Corrections: what we got wrong

What we said: A set of claims in our funding-transparency answer were stronger than the evidence behind them, and they overstated in both directions. We said the industry's trade association runs campaigns in at least eight states whose spokespeople decline to say who pays: the source names seven, and the association's president did answer questions. We called two opposition groups the most transparent finances in the debate, when neither names a single donor. We said foundation grant lists contain no data center groups, which rebuts an accusation nobody made while the entity actually named in it publishes no list at all. We credited a company's funding to tax filings that name no donor. And two source citations carried headlines that were paraphrases of web addresses rather than the headlines the pieces were published under.

What it says now: Each claim says what the document supports and no more, including that one of the newsrooms we cite discloses more than the two groups we had called the most transparent. The citations carry the real headlines.

How it was caught: Our own audit. Worth stating plainly: the overstatements ran in both directions, one favouring each side, which is what you would expect from carelessness rather than bias, and is exactly why we check rather than trusting that our intentions were even-handed.

What we said: Correcting our own correction, made the same day. Having removed an unsourced claim that the plants were built to run 30 to 40 years, we replaced it with depreciation periods of 20 and 32 years. Those figures are real but belong to a different case covering different plants. We had just fixed an error of attaching figures to the wrong plants, and then made it ourselves.

What it says now: What the filing for these plants actually says: Entergy modelled a 30-year depreciable life, and the commission set no depreciation rate when it approved them. The contract runs 15 years with automatic five-year renewals, and the contract itself is under seal, so the often-quoted 30-year ceiling is the opposing parties' reading rather than a documented term.

How it was caught: A second, deeper check of the same docket, run because the first fix relied on a filing we had not read end to end. The lesson is the plain one: a correction is a change to the site and deserves the same sourcing as the thing it replaces.

What we said: We wrote that the commission's own judge subpoenaed Meta for its power demand and job numbers, that Meta fought it, and that the commissioners overruled their judge 3 to 1. The vote and what Meta was asked for were right, but the framing was wrong in three ways: the judge did not act on her own, she was responding to a motion from groups opposing the deal, and she granted only two of their five requests; Meta was not alone in opposing it, because the utility and the commission's own staff opposed it too; and the missing fifth vote was an absent commissioner, which we left unexplained.

What it says now: The entry says who asked for the subpoena, that three of five requests were refused, who else opposed it, the date of the vote, and that the fifth commissioner was absent with no reason recorded. The underlying facts were strong enough without the framing.

How it was caught: Our own audit, against the commission's transcript and the judge's referral rather than press coverage.

What we said: Several numbers around the Louisiana project were loose. We said the gas plants are built to run 30 to 40 years, which came from an opinion column and is contradicted by the utility's own filing. We attributed a $21 billion package to seven gas plants when it also covers batteries and transmission. We placed a $3.4 billion stranded-cost estimate on the wrong plants. And we described a state senator as buying land near the site, which is the weaker half of what the reporting found and omitted that he has held much of that land for over a decade.

What it says now: The filing's actual depreciation periods, the real scope of each dollar figure, the correct case for the stranded-cost estimate along with the fact that the sworn version of it is redacted, and the documented core of the land story: a sale to the utility agreed four weeks after a vote he had lobbied for, carried together with his denial in his own words.

How it was caught: Our own audit. Two of these errors ran against the side we were quoting, which is its own warning: sloppiness is not the same as slant, and it is not automatically safer.

What we said: When we generated the new funding tags, two sources were credited to the wrong payer. A nonprofit newsroom's report on a state ethics case was tagged as paid for by the data center developer and the technology company it was reporting on, and the tax filings and reporting that exposed one company's funding of an advocacy group were tagged as paid for by that same company. The page said so in plain words, 'paid by', next to each citation.

What it says now: Both say who actually paid for the document: foundations and advertising for the first, philanthropy for the second, with the money the reporting is about moved into the note where it belongs. Neither is now shown as compromised.

