Version history & corrections

We version the method and date every change here. When we find an error, we correct it in the open, date the correction, and, because the ledger is append-only and externally anchored, keep the original on the record rather than rewriting history. The method changes when the evidence changes; this is the log.

Oct 27, 2026Page text corrected after a review, no change to the number. On /queues, the summary and the NYISO entry still said each of the 74 load requests carries an end-use class; 51 do, as the Oct 3 entry below says, and both now give that count. The page called the show-cause orders “proposals, not mandates” and said nothing is required yet; the orders make preliminary findings and call for a response from each operator, and no posting obligation is in force yet. One sentence named only PJM as still open; SPP is open too. The PJM entry now reports the Commission’s order of Oct 9, 2026, on the Large Load Registry. On /dockets, six dockets carried a title taken from one filing made in several dockets at once; a docket now takes its title only from a notice filed in that docket alone. One filer had also been grouped with AI companies and data centers by a name rule; it no longer is. On /verify, the derivation said its figures were pulled live; the queue files are read once a day. On loadstar.energy/about, the account of past corrections said the number went down on Sep 29, 2026; four corrections in late September changed it, three raising it and one lowering it. The same page said every large-load proceeding starts from the record this index reads; the generation queues are one record that bears on them. And loadstar.energy/experts said the source archive comes with a data license without the condition that loadstar.energy/license states: only where the operator’s terms permit redistribution.
Oct 22, 2026Two notes on FERC proceedings brought up to date, no change to the number. On /queues, the California entry said CAISO had asked the Commission to hold the show-cause proceeding in Docket EL26-71 in abeyance. The Commission granted that request on Aug 14, 2026. Its order of Sep 21, 2026, in Docket EL26-71-001 recites that responses to the show-cause order are due on Nov 16, 2026, and that the obligation to respond is suspended for any respondent that makes a section 205 filing addressing the Commission’s preliminary findings on or before that date. The entry now gives the Nov 16 date and the section 205 condition. On /dockets, the note on Docket ER26-3515 described only PJM’s filing. It now also reports the Commission’s order of Oct 9, 2026, which accepted the Large Load definition and the Large Load Registry and rejected the remaining proposed tariff revisions.
Oct 12, 2026Four days without a reading, and the older pages retired; no change to the number. Our pages say the number is read once a day. Between Jul 16 and Oct 12, 2026, four days have no reading: Jul 20, Aug 6, Aug 20 and Oct 4. The daily series has no entry for those dates. The scheduler waited 24 hours after the last reading but checked only every six hours, so the reading slipped later until a calendar day passed without one. Since Oct 11 the reading runs once per calendar day, and an internal check flags a day left without one. No published figure was wrong, and none changes. We also retired the older pages of index.loadstar.energy: the regional map (/map), the per-region pages (/r/), /energy, /projects, /factory, the regional heat widget (/embed/heat), their preview images, and the JSON endpoints under /api/v1/regions that served the regional index. They were earlier experimental work, not part of the LOADSTAR Number. They still carried a tagline we no longer use and the three words we do not use for requests without a signed agreement. Each of those addresses now answers with a short notice (HTTP 410) that points to this log, and the header, the footer, /api, /llms.txt and the sitemap no longer link to them. The region_reality tool of /mcp returns a retirement notice. /journal stays, with every sealed call. Under /api/v1/regions, queue-derate, known-objections and the methodology record stay. The ledger root (/api/v1/ledger/root) carries a sha256 fingerprint of the methodology record, so that record stays as written; it still carries some of those words. The Oct 10 entry said some of those words remained on the CFI methodology page of loadstar.energy and in the PDF linked from it. That page no longer carries them: the lowest band of its scale is now labeled “LOW”, and the PDF was removed.
Oct 10, 2026Old wording removed from /the-number, /verify and /api, and new sealed calls paused in fact; no change to the number. On /the-number we removed two lists, “The consequences already landing” and “Who pays”. They were commentary, and one line spoke of dozens of requests for every confirmable project, which the number on the same page does not support. The ready-made citation there is now signed “LOADSTAR”, with no tagline. On Oct 2 we wrote that pages say signed capacity “shows a signed agreement”; the text of /the-number, /verify and /llms.txt still used the older word in several places, and now follows the Oct 2 wording. /verify also still called the share a lower bound by construction; it now says we treat it as a floor, for the reason given on Oct 8. Three words we do not use for requests without a signed agreement were still on /the-number, /verify, /api and the archived CFI methodology page, most of them in sentences that denied them. Those sentences are rewritten without the words. Some remain: two inside the caveat of the headline sentence, and others on the archived CFI methodology page and in the PDF linked from it. We also removed from /api a sentence about when a predictive track record would be validated. /the-number and /ai-load carried, in their machine-readable data, questions and answers that the pages do not show; one asked about the AI data-center boom and answered with the generation-queue figure. Those blocks are gone, and the dataset description is now a sentence from the page, without figures. The footer of index.loadstar.energy now matches the one on loadstar.energy. Sealed calls: since Oct 3 this log and our pages have said that new sealed calls are paused. The weekly job kept running, and /journal shows 18 calls sealed on Oct 9 and Oct 10. The job is off as of Oct 10. Every call already sealed, those 18 included, stays on /journal and will be resolved on its date under the rule published before. The headline sentence and its caveats still say “unverified”; a change to them goes through the review that every change to that sentence goes through, and will be logged here. The arithmetic, the rules per queue and the figures are unchanged.
