PJM peak load vs forecast, 2017-2026 — does it corroborate the ERCOT-only anchor?
The question
"Does PJM's peak-load-vs-forecast data (2017-2026, via Wayback Machine) corroborate the ERCOT-only evidence anchoring the power-cycle thesis?"
Context: power-cycle-v1.1 has $20,000 of paper capital deployed. Its phase-marker #1 is stated as a PJM + ERCOT test, but the anchor dataset backing it contains ERCOT rows only — [[2026-05-17-power-anchors-backfill-notes]] flagged the PJM gap and never closed it.
Headline: PJM does NOT corroborate the ERCOT reading. The two ISOs point in opposite directions on realized peak load, and both directions are mostly weather artifacts. On the thesis's own marker-1 threshold, PJM's verdict flips depending on whether you use raw or weather-normalized peaks — meaning marker-1 as currently written is not decidable, not merely unconfirmed.
What we already know (from the vault)
- Marker 1, in the thesis's own words ([[2026-05-17-power-cycle-v1]], "Phase markers" table): "PJM + ERCOT annual peak-load growth vs forecast. Bullish state: Growth at or above 3.3% PJM / 9.6% ERCOT 2026 baseline. Bearish flip: Materially below baseline for 2 consecutive annual cycles — demand softening." In v1 a marker-1 bearish flip sustained 2 quarters = close the entire power bucket; [[2026-05-18-power-cycle-v1.1]] softened this to a founder-review gate.
- The ERCOT evidence is a miss, not a beat. The anchor CSV (
anchors/power-layer/iso-demand-growth-2017-2026.csv) records ERCOT peaks of 85.46 GW (2023) → 85.20 GW (2024) → 83.68 GW (2025): -0.31% then -1.78% YoY, against a 9.6% bullish baseline. 2025 actuals came in 1.19% below ERCOT's own ERCOT-Adjusted forecast (83,679 vs 84,686 MW) and 9.6% below the TSP-Provided forecast (92,582 MW). - The backfill notes already flagged both readings. Bull: 2025 was a mild Texas summer, weather-normalize it. Bear: "consistent with the bear-case in the thesis that the data-center demand surge is being over-counted in forecasts." The notes concluded "2025 alone doesn't trip it."
- The PJM gap was known and scoped. Same notes: "PJM data is the biggest gap... PJM's main domain (www.pjm.com) does NOT resolve from the scraper's network (DNS NXDOMAIN)", with a "~30 min" Wayback backfill queued as a next step.
- The thesis is timing-sensitive. [[2026-05-18-power-cycle-v1.1]] sizes the basket on being "in phase 2 (capacity announcement velocity accelerating) heading into phase 3 (capacity online)", citing +60-120pp historical returns for that transition "in 12-18 months."
What the web says
Retrieval note: Wayback was not needed. PJM keeps every Load Forecast Report vintage live at a stable URL pattern, so all ten vintages (2017-2026) were fetched directly from pjm.com in this run and parsed locally. Every number below comes from a PDF or XLSX I downloaded and read on 2026-07-28.
- PJM's realized summer peaks went UP hard in 2024-2025, the opposite of ERCOT. Non-coincident unrestricted RTO summer peak: 146,799 MW (2023) → 152,314 MW (2024, +3.76%) → 160,709 MW (2025, +5.51%). 2025 is PJM's highest summer peak since 2011 and its fourth-highest on record, behind only 2006 (166,929 MW), 2011 (165,524 MW) and 2007 (162,035 MW). Source: Table F1-Summer, 2026 Load Report Tables (XLSX).
- But PJM's own weather-normalized peaks did not. Same table, WN column: 149,830 (2023) → 151,025 (2024, +0.80%) → 153,503 (2025, +1.64%). The entire +5.51% raw 2025 print is the 6/23-6/24/2025 heat event — PJM's 5CP sheet shows 160,649 MW coincident on 6/23/2025 18:00 against a weather-normalized coincident RTO peak of 153,870 MW (Summer 2025 Peaks & 5CPs PDF, revised 11/14/2025).
