H1 2026 U.S. Power Market Update: Regional Responses to Rising Electricity Demand

In H1 2026, one theme came to define U.S. power markets: reallocation. As AI and data-center demand outpaced a delivery-constrained grid, the year turned less on adding capacity than on redirecting it — who pays for reliability, what provides it, and where the grid gets built next.

PJM

A capacity shortfall, a record grid bill, and a $67B merger

PJM's capacity market reached its limits in H1 2026. The 2027/28 capacity auction cleared at its price cap yet still recorded the first RTO-wide capacity shortfall, driven by rising demand rather than generation retirements. As a result, the cost of maintaining reliability is increasingly shifting beyond the capacity market and into transmission investments and large-load cost recovery.

Figure 1. PJM Base Residual Auction clearing price, $/MW-day

KEY DEVELOPMENTS:

• The 2027/28 Base Residual Auction (results Dec 2025) cleared at its $333.44/MW-day cap yet still fell 6,623 MW short — PJM's first RTO-wide shortfall, at a record-low ~15% reserve margin. Forecast peak load for 2027/28 rose to about 164.6 GW — up ~5,250 MW from ~159.3 GW in the prior auction — and PJM attributes roughly 5,100 MW of that increase to data centers; total cleared capacity payments reach about $16.4B.

  
  

• To build out of the constraint, the Board approved an $11.8B RTEP — its largest ever — including ~$4.8B to Dominion (its total award across multiple projects), led by a 525-kV underground HVDC line into Loudoun County and a new 765-kV line in Ohio. Socializing those costs faces state pushback.

  
  

• On June 18, 2026, FERC issued §206 show-cause orders directing all six FERC-jurisdictional RTOs to revise their large-load tariffs within 60 days; PJM's behind-the-meter / co-located rewrite came earlier, in FERC's Dec. 18, 2025 order — together settling who pays for, and who can bypass, the grid.

  
  

• On May 18, 2026, NextEra and Dominion agreed to a $67B all-stock combination (NextEra the acquirer) to fund AI-era load — but a $150M securities settlement over FPL political-spending allegations clouds approval at the Virginia SCC.
  

ERCOT

Solar overtakes coal, storage scales, and data centers must pay

Figure 2. ERCOT 2026 generation, solar vs. coal, TWh

ERCOT offers an early blueprint for the AI-load era. While utility-scale solar is projected to surpass coal generation for the first time in 2026 (78 vs. 60 TWh), natural gas remains the dominant source of electricity, supplying more than 40% of all generation. The more significant structural shift is SB6, which requires large-load customers to fund their own interconnection costs.

KEY DEVELOPMENTS:

• Utility-scale solar generation is set to outproduce coal for the first time in 2026 per EIA STEO (78 vs. 60 TWh). Solar's share has climbed from 4% of the mix in 2021 to ~12% by 2025, and Texas now accounts for roughly 40% of all U.S. solar additions.

• Solar growth has helped make ERCOT the largest U.S. battery market: ~20 GW (June 2026, up from ~14 GW in 2025), tracking toward ~37 GW by 2027 — with storage repeatedly carrying the evening peak and damping the price spikes that defined past summers.

  
  

• SB 6 (enacted June 2025) makes loads ≥75 MW fund their own interconnection. The PUCT's draft rule would require a $50,000/MW security and 100% of interconnection costs upfront, and on June 18, 2026 — the same day as FERC's §206 order — the PUCT approved ERCOT's batched "Batch Zero" study process. Ahead of FERC’s order, Texas assigned large-load interconnection costs directly to the customer, unrecoverable through regulated rates.

  
  

• Real-Time Co-Optimization plus Batteries (RTC+B), ERCOT's biggest market redesign since 2010, went live in December 2025 — co-optimizing energy and ancillaries every five minutes and modeling battery state-of-charge. This has reshaped how batteries earn revenues and how the system manages evening peaks.

MISO & CAISO

Two divergent bets: coal economics vs. a solar-led grid

Figure 3. MISO spreads, $/MWh (Jan–Apr 2026)

Same demand pressure, opposite responses. MISO's coal fleet remained economically competitive amid elevated winter natural gas prices, while CAISO prioritized storage and grid-enhancing technologies over new transmission. Neither approach reflects technology preference; both are shaped by regional fuel economics and existing infrastructure.

KEY DEVELOPMENTS:

MISO — coal holds on economics

• Coal stayed the most profitable fuel — a ~$28/MWh dark spread vs. ~$9 for gas (Jan–Apr 2026 avg) — after Winter Storm Fern drove gas-fired generation cost to a peak ~$549/MWh in late January (from ~$25/MWh a week earlier); on-site fuel is a physical hedge against winter gas volatility.

