Same question, same model, same tools: the two reports

One run per harness, the run closest to its three-run average. Reports are shown as the harnesses wrote them. Scroll each column.

Claude Code

Which S&P 500 utilities actually benefit from data-center demand (as of November 10, 2026)

Twenty of the 31 have contracted data-center load large enough to rank. Southern and AEP come out on top. Southern combines 21 GW of contracts that carry minimum bills with a very small remaining equity need. AEP has by far the most contracted load, though its own figure fails the grid-operator check and has to be cut heavily.

Most figures come from each company's Q3-2026 call, deck and 10-Q, retrieved through Samaya's summaries; I did not open the primary documents. Market caps are November 10 closes.

Step 0: Who doesn't benefit (11 not ranked)

Group Why it isn't ranked
American Water, Atmos No data-center load or data-center capex found in their latest calls, decks or 10-Qs.
Con Ed, Eversource, Edison International Data-center load is essentially nil. New York and New England grid forecasts show roughly 0.1–2.6 GW of large load in total. Edison also carries Eaton fire liability.
PSEG No contracted load, only about 11.8 GW of inquiries, which management calls mostly small edge sites. Any upside sits in its unregulated nuclear plants.
AES Agreed to be taken private at $15.00 cash; the stock is at $14.92, so shareholders get no upside.
Dominion 12.2 GW under signed contracts in Virginia, the strongest data-center zone in PJM. But it is being bought by NextEra in an all-stock deal (0.8138 NEE shares per share; Virginia hearings began in November). The stock now trades on NextEra's price.
Constellation, Vistra, NRG Power producers, not regulated utilities, so they're listed separately below. Their deals sell power from existing plants, which raises margins rather than a regulated asset base. None plans to issue equity.

Step 1: Contracted load and funded capex plan

Company Contracted data-center load Capex plan
AEP 84 GW "contracted" through 2031: 46 GW Texas (letters of agreement, not firm contracts), 31 GW PJM, 7 GW SPP. Only about 35 GW is firm take-or-pay. $100B, 2027–31; rate base growing over 13% a year
Southern 21 GW across 33 projects, all with minimum bills, 31 under construction. Its 10-Q says about 20 GW. $81B through 2030, could reach about $100B
PPL 11.4 GW under signed contracts in Pennsylvania (wires-only there) $25B, 2027–30
Sempra / Oncor Texas: 45 GW submitted, about 24 GW classified as firm "base" load by the grid operator (10 GW already connected). Not signed contracts. Oncor $47.5B and Sempra about $65B, 2026–30
FirstEnergy 7.9 GW; first sites come online around 2030 $36B, 2026–30
Pinnacle West 5.8 GW committed, including about 1.8 GW for TSMC's chip plant; 4.9 GW not yet served About $10.4B, 2025–28. TSMC-related capex excluded.
CenterPoint About 10 GW firm and 2 GW still under study, per the Texas grid operator's classification. Not signed contracts. $38.7B, 2026–30
NiSource 3.9 GW signed (Amazon, Alphabet, Microsoft). Management also says "about 5 GW"; the two aren't reconciled. $29B, 2027–31 ($40B through 2033)
Alliant 3.4 GW (QTS, Google, Meta) $15.9B, 2027–30
Exelon About 11 GW "high-probability", 45% backed by collateralized agreements. Only about 10% online by 2028. $41.7B, 2026–29; wires-only
Evergy About 2.6 GW (Google, Meta, Digital Realty) $22.5B, 2026–30
WEC About 4.5 GW forecast under its large-customer tariff: Microsoft 3.2 GW, Vantage/Oracle 1.3 GW. Not stated as contracted MW. $47.5B, 2027–31
Entergy No total disclosed. I estimate about 7 GW from the Meta, Google, AWS and other contracts. $55B, 2027–30
Ameren 2.8 GW (Google and Amazon sites in Missouri) $38.1B, 2027–31
DTE 2.4 GW approved by regulators (Oracle 1.4 GW, Google 1.0 GW) $44.2B, 2027–31
Duke About 7.8 GW signed, plus 15.4 GW late-stage $103B, 2026–30
Xcel About 2 GW (Google, QTS, Fermi) $70B, 2027–31
CMS One site over 1 GW, not yet in the capex plan $24.1B, 2026–30
NextEra Florida utility: nothing signed (targets 8 GW by 2032). Generation arm: about 1.6 GW (Google at Duane Arnold, Point Beach). About $94.5B, rest of 2026–30
PG&E 0.64 GW with construction agreements; 12.9 GW pipeline Guidance beyond 2027 withdrawn for a strategic review

Step 2: Load checked against grid-operator forecasts

The haircut column is the share of claimed load I kept: 1.0 if the claim is consistent with the regional forecast, 0.75 if it is partly consistent or the timing is at risk, and 0.5 if it exceeds what the region's forecast can hold.