How it was caught: Our own audit, the same day the tags went live. Implying a conflict of interest that does not exist is the one error this site cannot afford, and generating funding labels in bulk reintroduced exactly that. The nightly review now flags any source that credits a company with paying for journalism about that company.

What we said: We named Bloomberg Philanthropies as one of four funders of the nonprofit newsroom behind the Louisiana land-deal investigation. Bloomberg Philanthropies does not appear anywhere on that newsroom's disclosed funder list. We also listed four funders when the newsroom discloses thirty-one, and named two of its five co-publishers.

What it says now: The note gives the disclosure as it stands: thirty-one funders of $5,000 or more, including the Bloomberg-launched Beyond Petrochemicals campaign, which is how that money actually reaches them, and Armature Philanthropy at Foundation for Louisiana, which funds work in this state specifically. It also says that several entries are donor-advised funds whose underlying donors are not disclosed, and names all five co-publishers.

How it was caught: Our own audit. This is the error this site exists to catch in other people's work, made in our own: a funder named who was not a funder, and a short list presented as the list.

What we said: We described the site that republishes the Census figures for Richland Parish as 'a nonprofit funded by journalism grants'. It is an independent open-source project incorporated as an Illinois not-for-profit, and we could find no federal tax filings for it, so we should not have implied charitable status.

What it says now: The note says what it is: an independent open-source project built with a Knight Foundation grant and later Google News Initiative support, with development and hosting help from two universities, republishing federal statistics rather than producing them.

How it was caught: Our own audit of sequential source numbers.

What we said: Our funding note for a Louisiana news source described Straight Arrow News as paid for by its founder plus paid memberships, and left out that it also carries advertising for anyone who has not subscribed.

What it says now: The note says it is funded by its owner, by subscriptions and by advertising shown to non-subscribers, and points out that the outlet markets itself as serving readers rather than advertisers while doing so.

How it was caught: Our own audit of two clusters of sequential source numbers, checking for exactly this kind of drift.

What we said: We said Virginia's data center tax exemption returns 48 cents per dollar given up, 'nearly three times the 17-cent average of the state's other sales tax exemptions.' The audit's own table compares it with 41 cents for the average Virginia incentive, and the report rates the exemption 'moderate' and 'similar to' that average. The figure 17 cents appears nowhere in it.

What it says now: 48 cents against 41 cents, with the auditors' own characterisation. The exemption does beat the average, but not by three times, and it still returns less than it gives up.

How it was caught: Our own audit. A first check appeared to confirm the 17-cent figure from an opinion column that cited the audit; reading the audit gave the opposite answer.

What we said: Ten sources added during a renumbering were never linked to the claims they supported, while the pages needing them cited unrelated sources. Clicking a citation on the data center noise question produced a story about county property taxes, and the questions about the AI race, which quote Stanford's AI Index, cited two articles about water scarcity in Oregon and Arizona.

What it says now: Each of those pages cites the sources that actually support it. The nightly review now flags any source that no page cites, which is the cheap half of the same signal.

How it was caught: Our own audit, prompted by Jon asking us to look for more errors of the kinds already found.

What we said: We said the industry's commissioned study modelled 'over $200 billion' in annual tax contributions.

What it says now: The study reports $162.7 billion in federal, state and local tax contributions in 2023, its most recent year.

How it was caught: Our own audit.

What we said: We gave 700,000 litres as the freshwater cost of training one AI model without saying that this is the water evaporated at the building only.

What it says now: Both figures, labeled: about 700,000 litres on site, and about 5.4 million litres once the water used to generate the electricity is counted. They differ by roughly eight times, and the smaller one is the one that travels.

How it was caught: Our own audit.

What we said: We described Virginia's new rules for large customers as requiring them to pay for '60% of generating capacity reserved, used or not', which reads as a block of power plants set aside for data centers.

What it says now: The 60% figure is the share of minimum charges a customer must guarantee with collateral, and only when its credit is not strong enough on its own. It is a credit requirement, not reserved generation.

How it was caught: Our own audit.