- PJM's near-term forecast error flipped sign in 2024 — on raw actuals only. Vintage-N forecasting year N over-shot realized peak in 7 of 7 years, 2017-2023 (mean -2.0%), then under-shot in 2024 (+0.71%) and 2025 (+4.26%). Measured against PJM's own weather-normalized actuals, however, there is no regime change: the 2024 and 2025 errors are -0.15% and -0.42% (still slight over-forecasts), and error magnitude has shrunk, not grown. See Table 1.
- The real PJM signal is in the forward curve, not realized load. PJM's forecast for target year 2030 went 153,775 MW (Jan-2022 vintage) → 157,899 (Jan-2023) → 167,873 (Jan-2024) → 183,883 (Jan-2025) → 183,008 (Jan-2026): a +19.0% revision in four vintages. PJM's stated 10-yr summer CAGR moved from 3.1%/yr (Jan-2025 report, p.6) to 3.6%/yr (Jan-2026 report, p.6).
- The Jan-2026 vintage pushed demand OUT in time. Against the Jan-2025 vintage on identical target years: 2026 -1.61%, 2027 -2.27%, 2028 -2.60%, 2029 -2.59%, 2030 -0.48%, then higher from 2033 (+0.92%), 2034 (+2.23%), 2035 (+3.31%). PJM did not cut the total; it delayed roughly 1-2 years of it. Both rows computed from the Jan-2025 and Jan-2026 Table B1 sheets.
- PJM attributes the growth to data centers by name. The Jan-2026 report's zonal load-adjustment table cites "Growth in data center load" explicitly for DOM (Dominion/Northern Virginia) and PS (PSE&G), among others (2026 Load Report, p.5).
- Independent confirmation of the 2025 print. PJM's Independent Market Monitor records the 2025 summer peak-load hour at 156,256 MWh on Monday 6/23/2025 HE18 vs 148,890 MWh on 7/16/2024 — 2025 State of the Market, Section 3, Table 3-6. Different metric (MMU accounting load excluding losses, not PJM unrestricted non-coincident peak), so the levels differ from PJM's; the direction and the 6/23/2025 date agree.
Table 1 — PJM RTO summer peak: vintage-N forecast for year N vs realized
Error = actual − forecast. Positive = PJM under-forecast. All forecast figures are the PJM RTO row of Table B-1 (summer peak, MW, non-coincident unrestricted) in the January report of that vintage. All actuals are Table F1-Summer of the Jan-2026 tables workbook (PJM restates prior-year WN values each vintage; using one vintage throughout avoids mixing restatements).
| Vintage (Jan) | Target yr | Forecast MW | Actual MW | Err MW | Err % | WN actual MW | Err vs WN % |
|---|---|---|---|---|---|---|---|
| 2017 | 2017 | 152,999 | 145,434 | −7,565 | −4.94% | 149,573 | −2.24% |
| 2018 | 2018 | 152,108 | 150,573 | −1,535 | −1.01% | 149,707 | −1.58% |
| 2019 | 2019 | 151,358 | 151,302 | −56 | −0.04% | 149,088 | −1.50% |
| 2020 | 2020 | 148,092 | 144,320 | −3,772 | −2.55% | 147,175 | −0.62% |
| 2021 | 2021 | 149,224 | 148,433 | −791 | −0.53% | 148,912 | −0.21% |
| 2022 | 2022 | 148,938 | 147,361 | −1,577 | −1.06% | 149,000 | +0.04% |
| 2023 | 2023 | 149,059 | 146,799 | −2,260 | −1.52% | 149,830 | +0.52% |
| 2024 | 2024 | 151,247 | 152,314 | +1,067 | +0.71% | 151,025 | −0.15% |
| 2025 | 2025 | 154,144 | 160,709 | +6,565 | +4.26% | 153,503 | −0.42% |
| 2026 | 2026 | 156,373 | not retrieved | — | — | not retrieved | — |
2026 actual is not retrieved: the 2026 summer season is still in progress as of 2026-07-28, and PJM has not posted a summer-2026-peaks-and-5cps file (URL probed, returns /not-found). Do not evaluate the 2026 row until PJM's Jan-2027 report or the summer-2026 5CP sheet publishes. This is the same partial-year trap the backfill notes flagged on ERCOT's 2026 row.