• The 2026/27 auction was two-speed: annualized prices fell ~42% to ~$126/MW-day on a 4.6 GW surplus, but summer (Zones 1–7) cleared $424.30 — risk now concentrates in summer

• The Board approved the $21.8B Tranche 2.1 portfolio (3,631 mi of 765/345-kV); five states have challenged the cost allocation at FERC.

CAISO — a leaner, load-serving grid

• Utility-scale solar surpassed gas for the first time over Jan–May 2026, beating it on 82% of days (up from 21% in 2024–25) — making daytime-to-evening storage increasingly critical.

• The $6.7B transmission plan (38 projects) was driven increasingly by load growth (upgrades near fast-growing demand centers) rather than connecting remote renewables.

• CAISO canceled the Serrano–Del Amo– Mesa 500-kV line after its estimate rose from $1.2B to ~$5.0B, choosing an alternative to reconductor smaller lines + targeted energy storage.

  
  
  
  

Other Notable Developments

NYISO

• Large-load interconnection requests jumped from 1,045 MW (2022) to 12,670 MW by May 2026

• Power Trends 2026 warns margins are narrowing as retirements outpace supply.

ISO-NE

• FCA 18 (2027/28) cleared at $3.58/kW-month — about $1 higher than the prior four auctions.

• FERC approved Phase 1 of ISO-NE's Capacity Auction Reforms in March 2026 — a shift to a prompt, ~month-ahead auction. Winter fuel-security accreditation is Phase 2, not yet filed.

SPP

• On April 1, 2026, SPP extended its RTO into the Western Interconnection — the first U.S. RTO to span both the Eastern and Western Interconnections.

• Nine western utilities joined, gaining access to a broader resource portfolio and coordinated reliability services across a wider footprint.

Southeast

• No organized market: utilities recover costs via IRPs, with large loads increasingly placed on minimum-bill and take-or-pay style tariffs.

• Georgia Power (Dec 2025) adds ~10 GW (~$16B), ~80% data-center-driven; loads above 100 MW face 15-year minimum-bill deals.

• Duke’s Carolinas plan carries ~14 GW of new build; its advanced-development large-load pipeline (mostly data centers) reached 5.6 GW across 38 projects, up from 1.4 GW in 2023.

• TVA’s draft 2026 IRP weighs 7–26 GW of new gas by 2040; its board voted to keep Cumberland and Kingston coal units running past prior retirement dates.

Nuclear

• Hyperscalers have announced ~13 GW of nuclear, but ~6.9 GW is likely to materialize by the early 2030s (Carnegie, 2026).

• Meta’s 6.6 GW package (Jan 2026) pairs existing plants with SMRs, majority into PJM.

• Nuclear PPAs near ~$100–115/MWh sit above merchant; restarts still face timing risk, with Crane (TMI-1) targeting 2H 2027.

Takeaways

1. Interconnection is the binding constraint. FERC's §206 order rewrites large-load rules, and an executed Interconnection Agreement (IA) is now the gate to financing — no IA, no non-recourse debt. With grid access scarce, some large loads go behind-the-meter to secure power while staying grid-connected.

2. The ITC begin-construction window has now closed. Wind and solar that missed the July 4 start must now reach service by end-2027 — so underwriting shifts to safe-harbor proof and COD certainty. A June 6 ruling restored the 5% safe harbor (subject to appeal). Overall, solar and wind have now overtaken coal generation.

3. Demand remains a thesis, not a given. In the U.S., the majority of data center capacity is still in the announced or early development stage with a fraction of projects operating or under construction. Sightline Climate estimates that ~30–50% of planned capacity could be delayed or canceled. The queue signals intent, not committed demand.

Looking Ahead

Jul 2026	PJM's 2028/29 auction results (Jul 14) test whether the shortfall repeats.
Aug 2026	Six RTOs file show-cause and tariff responses (Aug 17; adequacy reports Jul 20), shaping who funds interconnection.
Dec 2026	PUCT finalizes Texas SB6; PJM’s 2029/30 auction runs; MISO’s $21.8B Tranche 2.1 allocation stays contested at FERC.
Winter 2026–27	NYISO and ISO-NE capacity redesigns face their first winter. Tight margins would invite NERC alerts and fresh market reform. FEOC content step-up — foreign-content thresholds ratchet (~40→45%); sourcing becomes a credit-eligibility gate.
H2 2027	Crane (TMI-1) restarts target 2H 2027; Palisades already slipped. On-schedule nuclear sets the price floor for firm AI load. NextEra–Dominion close — $67B merger (largest US regulated utility) pending FERC / NRC / state review — watch the FERC filing as the trigger.
End-2027	U.S. projects without a durable BOC position face a 2027 ITC in-service deadline — even though only a fraction of announced data-center capacity is under construction.

Simcore H1 2026 Power Market Update.pdf

Download PDF • 1.13MB