with the Federal Energy Regulatory Commission, and the company has not publicly identified which two grids are the targets. But the utility operates across both PJM Interconnection, the nation’s largest regional grid operator covering 13 states and the District of Columbia, and the Southwest Power Pool, which coordinates the grid across 14 central U.S. states. Both systems are grappling with interconnection backlogs that have ballooned in recent years.
The frustration is not abstract. According to data from Lawrence Berkeley National Laboratory, the average time from interconnection request to commercial operation in the United States has stretched beyond five years, and more than 2,600 projects representing roughly 300 gigawatts of capacity were sitting in PJM’s queue alone as of late 2025. For a utility like AEP, which has publicly discussed billions of dollars in planned capital spending to meet surging load from data centers and industrial customers, those delays threaten to strand investments and break commitments to major customers.
FERC’s nationwide overhaul
Federal regulators have not been idle. FERC finalized Order No. 2023, a sweeping overhaul of the interconnection process that replaces the old first-come, first-served queue model with a cluster-based study process. Instead of evaluating one project at a time in strict queue order, transmission providers now assess groups of projects together, with higher financial-readiness requirements designed to weed out speculative proposals that clog the pipeline.
The order also imposes firm study deadlines on grid operators and attaches financial penalties when those deadlines slip. Transmission providers must complete feasibility, system impact, and facilities studies within specified timeframes. Miss those windows, and the penalties scale with the duration and severity of the delay. The intent is blunt: process interconnection requests faster or pay for the failure.
Whether those penalties will actually change behavior remains an open question. No grid operator has yet been publicly penalized under the new framework, and the mechanics of enforcement, including how FERC will assign blame when delays stem from disputed study assumptions or cascading project withdrawals, are untested. The rule is on the books, but its teeth have not been shown.
PJM under the spotlight
FERC has also singled out PJM for targeted action. In a separate directive, the commission ordered PJM to draft new rules addressing co-location arrangements, where data centers or large industrial loads sit physically adjacent to power plants and draw electricity before it ever reaches the broader grid. These setups have raised reliability and market-fairness concerns because the load effectively bypasses the transmission system and its associated costs.
The same order requires PJM to file reports on faster capacity additions and demand-side tools, with a specific focus on improving load forecasting and integrating demand flexibility into planning and market operations. The directive is a clear signal that FERC views the data-center demand wave not as a distant planning exercise but as an immediate operational challenge for the grid’s largest operator.
For AEP, which serves a substantial portion of its customers within PJM’s footprint, the outcome of these proceedings matters directly. If PJM can accelerate its queue processing and accommodate large, fast-moving loads without reliability trade-offs, AEP’s interconnection timelines could shorten meaningfully. If the reforms stall in implementation, the utility’s threat to walk away becomes more than posturing.
The data-center pressure cooker
The backdrop to all of this is a dramatic reversal in U.S. electricity demand trends. After roughly a decade of flat or declining load growth, utilities across the country are projecting sharp increases driven by data centers, manufacturing reshoring, and electrification. The U.S. Energy Information Administration has noted that data centers alone could account for a significant share of new electricity consumption through the end of the decade, and several utilities, AEP among them, have flagged data-center load as the single largest driver of their revised demand forecasts.
AEP’s service territory is a particular hotspot. Northern Virginia, which falls partly within PJM and adjacent to AEP’s operating companies, already hosts the densest concentration of data centers on the planet. But the buildout is spreading into Ohio, Indiana, and other parts of AEP’s footprint, where land is cheaper and power is more available, at least on paper. The catch is that “available” power means little if the generation cannot physically connect to the grid for half a decade.
That mismatch between demand urgency and supply timelines is what makes AEP’s threat significant, even without formal filings. Large utilities do not publicly float the idea of abandoning grid agreements without serious internal deliberation. The signal is aimed at multiple audiences simultaneously: grid operators who control queue processing, regulators who set the rules, investors who fund the capital plans, and large customers who need certainty about when power will actually arrive.
What AEP’s move signals for the broader grid
AEP is not the only utility wrestling with this problem. Dominion Energy, Duke Energy, and Georgia Power have all disclosed surging data-center demand in their service territories, and each faces its own version of the interconnection bottleneck. But AEP’s willingness to publicly threaten withdrawal from grid agreements raises the stakes for the entire system. If one of the country’s largest utilities concludes that the interconnection process is too broken to rely on, it could accelerate a shift toward behind-the-meter generation, where large customers build dedicated power plants that never touch the grid at all.
That outcome would carry its own risks. Behind-the-meter generation sidesteps the queue but also sidesteps the shared reliability framework that keeps the lights on for everyone else. It can shift costs onto remaining ratepayers and reduce the pool of resources available to grid operators during emergencies. FERC’s co-location directive to PJM is, in part, an attempt to get ahead of that trend by creating rules that accommodate large dedicated loads without fragmenting the grid.
For now, the most grounded reading of the situation is this: federal regulators have acknowledged the interconnection crisis and begun restructuring the rules, but the reforms are untested and the backlog is enormous. AEP’s threat is a measure of how severe the pressure has become, not a resolution. The practical question for the months ahead is whether FERC’s penalty framework and PJM’s new planning obligations can shrink queue timelines fast enough to keep utilities and their largest customers from finding workarounds that leave the rest of the grid worse off.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
