Coal isn't retiring as planned. The summer power-burn consensus is over-stated.

The retirement schedule consensus baked into summer power-burn forecasts assumes coal comes off as planned. The data through Q1 2026 says it isn't.

What's visible by today:

• 2025 coal retirements came in at 2.6 GW, substantially short of the 8.5 GW plan. DOE has issued 43+ Section 202(c) emergency orders since May 2025, effectively stalling roughly 4.4 GW of planned coal retirement through April 2026.
• Q1 2026 saw zero major coal retirements complete. Centralia, Craig, and J.H. Campbell all kept online under fresh DOE 202(c) orders.
• Today — Colorado's governor backs a bill to delay the 207 MW coal-burning unit at Ray Nixon to 2032. February's TVA board vote extended Kingston and Cumberland (~3.8 GW combined) beyond plan.
• MISO coal dark spreads averaged $28/MWh in the first four months of 2026, 39% above the same period 2025. Despite gas sub-$3 Henry Hub, coal remains competitive in the merit order.

Collectively, this suggests a ceiling on the expected gas burn lift from coal-to-gas switching into summer.

The structural mechanism: on standard heat-rate math, if 3 GW of scheduled-to-retire coal stays online through summer at typical 50–70% capacity factors, that displaces roughly 0.25–0.4 Bcf/d of gas burn. If slippage scales to 5–6 GW by mid-summer through additional 202(c) renewals and state-level actions, the displacement reaches 0.4–0.7 Bcf/d.

The EIA April STEO and broker-desk consensus for summer 2026 power burn implicitly assume the retirement schedule holds. The data says it doesn't. Summer power burn could under-deliver the 40 Bcf/d-plus consensus by 0.3–0.6 Bcf/d. That gap removes much of the implied upside in the summer 2026 power-burn story, particularly at the upper end of the band.

What invalidates it: a federal court ruling vacating the 202(c) order framework before summer, a coal-plant forced-outage cluster, or a sudden retirement-acceleration policy reversal. None visible today.