Company What the regional forecast shows Haircut
AEP Fails. Its 46 GW of Texas agreements compares with 66 GW of firm load classified for all of ERCOT. Its 31 GW in PJM compares with about 34 GW of data-center growth PJM accepted for the whole region by 2030; PJM cut the AEP zone by 3.7 GW for 2031. AEP's own 2030 submission to PJM is 19.6 GW. 0.5, applied to the 35 GW firm subset
Southern No grid operator in its region. Against the Georgia regulator's plan (about 8.5 GW of growth by 2031; 9.9 GW of new resources certified), Georgia's 13.5 GW runs ahead of the plan. Minimum bills and construction underway support it. 0.75
PPL PJM has its zone as the fastest-growing (6.4% a year); its 12.8 GW submission for 2030 lines up. PPL, FirstEnergy and AES Ohio together cut near-term requests by 6.5 GW. 0.75
FirstEnergy PJM cut its Ohio zone 8.1% for 2031. Its submissions total about 5 GW by 2029, and its load only starts around 2030. 0.75
Exelon PJM added only 3.7 GW for ComEd by 2031, against about 9 GW claimed there. 0.5
Oncor / CenterPoint Both claims sit inside ERCOT's firm classification, but the grid operator's audit is due in December 2026. Oncor, CenterPoint and AEP Texas together claim more than the 66 GW of firm load ERCOT has classified across its whole grid. Independent analysts expect only about 33 GW of Texas data centers by 2030. 0.75
NextEra (Florida) The Florida regional forecast has about 3 GW of data centers statewide over 10 years, versus FPL's 8 GW goal. I excluded the unsigned 8 GW entirely. Unsigned load excluded
PG&E California's energy commission forecasts 5.8 GW for the whole state by 2040, against 14.7 GW of requests at PG&E. Only the firm 0.64 GW is used. 1.0 on the firm figure
MISO and SPP utilities (Ameren, WEC, Alliant, NiSource, DTE, CMS, Evergy, Xcel, Entergy) Consistent. MISO expects about 20 GW of data centers by 2030 and lists Ameren at 2.9 GW, Alliant at 3.4 GW and WEC at 3.9 GW. SPP shows 9.6 GW under signed service agreements region-wide. 1.0 (Entergy 0.75, since it gives no total)
Duke, Pinnacle West Consistent with their state resource plans. 1.0

The wider evidence points the same way. PJM accepted only about 57% of what utilities submitted for 2030. NERC, Grid Strategies and FERC (June 2026 orders) all flag projects being counted more than once.

Step 3: New equity need

This is planned new common and equity-like issuance over each plan period, as a share of today's market cap, per year.

  • Lowest (about 0.5% a year or less):
  • PG&E: 0%, needs already covered.
  • Sempra: about 0.2%; its KKR stake sale replaces common equity.
  • Southern: about 0.5%; only about $2B left after $5.3B of forward sales settled.
  • Moderate (about 1.3–2.5% a year):
  • NextEra: 1.3%
  • AEP: 1.5%; most of its $5.2B is already sold forward.
  • FirstEnergy: 1.5%
  • Exelon: 1.8%
  • NiSource: about 2.1%
  • Entergy: 2.4%
  • Duke: 2.5%; $10B over 2027–30.
  • Heaviest (about 2.7–3.7% a year):
  • WEC: 2.7%, plus $4–5B of hybrid securities
  • PPL: 2.8%
  • Pinnacle West: 2.9%; likely understated, since TSMC capex isn't in the plan.
  • DTE, Xcel: 3.1% each
  • Ameren, Evergy: 3.2% each
  • CMS: 3.5%; the 1 GW site isn't in the plan yet, so more is likely.
  • Alliant: 3.6%
  • CenterPoint: 3.7%

Step 4: Ranking

The load score is credible GW per $10B of market cap. It applies the step 2 haircut, then halves the result for wires-only utilities, which earn on transmission spending but don't own the new generation. The composite weights the load rank at 60% and the equity rank at 40%.