Oct 8, 2026Wording of the headline sentence, no change to the number. The sentence that carries the number said “at least” before the signed gigawatts. We removed it: in NYISO we read “signed” from the project stage, so a project could count before its agreement is published, and an absolute floor claimed more than our reading supports. The sentence now says we treat the figure as a floor, and carries two caveats in line: NYISO’s status is read from the project stage, and MISO’s active list shows almost no signed capacity, with the share without MISO given next to it. We also unified MISO’s treatment (one text said we report no share for it while another page showed 0.1%), wrote “ISO/RTO” and “ISO-NE”, added the thousands separator, replaced “not a rate” with “not a forecast”, and gave the definition of “signed” the stage reading used in NYISO and PJM. The arithmetic, the rules per queue and the figures are unchanged.
Oct 7, 2026The project-stage caveat now applies to NYISO only, no change to the number. On loadstar.energy/signed-share we said that in PJM and NYISO we read “signed” from the project stage, not from an agreement field, and called that weaker evidence, because a project could count before its agreement is published. We checked PJM’s files on Oct 7: every row we count as signed there also lists a posted agreement document. The caveat now applies to NYISO, whose file carries no agreement field to check against. For PJM the page says we read the stage and that each counted row also lists a posted agreement. We checked the column in PJM’s file, not the documents themselves.
Oct 5, 2026A leftover sentence on /the-number, no change to the number. Under “Who verifies intent?”, the page still said that several queues could not be turned into a comparable share, that some publish no usable field, and it named ERCOT next to MISO as near-zero. That described the rule we replaced on Sep 29. All seven queues are read, and ERCOT shows a signed share on the same page. The sentence now names only the queue our reading places near zero, taken from the same data as the citation on that page.
Oct 3, 2026A newer NYISO file on /ai-load, no change to the number. The /ai-load page moved from the July 31, 2026 edition of NYISO’s interconnection queue file to the file read on Oct 3, 2026, whose rows carry updates through Sep 30, 2026. Load in NYISO’s “AI Datacenter” class went from 5,290 MW to 5,266 MW, across 12 requests in both files. The total NYISO classifies by end use went from 13,568.9 MW to 13,611.9 MW. The new file has no date in its name, so the page gives the day we read it and the latest update date on its rows. The NYISO entry on /queues was updated from the same file. The NYISO note on /queues also said every load row carries an end-use code; 51 of the 74 rows do, and the note now gives that count. This does not change the LOADSTAR Number, which is computed from the generation queues.
Oct 3, 2026An out-of-date page, no change to the number. The front page of index.loadstar.energy still described the number under the rule we replaced on Sep 29: it said the figure came from six of seven queues and treated ERCOT as not comparable, and it announced sealed predictions as live although they are paused. The front page now opens the number itself, the older regional map is marked as an archive, and the out-of-date note and the predictions line are gone.
Oct 2, 2026Wording and rounding, no change to the number. Each queue’s share is now computed from megawatts, not from rounded gigawatts, so ISO-NE shows 67% where it showed 66%. Where pages said signed capacity was “verified”, they now say it “shows a signed agreement” in the operator’s queue file: we read each operator’s status field, we do not inspect contracts. Pages also show the share without MISO next to the headline. On the home page, the PJM capacity-cost note was corrected to the market monitor’s own wording: the $9.3 billion is how much lower total revenue in the 2025/26 capacity auction would have been without data-center load, not a share of the price increase. This log was put in date order; entries written for the same date were merged, with both texts kept.
