Across PJM, MISO, and several Southeast balancing authorities, coal plants that were penciled in for retirement are running. Not because the politics shifted back, but because the load math did. Data center buildout, onshoring of energy-intensive manufacturing, and the slower-than-projected ramp of firmed renewables have created a capacity gap that coal, already depreciated and still connected, is filling by default.

Where the Lights Are Still On

The operating picture is not uniform. Roughly 170 GW of coal capacity remained on the U.S. grid entering 2024, down from a peak above 300 GW, but retirements have decelerated noticeably since 2022. PJM has granted retirement deferrals to plants that failed to secure deactivation approval, citing reliability concerns. In MISO, several utilities have filed to extend operating licenses on units originally slated to close in 2025 and 2026. Internationally, Germany restarted idled hard coal capacity during the 2022 energy crisis and kept units available through 2023. Poland has legally committed to coal through 2049, though that date is under revision.

India and Indonesia present a structurally different picture. Both are commissioning new coal capacity, not extending old. India added approximately 15 GW of coal-fired generation between 2022 and 2024. The IEA’s observation that global coal power generation hit record highs in 2023 is the detail that European energy policy tends to bracket out of the conversation.

The Policy Logic, Such As It Is

Regulators operating inside organized markets face a structural bind. Capacity markets were designed to price reliability, but they were not designed for a transition scenario where thermal retirements accelerate faster than storage and transmission can absorb the gap. The result is what FERC has described, cautiously, as a reliability gap risk in portions of the Eastern Interconnection through the late 2020s.

The policy response is essentially a managed delay. Utilities receive cost-of-service recovery in some jurisdictions to keep units available even when they are uneconomic on an energy basis. This is not a reversal of decarbonization targets on paper. In practice, it extends the operational runway of assets that were supposed to be off the grid.

The emissions accounting on extended coal operations versus the counterfactual is genuinely contested. Running an old coal unit at low capacity factor to provide grid insurance is a different calculus than running it as baseload. The structural tension between reliability mandates and emissions trajectories has not been resolved; it has been deferred.

What the Asset Picture Reflects

Thermal generation assets trading below replacement cost, which most coal plants do, carry a specific kind of optionality. The optionality is not that coal makes a structural comeback. It is that the transition timeline has a range of outcomes, and fully depreciated assets with existing grid interconnection and fuel supply agreements sit inside that range with low carrying cost.

Several infrastructure-focused capital allocators have noted, without publicity, that distressed coal plant acquisitions in PJM corridors have drawn interest from buyers who are not coal operators at all. The thesis is site control, interconnection queue position, and existing transmission access, assets that take years to replicate from greenfield.

The Operator Read

The structural observation for operators and allocators is not that coal is resilient as a fuel story. It is that the grid transition has dependencies that are running behind schedule, and that creates an environment where assets positioned as reliability infrastructure rather than generation assets carry different logic than the headline narrative suggests. The capacity market reform conversations happening inside FERC and at the state commission level will determine how long that logic holds. Operators watching those dockets are watching the real timeline.