Rank Company Credible GW per $10B Equity % per year Main risk
1 AEP 2.5 1.5 Its own headline load figure fails the regional check
2 Southern 1.5 0.5 Rates frozen in Georgia and Alabama through 2029
3 NiSource 1.8 2.1 Gas plant delays; Microsoft contract still needs approval
4 Pinnacle West 3.8 2.9 Includes TSMC; rate case pending; plan runs only to 2028
5 PPL 1.6 2.8 Wires-only in Pennsylvania; near-term cuts
6 FirstEnergy 1.1 1.5 Load arrives in 2030 or later; mostly wires-only
7 Sempra 0.8 0.2 Texas audit; California drag
8 WEC 1.2 2.7 Wisconsin regulator cut its allowed return to 9.5%
9 Alliant 1.8 3.6 Heaviest dilution among those with solid load
10 Entergy 1.0 2.4 No total disclosed; Meta approvals pending
11 Evergy 1.3 3.2 Kansas and Missouri regulatory decisions pending
12 PG&E 0.2 0 Wildfire risk; strategic review
13 CenterPoint 1.4 3.7 Texas audit; highest dilution
14 Exelon 0.6 1.8 Slow ramp; wires-only
15 Duke 0.8 2.5 Data centers are small relative to its size
16 Ameren 0.9 3.2 Construction timing
17 NextEra 0.1 1.3 Florida utility has nothing signed yet
18 DTE 0.8 3.1 Rate case due February 2027
19 Xcel 0.4 3.1 Most upside not in the plan; negative Moody's outlook
20 CMS 0.5 3.5 Site not in the plan; customer unnamed

Power producers (separate list): 1. Constellation: about 4.6 GW signed with named hyperscalers (Microsoft, Meta, Amazon, Google), plus a 2.7 GW Google load-following deal, and a $5B buyback. 2. Vistra: about 4.4 GW signed (AWS, Meta, Microsoft), with at least $1B a year of buybacks. 3. NRG: nothing signed yet.

What could change the ranking

  • AEP's lead depends on my 0.5 haircut. Its own 84 GW figure would put it far ahead, and a deeper cut would let Southern overtake it.
  • The haircut factors and the 0.5 wires-only discount are my own judgment calls.
  • Pinnacle West excluding TSMC: using roughly 3 GW of data-center load instead of 4.9 GW still leaves it in the top five.
  • Disclosures that don't reconcile:
  • Southern: 21 vs 20 GW
  • NiSource: 3.9 vs 5 GW
  • Duke: 7.6 vs 7.8 GW
  • Oncor: inconsistent figures within its own deck
  • Unconfirmed item: a reported AEP–Google fuel-cell deal placed in Wyoming, outside AEP's territory.
  • Events that could reorder things:
  • The Texas grid operator's audit, due December 10.
  • PJM's 2027 load forecast, due in January.
  • The February 2027 plan updates from Southern, CMS and PSEG.

The scoring script is at /tmp/cc_work/scripts/rank.py if you want to change the haircuts or the 60/40 weights.

Sources: company Q3-2026 calls, decks and 10-Qs via Samaya. Grid forecasts: PJM 2026 Load Report, PJM Inside Lines, Utility Dive on the PJM forecast, ERCOT preliminary forecast, ERCOT adjustment notice, NERC 2025 LTRA, Grid Strategies LTRA review. Merger: Switchgear Magazine, VPM.

Gen 3

S&P 500 Utilities: Contracted Data-Center Load vs. Grid Forecasts and Equity Financing Need

Executive Summary

  • Contracted load is highly concentrated. 17 of 31 constituents disclose contracted DC load, totaling ~178 GW; AEP alone holds 84 GW (Transcript) 1, or 47%. 12 electric utilities and the 2 non-electric names disclose none.
  • Aggregate equity need is about $115.5B across 30 constituents (AES excluded) over their plan windows. SO, DUK, AEP, ATO, XEL and NEE account for $55.6B (48%). Seven expect no new common equity: CEG, EIX, NRG, PCG, PEG, SRE, VST.
  • Top composite ranks: CEG, NI, VST, NRG, FE, LNT, PPL, WEC. Among regulated names, NiSource, FirstEnergy and Alliant pair a 12 to 19% share of their regional forecast with equity plans of $2.0 to $2.7B.
  • Largest books, heaviest dilution: AEP, SO and DUK rank top 3 on load share but 28th to 30th on equity need (~$9.7B, ~$12B, ~$10B), landing at composite 13, 15 and 12.
  • Biggest narrative gaps: NEE, SRE, CNP, ETR and PCG lead with DC narratives and no quantified signed load; XEL (0.10) and EXC (0.06) have the weakest contracted-to-pipeline ratios; Dominion has only ~22% of its 54.9 GW headline under ESAs. Best ratios: LNT 1.13, NI 0.98, DUK 0.51.