All 9 non-2026 rows are backed by a document retrieved and parsed in this run. No figure is interpolated, estimated, or recalled.
Table 2 — Forecast for fixed target year 2030, by vintage
| Vintage | 2030 forecast MW | Δ vs prior vintage |
|---|---|---|
| Jan-2020 | 157,132 | — |
| Jan-2021 | 153,536 | −3,596 |
| Jan-2022 | 153,775 | +239 |
| Jan-2023 | 157,899 | +4,124 |
| Jan-2024 | 167,873 | +9,974 |
| Jan-2025 | 183,883 | +16,010 |
| Jan-2026 | 183,008 | −875 |
Jan-2017 through Jan-2019 vintages do not reach 2030 (11-year horizon) → not applicable, not missing.
Table 3 — ERCOT side, for comparison (vault-sourced, NOT re-verified in this run)
| Year | ERCOT actual peak MW | ERCOT-Adj forecast MW | YoY | vs forecast |
|---|---|---|---|---|
| 2023 | 85,464 | not in dataset | +6.78% | — |
| 2024 | 85,199 | not in dataset | −0.31% | — |
| 2025 | 83,679 | 84,686 | −1.78% | −1.19% |
| 2026 | 75,572 (partial-yr) | 93,504 | — | do not use |
Source: iso-demand-growth-2017-2026.csv, scraped 2026-05-17 by power-anchors-scraper.py. I did not re-fetch ERCOT primary data in this run — these rows are carried from the vault as-is. ERCOT forecast columns are blank pre-2025 because the Peak Demand Scenarios workbook only publishes forward years.
Convergences and contradictions
- Direct contradiction on realized peak direction. ERCOT's realized peak fell two years running (−0.31%, −1.78%); PJM's rose two years running (+3.76%, +5.51%). PJM does not corroborate ERCOT — it points the other way. Anyone who read the ERCOT rows as "demand surge is over-counted" would read the PJM rows as "demand surge is real and under-counted," and both would be over-reading weather.
- Convergence once you weather-normalize — and it is unflattering to the thesis. PJM's weather-normalized growth is +0.80% (2024) and +1.64% (2025), against a 3.3% bullish baseline. ERCOT's raw −1.78% is against a 9.6% baseline. Neither ISO's realized peak load is growing at the rate marker-1 requires. The 2025 divergence is a mild Texas summer against a June heat wave in the Mid-Atlantic.
- Convergence on the forecast side, which is the weakest evidence. Both ISOs show large upward forecast revisions driven by large-load/data-center interconnection requests (PJM 2030: +19.0% over four vintages; ERCOT's TSP-Provided 2027 forecast of 137,034 MW vs an 83.7 GW realized 2025). But forecast revisions are downstream of the same speculative-load submissions the thesis's own bear case says are double-counted across utility territories. Corroboration between two ISOs sharing an input bias is not independent corroboration.
- The thesis's 3.3% PJM figure does not match either retrieved vintage. PJM's Jan-2025 report states 3.1%/yr; Jan-2026 states 3.6%/yr. The 3.3% in [[2026-05-17-power-cycle-v1]] traces to footnote [2] (an Inside Lines year-in-review post), not to a Load Forecast Report I retrieved. Flagged as unverified.
Synthesis for RDCO
The 30-minute follow-up was worth more than 30 minutes. The PJM backfill does not turn marker-1 from "unconfirmed" into "confirmed" or "broken" — it shows marker-1 is specified badly enough that it cannot be evaluated as written, and that the ERCOT-only version was giving a falsely clean signal in the bearish direction just as a PJM-only version would have given a falsely clean signal in the bullish direction. Marker-1 says "annual peak-load growth vs forecast" without naming a basis. On raw non-coincident unrestricted peaks, PJM clears the 3.3% bullish threshold in both 2024 (+3.76%) and 2025 (+5.51%), and ERCOT trips the bearish flip on a literal read (materially below 9.6% for two consecutive annual cycles). On weather-normalized peaks — the basis PJM itself publishes precisely so that year-over-year comparisons mean something — PJM fails the bullish threshold in both years (+0.80%, +1.64%), and the whole apparent regime change evaporates: PJM's forecast error against WN actuals has been within ±2.3% for a decade with no directional break. Two metrics, two opposite verdicts, on a marker whose v1 consequence was closing a $20,000 bucket. That is the finding.
