Which coal plants actually retire in 2026?

A unit-by-unit assessment reveals most of the scheduled retirements won't happen this year

In Part 1, I argued that coal retirements are no longer primarily an economic decision, but a reliability one. Whether a plant can retire now depends less on its average utilization and more on whether the system can replace its output during winter peaks, when gas generators compete with LDCs and LNG for both volume and deliverability.

Applying that framework screens out most scheduled 2026 retirements: of the more than 7 GW currently slated to retire, I expect only 3–5 GW to actually come offline on schedule. The balance is likely to be delayed until replacement capacity catches up.

Below, I evaluate each plant unit by unit, grouping them by retirement likelihood. For each, I focus on winter peaking behavior, nearby replacement capacity, and fuel deliverability constraints — the factors that now determine whether retirement is realistic rather than merely planned.

Replacement capacity isn’t the whole story

Gas CC utilization at units within 100 miles¹ of the coal plants scheduled to retire this year is generally lower than at the units that either the DOE or utilities delayed retirement. In all cases, nearby gas steam and peaking plants have significant spare capacity.

However, relying on gas steam and peaking units introduces a different constraint: gas pipeline capacity. With lower utilization rates, these units are much less likely to have contracted for pipeline capacity. On peak days, these units compete with nearby CCs for gas supplies, even when power capacity is sufficient. And retiring coal plants run hardest in the winter, when LDCs are likely to fully utilize their contracted capacity.

Figure 1 | Capacity factors for retiring coal plants and nearby gas capacity

Potential January constraints

Furthermore, capacity factors indicate only whether spare capacity is available, not its magnitude. For example, a 30% utilized 500 MW CC cannot offset the generation from a 30% utilized 1,000 MW coal unit. In January, when these coal plants run hardest, only one DOE-delayed plant has materially more spare CC capacity available than the coal generation retiring. Plants where utilities have delayed retirements also feature tighter January CC availability. Some scheduled retirements in green also appear to be at risk unless new baseload capacity is added.

Even this analysis understates potential stress, as the analysis reflects the peak month rather than the peak day. Peak-day power capacity is very likely available from steam and peaking units, but gas pipeline capacity is in even shorter supply than in January on average. Plants with dispatchable replacement capacity under construction are far more likely to proceed than those relying on under-utilized capacity elsewhere.

Figure 2 | January coal generation and nearby CC spare capacity

Plant-level outlook

Likely to retire as scheduled

Cumberland (TN): Replacement CC under construction, so retirement likely proceeds once the CC is online; any risk to retirement is associated with CC construction.

GREC (OK): CT under construction and coal unit sporadically utilized, so CT likely sufficient.

Sherburne County (MN): No on-site replacement capacity planned, but modest utilization and nearby CC availability.

Comanche 2 (CO): Delayed for Comanche 3 outage. Once Comanche 3 returns mid-year, retirement likely proceeds, but net negative for gas demand given Comanche 3’s larger size.

Could retire this year, but less certain

F B Culley 2 (IN): Small (~100 MW) size gives it best odds of DOE’s order not being extended, especially as new MISO capacity comes online; minimal gas market impact either way.

R M Schahfer 17 and 18 (IN): Low utilization and nearby CC capacity available, but no major change since DOE delay.

South Oak Creek 7 and 8 (WI): Small units not running very hard; nearby CC capacity available but no material change since utility delay.

Kingston 7, 8, and 9 (TN): EIA lists three units retiring this year and remaining six in 2027, but replacement CC not online until 2027. Possible that all nine retire next year, but small size and nearby CC availability may allow some to proceed this year.

Likely delayed

Craig 1 (CO): Isolated in northwest Colorado; replacement plan weighted to wind and solar, with little CC availability nearby.

Centralia (WA): TransAlta recently announced 2028 gas conversion rather than retire it, and the retirement may wait until that happens, currently targeting 2028.

J H Campbell (MI): Already delayed multiple times by DOE; MISO capacity additions could facilitate the retirement of smaller units before the ~900 MW unit 3.

Figure 3 | Coal retirement screening

Gas implications

Altogether, I expect 3.0-5.2 GW of retirements in 2026, adding 200-400 MMcfd of gas demand. This compares with 7.3 GW in the most recent EIA 860M, including the DOE-delayed plants still listed as planned for 2025.

Regardless, whereas in the past coal retirements consistently surprised to the upside, going forward the schedule looks likely to be pushed to the right — no matter who is elected in 2028. For the gas industry, this slowdown in retirements means that gas’s share of fossil generation is likely to be relatively flat, as modest retirements offset declining coal-gas displacement as gas prices rise.

Figure 4 | Gas impact from 2026 coal retirements

This slower pace — roughly half of what’s currently planned — is modestly bearish for near-term gas demand. But stretched over a longer timeline, it’s likely healthier for the gas industry. Growing reliance on gas-fired generation, especially before commensurate new pipeline capacity is developed, increases the risk of gas price spikes to $8+/MMBtu — or worse, reliability failures. The same winter deliverability constraints that delay coal retirements also cap how much incremental gas demand the power sector can absorb without destabilizing prices.

While very high gas prices may screen positive for gas E&Ps and midstream operators, in reality the industry should be — and mostly is — terrified of them. Gas price spikes generally and winter reliability issues in particular not only erode utilities’ confidence in shifting toward gas but also, in extreme cases, risk precipitating curtailment of liquefaction capacity additions or even of currently operational liquefaction. Given the scale of US LNG investment underway, maintaining moderate prices is important not just for US LNG offtakers but for US E&Ps and the rest of the US gas value chain.

1. Admittedly an arbitrary threshold, but nonetheless useful analytically. In reality, transmission constraints are highly localized.