Ranking method: Composite = rank on contracted DC load as a share of the regional grid forecast (D1) + rank on estimated equity need (D2), equal weights. Utilities with disclosed contracted load (17) rank above electric utilities without it (12), with water and gas-only names last. Without that gate, EIX, PCG, PEG and SRE (rank-sum 27.5, zero equity) would rank ahead of DUK, SO and AEP. Equity need uses management's equity plan where disclosed; otherwise Equity Need = 5-yr Capex - (5-yr Operating Cash Flow - Dividends) - Planned Debt Capacity using Visible Alpha SD line items. The formula proved unstable where only one year of debt capacity is disclosed (AEP $2.6B to $62.6B against a ~$9.7B company plan), so it is used as a cross-check.

Utility Data-Center Exposure and Financing Table

UtilityTickerGrid OperatorContracted DC Load (GW)Pipeline DC Load (GW)5-yr Capex Plan ($B)Equity Need ($B)EPS Growth GuideRankSource Citations
Constellation EnergyCEGPJM (also ISO-NE, NYISO, ERCOT, MISO)8.5 (generator PPAs, mostly existing fleet)Not disclosedNot disclosed (2-yr: ~5.7 2026, ~4.7 2027)0 (no equity program; $5B buyback)Base EPS ≥20% CAGR 2026-20291Transcript Q3'26 2; 10-K 3; Inv. pres. 4
NiSourceNIMISO3.9 signed customer requirement (~5 GenCo capacity)Up to 9 total by 2035 (~4 beyond signed)29.0 (2027-2031)~2.5 ($0.4-0.6B/yr maintenance)10-11% (2027-2033)2Transcript 5; Inv. pres. 6,7 ; PR 8; Equity 9
Vistra Corp.VSTERCOT / PJM1.2 (AWS Comanche Peak PPA)Not disclosedNot disclosed (~2.75 2026E, ~3.15 2027E)0 (VA SD OCF less dividends exceeds capex)Not given310-Q 10; Transcript 11; Inv. pres. 12; VA SD 13
NRG EnergyNRGERCOT / PJM0.445 signed1.2 (expandable to 2.4; not definitive)Not disclosed (VA SD ~4.7 FY26-30)0>14% adj. EPS CAGR through 20304Transcript 14; Inv. pres. 15; 10-Q 16; VA SD 17
FirstEnergyFEPJM7.919 (~4 in active negotiation)36.0 (2026-2030)~2.0 (up to $2B through 2030)Near top of 6-8% through 20305Transcript 18; Inv. pres. 19,20 ; PR 21; Equity 22
Alliant EnergyLNTMISO3.42-4 (advanced prospects)15.9 (2027-2030, 4-yr)~2.7 (2027-2030)8%+ 2027-203068-K 23; Transcript 24; Inv. pres. 25; 10-Q 26
PPLPPLPJM11.4 PA ESAs (+2.3 KY reimbursement agreements, not ESAs)35.3 PA advanced-stage; 12.8 KY25.0 (2027-2030, 4-yr)~4.5 ($3B remaining + ~$1.5B forwards)7-9% through 20307Transcript 27,28 ; Inv. pres./PR 29; Equity 30
WEC Energy GroupWECMISO4.5 tariff-backed (Microsoft 3.2 + Oracle 1.3)No queue figure47.5 (2027-2031)~5.5 (2026-2030 plan, pre-$47.5B)7-9%+ 2027-20318Transcript 31,32 ; Inv. pres. 33; 8-K 34
EvergyEVRGSPP2.6 (5 ESAs)Tier 1 2.0-2.5; Tier 2 1.0-2.0; Tier 3 >1021.6 (2026-2030; raised to 22.5)3.36-8%+ through 20309Transcript 35; Inv. pres. 36,37 ; 10-K 38
AmerenAEEMISO2.85.6 (1.2 construction agreements + 4.4 completed studies)38.1 (2027-2031)~5.08-10% 2027-20311010-Q 39; Inv. pres. 40; Earnings materials 41,42
Dominion EnergyDPJM (DOM)12.2 (ESA tier)42.6 pre-ESA (ELOA 32.8 + CLOA 9.8)~65 (2026-2030)~6.5 (~10% common equity in mix)5-7% through 203011Inv. pres. 43,44,45 ; Transcript 46; VA SD 47
Duke EnergyDUKNon-RTO (Carolinas); Ohio PJM, Indiana MISO7.815.4 late-stage103 (2026-2030)10.0 (2027-2030)5-7% through 203012Transcript 48,49 ; 10-Q/IR 50; Equity 51
American Electric PowerAEPPJM / ERCOT / SPP84 through 2031 (ERCOT 46, PJM 31, SPP 7)~195 interconnection queue100.4 (2027-2031)~9.7 (ATM 4 + forwards 4.8 + DRIP 0.9)9-11% 2027-203113Transcript 1,52 ; Inv. pres. 53; 10-Q 54; Financing 55