Sep 29, 2026Correction: the national signed share goes down. One audit found errors in how we read four queues, and fixing them lowers the national signed share. PJM (three errors): PJM publishes Cycle 1 and the two transition cycles in a separate file from its older serial queue, and we had been reading only the older file, so Cycle 1 was missing. The older file also lists long-term transmission-service requests, which we had included in the count. And we had sized requests by “Maximum Facility Output”, which PJM defines as the output of the whole facility, so a request at an existing plant counted the entire plant (a 44 MW request at an existing nuclear plant counted as 2,709.8 MW). We now read both files, count only generation requests, and size each request by the larger of its “MW Energy” and “MW Capacity”. SPP: projects marked “TERMINATED” had been counted as active. ISO-NE: a 1,200 MW transmission line had been counted as a signed generator. ERCOT: a new rule excludes requests whose project name marks a fuel change, an upgrade or an uprate. Two of the requests it matches are 50 MW or more, the size we count: a fuel change and a turbine upgrade, each listed at the size of the whole existing unit (655 MW and 1,336.5 MW). Projects named as repowers are not excluded and count at the size ERCOT lists. The number’s history marks this date as a change of method: the number on this date is not comparable with the days before it.
Sep 25, 2026Clarification: what MISO’s “Done” means. On Sep 24 we wrote that MISO does not publicly define its queue statuses and that projects reaching the agreement stage are listed as “Done”. MISO does define them, in the legend of its interactive queue: “Active” is a project still “under the study cycle”, and “Done” is one that “has completed all the studies”, after which agreement negotiation begins. “Done” therefore marks the end of study, not a signed agreement, and we do not count it. The number does not change; the explanation now uses MISO’s own words.
Sep 25, 2026Correction: ERCOT now counts. We read ERCOT’s queue through gridstatus, an open-source library that labels every ERCOT project with a date in the “IA Signed” column as “Completed”. We took that label to mean the project had left the queue, and said so on this site, attributing it to ERCOT. It was the library’s label, not ERCOT’s: ERCOT’s own queue report keeps signed projects on its list, with the signature date. As a result ERCOT counted 0 GW signed. ERCOT lists about 137 GW of signed projects of 50 MW and up. About 116 GW of that is not yet approved for synchronization, and it now counts; we treat a project as in service once ERCOT records that approval. ERCOT’s requested total rises by the same amount, because those projects were missing from it too. The national signed share rises by roughly 9 points from this correction alone, not from anything that happened in the queue, and the number’s history marks the change.
Sep 24, 2026Correction: PJM was undercounted by about 40 GW. We counted as signed only the PJM projects listed as “Under Construction”. PJM itself groups three queue statuses as projects with executed final agreements: “Engineering and Procurement”, “Suspended” and “Under Construction” (Markets and Reliability Committee, “Commercial Deployment of New Generation”, Sept. 25, 2024, slide 7). The 202 active projects in “Engineering and Procurement” — about 40 GW, all from the older serial queues, most with the agreement document posted in PJM’s own agreement column — now count. PJM files “Suspended” projects under executed agreements too, but a suspended project is not in PJM’s active queue, which is the population we measure, so it does not count here either. The national signed share rises by roughly 4 points from this correction alone, not from anything that happened in the queue, and the number’s history marks the change. PJM is now read from its named status field, so the note that part of the signed total is inferred from free text no longer applies.
Sep 24, 2026Correction: why MISO shows almost no signed capacity. Our explanation said MISO’s post-agreement status field was filled on “only a sliver” of rows, for a reason unrelated to ERCOT’s. Both parts were wrong. The field is filled on roughly four in ten active MISO requests, and nearly all of those read “Not Started” while still in study. The actual reason is the same as ERCOT’s: in MISO’s published data, projects that reach the interconnection-agreement stage are listed as “Done”, not “Active”. The number does not change — MISO still contributes almost nothing to the signed total — only the explanation does. MISO’s public queue documentation does not define these statuses; this reading comes from the data itself, and the check that watches it now tests the lifecycle, not just that the column exists. [Sep 25: two parts of this row were later corrected — ERCOT’s reason was our own misreading, and MISO does define these statuses. See the Sep 25 rows above.]
Sep 24, 2026Correction: NYISO now counts. NYISO publishes each active project’s stage in a column named “Project Status #”, with the key printed on its own sheet (11 = IA Completed, 12 = Under Construction). Our reader had lost that column, so NYISO’s queue counted 0 GW signed, and our text said we had not resolved whether NYISO publishes such a field at all. It does. NYISO now contributes roughly 7.5 of its 30 GW, and the national signed share rises by roughly 0.8 points from this correction alone — not from anything that happened in the queue. The number’s history marks the change instead of cutting the series, and its range and trend are computed within the current method only. We also found why this went unnoticed: the automated check behind it tested that the column existed, not that it carried values on the active rows the number uses. It now tests those rows. Same day: the short caveat now states that PJM’s signed figure is read by keyword from free-text status, a weaker read than a named field.