The second finding is more actionable and cuts against the position's timing, not its direction. PJM's Jan-2026 vintage trimmed target years 2026 through 2032 versus the Jan-2025 vintage (−1.61% for 2026, −2.27% for 2027, −2.60% for 2028, −2.59% for 2029) and raised 2033-2035 (up to +3.31% for 2035). PJM did not lose faith in the data-center load; it moved roughly one to two years of it to the right. power-cycle-v1.1 sizes the basket on being in phase 2 heading into phase 3, and claims that transition historically pays +60-120pp "in 12-18 months." A one-to-two-year rightward shift in the largest data-center ISO's own load curve is a direct hit on that 12-18-month clock — not on whether phase 3 arrives, but on whether it arrives inside the window the sizing assumes. The v1.1 anchor list (five founder-review triggers) has no trigger that would catch this, because none of them watch forecast-vintage deltas; they watch queue levels, PPA cadence, GEV orders, 13F deltas, and power prices. A "delay, not cancel" revision is invisible to all five.
The third finding is about evidence quality. The strongest-looking PJM datapoint — a +19.0% upward revision to the 2030 forecast in four vintages — is the weakest evidence in this brief, because PJM's forward forecast is built from large-load interconnection requests submitted by developers, which is the same input stream the thesis's own bear case identifies as speculative and double-counted. When PJM's forecast rises and ERCOT's forecast rises, that is one signal observed twice, not two signals. The genuinely independent PJM contribution is the realized weather-normalized series, and it says load growth accelerated from +0.06% (2022) to +0.56% (2023) to +0.80% (2024) to +1.64% (2025) — a real, monotonic, four-year acceleration, but running at roughly half the rate marker-1 calls bullish. The honest position: the demand story is real and accelerating, and it is arriving slower than the thesis priced.
None of this is an exit signal. The basket's return case in [[2026-05-18-power-cycle-v1-v2-honest-rerun]] rests on structural supply scarcity (queue depth ~13 years at ERCOT's clearing rate) far more than on realized peak-load prints, and nothing here touches the queue. But the PJM backfill removes the option of citing marker-1 as supporting evidence in either direction until the basis is fixed, and it puts a dated, PJM-sourced number on a timing risk that the current anchor set cannot see.
Why this is in the vault
This closes the PJM gap explicitly named in [[2026-05-17-power-anchors-backfill-notes]] and directly changes how phase-marker #1 of [[2026-05-18-power-cycle-v1.1]] must be evaluated at the next quarterly review of the $20,000 power bucket — specifically, it establishes that the marker needs a weather-normalization basis before it can trigger anything, and it surfaces a PJM-sourced timing risk (demand pushed 1-2 years right) that none of v1.1's five founder-review triggers currently monitor.
Open follow-ups
- PJM's Jan-2026 near-term trim: real de-commitment or methodology? Did the −2.6% cut to 2028-2029 come from data-center projects actually withdrawing / slipping their in-service dates, or from a change in how PJM haircuts large-load requests (a Manual 19 revision or a new ramp-rate assumption)? Answerable from PJM Planning Committee materials, the Manual 19 revision history, and the Jan-2026 report's load-adjustment tables. This determines whether the delay is demand-side or accounting.
- Did IPP equities historically re-rate on realized load or on forecast revisions? In the 2000-2007 PJM/ERCOT build-out, what weather-normalized peak growth did each ISO actually realize, and did merchant-generator multiples track realized load or track forecast-vintage revisions? This is the test of whether marker-1 should key on realized load at all — if the historical re-rate tracked forecasts, marker-1's realized-load framing is measuring the wrong variable.