ExelonEXCPJM~2 (TSA-backed; Q2 call cited 4)~9 high-probability remaining + ~23 futureNot disclosed 5-yr (41.7, 2026-2029)3.4 (through 2029)Near top of 5-7% 2025-202914Inv. pres. 56,57 ; Transcript 58; 8-K 59
Southern CompanySONon-RTO (Southeast)21 (33 projects)>75 prospective + 5 late-stage + 2 finalizing81 base (deck shows 85)~12.0 (>$10B secured)7-8% from 2028 base15Transcript 60; Inv. pres. 61,62,63
DTE EnergyDTEMISO2.4 (Oracle 1.4 + Google 1.0)5-6~44 (2027-2031)~5.0 (~$1B/yr)>9% through 203116Inv. pres. 64,65 ; Earnings materials 66
Xcel EnergyXELMISO / SPP; PSCo non-RTO~2>20 high-probability70 (2027-2031)8.0 (~40% contracted)9%+ avg through 203117Transcript 67; Inv. pres. 68; 10-Q 69,70
SempraSREERCOT (Oncor)Not disclosed45 Batch Zero conditional64.9 (2026-2030)07-9%18Inv. pres. 71,72 ; Transcript 73; PR 74
PG&E CorporationPCGCAISONot disclosed (1.8 planning assumption)12.87Not disclosed (5-yr plan withdrawn)0 (through 2030)Withdrawn beyond 2027 (2027 +9%)19Transcript 75; Inv. pres. 76; 10-Q 77
Edison InternationalEIXCAISONot disclosedNot disclosed38-41 (2026-2030)05-7% through 203020Transcript 78; Inv. pres. 79; 10-Q 80
Public Service Enterprise GroupPEGPJM (PS)Not disclosedNot disclosed24-28 (2026-2030)06-8% through 203021Transcript 81; Inv. pres. 82,83
Eversource EnergyESISO-NENot disclosedNot disclosed26.5 (2026-2030)~0.95 ($0.8-1.1B)5-7% through 203022Conference transcript 84; Inv. pres./10-K 85; 8-K 86
Pinnacle WestPNWNon-RTO (Arizona)Not disclosed as DC-only (5.8 total committed incl. TSMC 1.88)~20 uncommitted queueNot disclosed (3-yr APS 10.35)~1.1 (2026-2028)5-7% (2024-2028)23Transcript 87,88 ; Inv. pres. 89,90
Consolidated EdisonEDNYISONot disclosedNot disclosed38 (2026-2030)~2.0 (ATM proxy; 5-yr need not disclosed)6-7% (2026-2030)24Transcript 91; 8-K 92
CMS EnergyCMSMISONot disclosed as GW (one large DC agreement signed)~924 (2026-2030)~3.756-8% (toward high end)25Transcript 93,94 ; Inv. pres. 95,96
CenterPoint EnergyCNPERCOTNot disclosed~14 Batch Zero eligible38.7 (2026-2030)~4.0 (through 2030)7-9% (mid-to-high end)26Transcript 97,98 ; 10-Q 99; Inv. pres. 100
EntergyETRMISONot disclosed as GW7-12 hyperscale (probability-weighted) + 3-5 industrial55 (2027-2030, 4-yr)5.2 (2027-2030)>8% through 203027Transcript 101,102 ; Inv. pres. 103; 10-Q 104
NextEra EnergyNEENon-RTO (FRCC)Not disclosed as DC load>20 large-load opportunity set72 (2026-2030)~7.2 (company: $12-17B through 2032)8%+ through 203228Transcript 105,106 ; Inv. pres. 107,108 ; VA SD 17
The AES CorporationAESPJM (Ohio) / MISO (Indiana)Not disclosed (Monrovia ESA, no MW)Not disclosedNot disclosed (growth capex 20-22, 2025-2027)Not disclosed (take-private pending)7-9% through 20272910-Q 109; Inv. pres. 110; PR 111; Proxy 112
American Water WorksAWKN/A (water)Not disclosed (no DC exposure)Not disclosed20+ (2026-2030)2.57-9%30Transcript/PR 113,114 ; Financing plan 115
Atmos EnergyATON/A (gas-only)Not disclosed (no DC exposure)Not disclosed~29 (FY2027-2031)~8.7 (formula)6-8% through FY203131Transcript 116; Inv. pres. 117,118 ; VA SD 119

Plan windows differ: LNT, ETR, PPL (2027-2030) and EXC (2026-2029) are four-year plans; AEE, AEP, DTE, NI, WEC, XEL and ATO run 2027-2031; others with plans run 2026-2030. CEG, NRG, PCG, PNW, AES and VST have no five-year capex plan. Visible Alpha SD data was not usable for DTE or PPL, so their equity figures rest on company plans.