Aug 23, 2026Correction on /ai-load. The page's opening line said the NYISO load tab carries an end-use code “for every request”. It does not: in the July 31, 2026 edition, 22 of the 73 rows carry no code — all 17 withdrawn rows plus 5 not-withdrawn rows (4 already in commercial service, 1 in study). The page's own reconciliation section already stated the exception; the lead overclaimed past it. The line now reports the counted figure (51 of 73), derived from the file rather than typed. No published MW figure changes: the 5,290 MW and 13,568.9 MW totals were always computed over the coded rows only.
Aug 21, 2026A correction we had to correct, and a rule we broke. Earlier the same day we moved the LBNL page cite from p.55 to p.35, announced here that “the page was not” right, and — against the rule stated directly above this table — edited the dated Jul 19 row to match. Re-reading the full report showed p.55 was correct all along: it carries the sentence we paraphrase, verbatim (“Only 13% of capacity that submitted interconnection requests from 2000–2020 had reached commercial operations by the end of 2025; 75% of that capacity had been withdrawn and 10% was still active”). Page 35 is a chart carrying the same figures in different words — supporting evidence, not a replacement. Both changes are reverted and the original row restored. The figures themselves never changed. The method error is worth naming: we confirmed that p.35 supports the fact and inferred, without ever testing it, that p.55 did not. A positive finding on one page is not evidence against another. One further note that does stand: 13% and 75% are shares of capacity — the same source carries 19% and 71% by number of requests, and swapping the two would be easy and wrong.
v1.3 · Jul 20, 2026Fixed a second double-count, this one in the signal score. A single source document that matched two signal types was counted twice — measured: 148 of 231 utility_expansion signals were a duplicate of an event already carrying its correct type. Since 11 Jul the per-document de-duplication applied only to the diversity multiplier; it now applies to the score itself. One document counts once, at its strongest signal; the volume bonus still rewards distinct documents agreeing. Effect: regional scores fall a median of ~19% — that drop is the removal of inflation, not a loss of evidence. Weights (35/30/20/15) unchanged.
Jul 20, 2026Track-record verdicts made honest. Sealed calls that mature with no observable record in their window now resolve as unresolved — not as a miss. Absence of evidence is not error, and counting it as one would have published failure where no judgement took place; unresolved calls appear in the Journal and count in no accuracy rate. The horizon for new seals moved from 180 days to 540, to match the physical timeline of a data center (2–4 years); already-sealed calls are immutable and keep their original horizon.
Note (Oct 2, 2026): this rule changed after some calls had been sealed. When the track record is scored, it will be published under both rules: the rule in force when each call was sealed, and this one.
Jul 19, 2026The public sentence was rebuilt around the per-queue spread (the signed share differs enormously between grid operators) and the LBNL benchmark was updated to the 2026 Edition (13% operational / 75% withdrawn, 2000–2020 cohort, p.55), superseding the 2025 edition figures.

The headline sentence was rewritten to report the per-queue spread as a range with the named queues, moving the full per-queue table to /verify, and to state plainly that in the queues where our reading finds no machine-readable agreement field we cannot produce a signed share at all — silence in the record, which is not evidence that an agreement is missing.
Note (Oct 2, 2026): two entries were written for this date; they are merged here and both texts are kept as written.
Jul 16, 2026Correction, published in the open. Our reading of “signed agreement” had been keyword-based across all queues; six of seven operators in fact publish a named status column, which we now read directly. The ratio we had published (“50.5×”) was retracted and the headline number withheld until the corrected reading passed review. This was the largest correction in the number's history — logged here rather than quietly replaced.

Correction, and a retraction. Our per-queue reading of executed-agreement status was wrong for several queues: we had been reading a single status text where those operators publish a named column. The previously published national ratio (~50×) was withdrawn and the number was held offline until the corrected reading was re-derived and re-published. Method epoch for the corrected reading: 16 Jul 2026.
Note (Oct 2, 2026): two entries were written for this date; they are merged here and both texts are kept as written, including the ratio, which they round differently.
Jul 15, 2026The number was placed on hold after an internal audit found the per-queue reading above. Nothing was published while it was held — the surfaces showed the method, not a figure.
Jul 12, 2026Signal detection sharpened: cooling now catches cooling and water agreements (a physical-commitment signal), and financing catches joint-venture / fund-backed structure (committed capital, not just a headline dollar figure). Detection only — the CFI weights (35/30/20/15) are unchanged.
v1.2 · Jul 11, 2026Fixed a double-count: the queue's signed share had informed both the Energy and the Corroboration pillars. It now counts once (Energy); Corroboration measures independent convergence. Weights unchanged.
Jul 11, 2026Relabeled the national queue figure as the generation interconnection queue (it had been labeled “large-load”). Predictions sealed before this date carry a vintage note and stay immutable — disclosed, not rewritten (see §6).

The method →