- ERCOT weather-normalized peaks were never pulled. The vault's ERCOT rows are raw peaks only. Without an ERCOT WN series the ERCOT-vs-PJM comparison in Table 3 is apples-to-oranges against Table 1. ERCOT publishes weather-normalized peak demand in its Capacity, Demand and Reserves (CDR) and Hourly Load Data products.
- PJM 2026 realized summer peak — retrievable after the season closes; the Jan-2027 Load Forecast Report tables workbook will carry it in Table F1-Summer.
- Not examined here: PJM capacity-auction clearing prices and PJM interconnection-queue depth. Both are frequently cited as "PJM evidence" for this thesis and are different evidence from peak load. Nothing in this brief speaks to them, and no substitution was made.
Related
- [[2026-05-17-power-anchors-backfill-notes]] — the anchor backfill that flagged the PJM gap this brief closes
- [[2026-05-18-power-cycle-v1.1]] — the live thesis with $20k paper capital; marker-1 and the five founder-review triggers
- [[2026-05-17-power-cycle-v1]] — superseded v1, but the source of marker-1's verbatim wording and the 3.3%/9.6% baselines
- [[2026-05-18-power-cycle-v1-v2-honest-rerun]] — backtest the position sizing rests on
- [[2026-05-18-pressure-test-and-winners-survey]] — phase-2-into-phase-3 call that the timing finding bears on
- [[2026-05-18-stratechery-data-center-discontent]] — demand-side NIMBY / speculative-load skepticism
Sources
Retrieved and parsed in this run (2026-07-28):
- PJM Load Forecast Report, Jan-2017: https://www.pjm.com/-/media/DotCom/library/reports-notices/load-forecast/2017-load-forecast-report.ashx
- PJM Load Forecast Report, Jan-2018: https://www.pjm.com/-/media/DotCom/library/reports-notices/load-forecast/2018-load-forecast-report.ashx
- PJM Load Forecast Reports, Jan-2019 through Jan-2026: https://www.pjm.com/-/media/DotCom/library/reports-notices/load-forecast/YYYY-load-report.ashx (substitute 2019…2026 for YYYY; all eight fetched)
- PJM 2025 Load Report Tables (XLSX): https://www.pjm.com/-/media/DotCom/planning/res-adeq/load-forecast/2025-load-report-tables.xlsx
- PJM 2026 Load Report Tables (XLSX): https://www.pjm.com/-/media/DotCom/planning/res-adeq/load-forecast/2026-load-report-tables.xlsx — Table B1 (forecast) and Table F1-Summer (historical actual + weather-normalized peaks)
- PJM Summer 2025 Weather Normalized RTO Coincident Peaks & 5CPs: https://www.pjm.com/-/media/DotCom/planning/res-adeq/load-forecast/summer-2025-peaks-and-5cps.pdf
- Monitoring Analytics, 2025 State of the Market for PJM, Section 3 (Energy Market), Table 3-6: https://www.monitoringanalytics.com/reports/PJM_State_of_the_Market/2025/2025-som-pjm-sec3.pdf
Probed and confirmed unavailable:
https://www.pjm.com/-/media/DotCom/planning/res-adeq/load-forecast/summer-2026-peaks-and-5cps.pdf→ redirects to/not-found(season in progress)https://www.pjm.com/-/media/DotCom/planning/res-adeq/load-forecast/2024-load-report-tables.xlsx→/not-found(workbook companion begins with the 2025 vintage; 2017-2024 figures were parsed from the report PDFs instead)
Vault (not re-verified in this run):
~/rdco-vault/01-projects/investing/anchors/power-layer/iso-demand-growth-2017-2026.csv— ERCOT rows in Table 3~/rdco-vault/01-projects/investing/anchors/power-layer/2026-05-17-power-anchors-backfill-notes.md~/rdco-vault/01-projects/investing/theses/2026-05-18-power-cycle-v1.1.md~/rdco-vault/01-projects/investing/theses/2026-05-17-power-cycle-v1.md
Paywalled/skipped: none encountered.
Wayback Machine: not used. All ten forecast vintages were available live on pjm.com, which is a stronger provenance than an archive snapshot. Noted because the question presumed Wayback would be required.