Analysis of Load vs. Grid Forecasts

Share of regional forecast (D1)

TickerRegion / benchmarkContracted GWShare of benchmarkMateriality
AEPPJM + ERCOT + SPP8451.8% (PJM 47%; ERCOT 69%; SPP 23%)Dominant
DUKCarolinas Resource Plan 2026-2030 increment (3.6 GW)7.8218% (capped at 100%)Exceeds benchmark
SOGeorgia Power 2025 IRP (+9.7 GW by 2034)21216% (capped at 100%)Exceeds benchmark
WECMISO 2030 DC peak (20.5 GW)4.522.0%Meaningful
NIMISO3.919.0%Meaningful
DPJM 10-yr increase (65.7 GW)12.218.6%Meaningful
PPLPJM11.417.4%Meaningful
LNTMISO3.416.6%Meaningful
AEEMISO2.813.7%Meaningful
CEGPJM8.512.9%Meaningful (mostly existing fleet)
FEPJM7.912.0%Meaningful
DTEMISO2.411.7%Meaningful
XELMISO29.8%Moderate
EVRGSPP large-load requests (~30 GW)2.68.7%Moderate
EXCPJM~23.0%Immaterial
VSTERCOT Batch Zero base load (66.4 GW)1.21.8%Immaterial
NRGERCOT0.4450.7%Immaterial

CMS, CNP, ED, EIX, ETR, ES, NEE, PCG, PEG, SRE, AES and PNW do not disclose contracted DC GW, so their share is not disclosed; AWK and ATO have no DC exposure.

Where contracted load deviates from operator forecasts

  • PJM: contracts exceed the forecast. PJM projects a 65.7 GW 10-year summer peak increase to 222,106 MW in 2036 (PJM Load Forecast) 120, with data centers 61.1 GW of it 121. Constituents' disclosed PJM contracts (AEP 31, D 12.2, PPL 11.4, CEG 8.5, FE 7.9, EXC 2) equal 111% of that increment, though CEG's PPAs largely re-contract existing generation and may overlap utility load.
  • PJM is haircutting utility submissions. PJM cut the ATSI (FirstEnergy) 2031 summer peak by 8.1% (1.2 GW) and raised ComEd by 16.5% (3.7 GW) (PJM Load Forecast) 122. It reduced PPL's requested adjustment using a slower ramp and by discounting projects with signed letters of authorization 123, and imposed a minimum 36-month load-growth ramp 124. PL has PJM's highest 10-year zonal growth rate at 6.4% 125; DOM is at 5.4% 126.
  • MISO: six names nearly fill the 2030 DC forecast. MISO's 2026 LTLF sees DC peak rising from 1.2 GW in 2026 to 20.5 GW by 2030 (MISO LTLF) 127. LNT, AEE, DTE, NI, WEC and XEL contracted or tariff-backed load totals 93% of it, before CMS and Entergy's unquantified ESAs. MISO does not publish zone-level DC forecasts 128.
  • Non-RTO: planning lags contracting. Duke and Southern contracted load is more than double their latest official planning increments; Georgia's OpenAI contract pushes the state about 1 GW beyond recently approved capacity (Transcript) 129. FRCC projects 9,457 MW of firm peak growth over 2026-2035 130; FPL's 8 GW large-load expectation would be most of it but is not contracted.
  • ERCOT: queue far exceeds contracts. ERCOT tracks ~474 GW of large loads, ~90% data centers (ERCOT Senate briefing) 131; Batch Zero conditionally qualified 66.4 GW as base load and 127.9 GW as studied load 132. ERCOT says its 2032 LTLF "is not a prediction of what will be built" 133. Vistra sizes the real opportunity at 12-15 GW by 2030 (Transcript) 134.
  • CAISO, NYISO, ISO-NE: small forecasts, no contracts. The CEC forecasts 1.8 GW of DC growth by 2030 (CEC) 135, equal to PG&E's uncontracted planning assumption. NYISO's large-load queue is 13 GW by 2030 136; ISO-NE expects large loads to add only 110-130 MW of peak 137.

Regulatory and interconnection bottlenecks

  • ERCOT Batch Zero pause: ERCOT paused the study in August 2026 for a governor-requested data-center audit; the verification report is expected December 10, 2026 (10-Q) 138. It also paused energization of new large-load data centers pending verification 132. Exposed: AEP's 46 GW ERCOT book, SRE's 45 GW, CNP's ~14 GW.
  • PJM cost allocation: FERC's rejection of PJM's IRAS filing shifted large-load cost issues to the states (PEG Transcript) 139.
  • Transmission approvals: WEC's Vantage site depends on an ATC line pending at the PSCW (Transcript) 140; Evergy's 140 MW expansion awaits an SPP transmission study expected in Q4 2026 141. FirstEnergy assumes ~$250M of network upgrades per GW contracted 142.
  • State commission approvals: NiSource's Microsoft special contract is still pending approval (10-Q) 143; the IURC reopened the AES Monrovia record, with a hearing set for March 11, 2027 144. CMS's large DC agreement still faces siting and zoning conditions precedent 145.
  • Policy preconditions: PG&E ties DC growth to wildfire-liability reform (Transcript) 146.
  • Large-load tariffs work as contract filters: Dominion onboards customers under a Virginia SCC-approved large-load framework (Transcript) 147; Exelon's FERC-approved TSAs (~$1.2B collateral) screen out speculative requests (Inv. pres.) 56.

Equity Financing and Capex Deep Dive

Dilution risk: equity need as a share of plan capex

TickerEquity ($B)Capex ($B)Equity / capexSubsector
ATO8.72930%Gas LDC (formula-based)
PPL4.52518%Electric and gas
LNT2.715.917%Electric and gas
WEC5.533.3 (prior plan ex-ATC)16.5%Electric and gas
CMS3.752415.6%Electric and gas
EVRG3.321.615.3%Electric
SO128114.8%Electric and gas
AEE5.038.113.1%Electric and gas
AWK2.52012.5%Water
XEL87011.4%Electric and gas
DTE54411.4%Electric and gas
CNP438.710.3%Electric and gas
D6.56510%Electric and gas
NEE7.27210%Electric
DUK101039.7%Electric and gas
AEP9.7100.49.7%Electric
ETR5.2559.5%Electric
NI2.5298.6%Electric and gas
EXC3.441.78.2%Electric T&D
FE2365.6%Electric T&D
ED2385.3%Electric and gas (ATM proxy)
ES0.9526.53.6%Electric, gas and water
EIX, PEG, SRE, PCG0See master table0%No new common equity
  • Highest absolute dilution: SO (~$12B), DUK ($10B), AEP (~$9.7B), ATO (~$8.7B) and XEL ($8B). Southern has secured more than $10B of its plan (Inv. pres.) 63; AEP's plan equity is $5.2B above its prior plan (Financing) 55; Xcel has contracted about 40% of its $8B through ATM forward collars (10-Q) 70.
  • Highest relative dilution: ATO, PPL, LNT, WEC and CMS fund 15 to 30% of capex with equity. LNT ranks 6th despite this because its contracted share is high and its absolute need is small.
  • WEC understated: its $5.3-5.7B equity plan (8-K) 34 predates the $47.5B 2027-2031 capex plan, so dilution risk is likely higher than shown.
  • Zero-equity claims to monitor: Sempra says it is "eliminating the need for common equity" (Inv. pres.) 71; PG&E plans no equity through 2030 but withdrew its five-year capex plan (10-Q) 77.

Capex intensity: Electric vs. Multi-utility

SubsectorNamesAggregate plan capex ($B)Weighted equity / capex
ElectricAEP, EIX, ETR, EVRG, EXC, FE, NEE366.28.4%
Multi-utility (electric and gas)LNT, AEE, CMS, CNP, ED, DTE, D, DUK, ES, NI, PPL, PEG, SRE, SO, WEC, XEL736.69.8%

Subsector grouping follows each company's business mix (electric-only vs. electric and gas); PCG and PNW are excluded for lack of a five-year plan, and plan windows differ. Multi-utilities fund a slightly larger share of capex with equity, driven by SO, DUK and XEL; T&D-heavy electric names (FE, EXC) carry the lowest intensity. IPPs (CEG, VST, NRG) fund growth from free cash flow and need no equity.

Narrative vs. Contracted Reality

Utilities with narrative gaps (strongest first)

  • NextEra (NEE): No contracted DC load disclosed; 16.5 GW of federal hubs are development agreements, and management said "the targets are up, but the deals aren't signed yet" (Transcript) 106. FPL's 8 GW is a 2032 expectation 148.
  • Sempra (SRE): 45 GW Batch Zero load is "conditional" and subject to ERCOT's audit (Inv. pres.) 149; management concedes "a request is not a guarantee" 150. Hard evidence is nearly $6B of collateral (Transcript) 151.
  • CenterPoint (CNP): ~14 GW of Batch Zero submissions are not signed contracts; connected GW "may vary and be less" (10-Q) 152.
  • Entergy (ETR): References signed ESAs without quantifying GW; 7-12 GW pipeline is probability-weighted (Transcript) 102.
  • NRG: The 1.2 GW Texas project is not definitive; only 445 MW is signed (Transcript) 153.
  • PG&E (PCG): 1.8 GW is a planning assumption; only 400 MW is at Interconnection Construction Agreement stage and 174 MW online, against a 12.87 GW pipeline (Inv. pres.) 76.
  • Xcel (XEL): ~2 GW contracted against >20 GW "high-probability"; the November deck marks contracted load "Not publicly disclosed" (Inv. pres.) 68. Mitigant: base plan does not depend on new DC contracts 154.
  • Exelon (EXC): TSA figure inconsistent, 4 GW on the Q2 call vs. ~2 GW in the deck (Transcript) 58; only a "small amount" is in plan 155.
  • Dominion (D): "54 GW in various stages of contracting" vs. 12.2 GW under ESAs (~22%); ELOA/CLOA authorizations are pre-contract (Inv. pres.) 156.
  • Vistra (VST): Only the 1.2 GW AWS PPA is quantified; "a lot of megawatts still to contract on nuclear" (Transcript) 11.
  • PSEG (PEG): Cites "optionality around data centers" with no counterparty or GW (Transcript) 157.
  • Milder gaps: AEP's 84 GW book sits against a ~195 GW queue, and "not all of that will come on" (Transcript) 158; WEC's 4.5-5 GW Microsoft upside is its own estimate 159; Evergy Tier 3 >10 GW; PNW gives no DC-only figure.
  • Reverse gap: CMS's signed large DC agreement (>1 GW) is excluded from its capital plan (Transcript) 94. ED, ES, EIX, AWK and ATO make no DC growth claims.

Top performers: contracted load to pipeline

RankTickerContracted ÷ pipeline (GW)Ratio
1LNT3.4 ÷ 3 (midpoint of 2-4)1.13
2NI3.9 ÷ 40.98
3DUK7.8 ÷ 15.40.51
4AEE2.8 ÷ 5.60.50
5DTE2.4 ÷ 5.5 (midpoint of 5-6)0.44
6AEP84 ÷ 1950.43
7FE7.9 ÷ 190.42
8NRG0.445 ÷ 1.20.37
9PPL11.4 ÷ 35.30.32
10D12.2 ÷ 42.60.29
11SO21 ÷ 820.26
12EVRG2.6 ÷ 13.750.19
13XEL2 ÷ 200.10
14EXC2 ÷ 320.06

The ratio cannot be computed for CEG, VST, WEC and names without a contracted or pipeline figure. Contract discipline is explicit at the leaders: Alliant aims "not to present the biggest number" (Transcript) 160; Duke is "not counting anything until it hits the ESA level" 161; FirstEnergy puts no uncontracted demand in plan 162; Ameren includes only ESA-backed sales 163.

Appendix: Regional Grid Forecast Sources

RegionSourceProjection used
PJM2026 Long-Term Load Forecast Report+65.7 GW summer peak over 10 years to 222,106 MW (2036) 120; data centers 61.1 GW of the increase 121
MISO2026 LTLF (Apr 2026 update)DC peak 1.2 GW (2026) to 20.5 GW (2030) 127; 63% load growth 2026-2046 164
ERCOTBatch Zero classifications (Sep 3, 2026); preliminary 2026-2032 LTLF; Senate briefing66.4 GW base load, 127.9 GW studied 132; LTLF "not a prediction" 133; ~474 GW queue 131
SPP2026 ITP scoping; HILL requests~30 GW large-load requests 165; 27 HILL requests, 11,163 MW 166
CAISO / CECCEC data-center forecast (Jan 2026); 2025-2026 Transmission Plan1.8 GW DC growth by 2030 135,167
NYISO2026 Gold Book; Power Trends; large-load queue13 GW large-load queue by 2030 136,168,169
ISO-NE2026 CELTLarge loads add 110-130 MW of peak 137,170
Carolinas (non-RTO)Duke 2025 Carolinas Resource PlanDEC 18,633 to 20,998 MW; DEP 12,739 to 13,950 MW (2026-2030) 171
Southeast (non-RTO)Georgia Power 2025 IRP; NERC 2025 LTRA+9.7 GW summer peak by 2034 172; SERC-Southeast +4,468 MW 2027-2029 173
Florida (FRCC)FRCC 2026 Ten-Year Site Plan overview+9,457 MW firm peak 2026-2035 130
Arizona (APS)APS load forecast presentationPeak 8.7 GW (2025) to 12.0 GW (2035); 19 GW uncommitted queue 174
NationalNERC aggregate224 GW 10-year summer growth 175; an EIA regional load forecast is not available here