What we said: We said a PJM capacity auction price went from $28.92 to $329.17, 'more than ten times higher.' That pairs the starting point of one auction cycle with the result of a different one.

What it says now: The actual sequence is $28.92, then $269.92, then $329.17, then the $333.44 ceiling. The first jump is more than ninefold, which is dramatic enough without combining two auctions.

How it was caught: Our own audit.

What we said: We repeated that Microsoft's newest data center designs 'use no water at all' for cooling in normal operation, and that Google has committed to the same by 2030, citing an article that says neither.

What it says now: The site states the reduction Microsoft does report, and says plainly that the stronger versions of the claim are widely repeated but are not in that source.

How it was caught: Our own audit.

What we said: We said an electricians' union local worked 28 million hours, 'double what they worked a decade earlier', with '600 new apprentices in a single year'. The cited newspaper article contains neither of the latter two figures.

What it says now: Only what the article supports. The other two are left out until they can be sourced.

How it was caught: Our own audit.

What we said: We gave Ohio's tax exemption cost as '$1.6 billion in 2025, roughly $330 per Ohio household' in a way that implied both figures came from the organization cited. The per-household number was our own arithmetic.

What it says now: The page says the per-household figure is this site's own division of the state total, and is a way to feel the size of it rather than a bill anyone receives.

How it was caught: Our own audit.

What we said: We credited a flexible-load pilot with '9 sites, 49 participants' including Meta and Microsoft. The launch announcement names nine organizations and neither of those two, and gives no participant count.

What it says now: Nine sites, and the nine organizations actually named, with a note that the roster changes.

How it was caught: Our own audit.

What we said: We labeled $64 billion of blocked and delayed projects as a 2025 figure. The tracker's own window runs from 2024 into early 2026. We also stated flatly that the tracker's clients are AI companies.

What it says now: The label gives the real two-year window, and the funding note reports what the tracker says about its own independence, marked as its position rather than an audited fact.

How it was caught: Our own audit.

What we said: Two different projections of 2030 data center electricity use were presented as one range, mixing a sensitivity analysis with a compounded-growth scenario.

What it says now: Both are shown as separate rows with their methods named: 578 to 782 TWh, and 521 to 843 TWh.

How it was caught: Our own audit.

What we said: The size of the Prince William Digital Gateway project was given as 20 million square feet in one place and 23 million in another.

What it says now: About 23 million square feet throughout, matching the approved application.

How it was caught: A check for numbers that contradict themselves across pages.

What we said: Several figures cited documents that do not contain them: two electricity-sales forecasts cited a press release rather than the outlook table they come from, and a PJM demand-growth figure cited a news article about auction prices rather than PJM's own load forecast.

What it says now: Each cites the document the number is actually in, including a newly added citation for PJM's load forecast.

How it was caught: Our own audit.

What we said: Four specific figures could not be found in any cited source: an opposition-group count available only in a paid report, an analyst estimate that 30 to 50% of the 2026 pipeline was at risk, a date attached to a Georgia rate estimate, and four source links that had gone dead.

What it says now: The unsupported figures are removed or attributed to whoever claims them. The dead links are replaced or the sources dropped.

How it was caught: A link check across every source, plus our own audit.

What we said: We removed a true claim by mistake. Finding that the figure 'the US holds about three-quarters of world AI compute against China's 15%' was absent from the source cited for it, we deleted it from four places. The claim is real and was already on this site under a different source, which states exactly that.

What it says now: Restored, cited to the research group that actually measured it, and narrowed to what they measured: share of global GPU cluster performance, by where the hardware physically sits. A neighbouring forecast of US capacity growing to 152 gigawatts by 2030 really is unsupported and stayed out.

How it was caught: Re-checking our own correction before publishing it. Deleting a supported claim is the same kind of error as publishing an unsupported one.

What we said: A published link had a fragment of HTML stuck to the end of it and did not work, and two different documents were stored under the same report identifier.

What it says now: Both fixed, and the daily updater now sanitises URLs and rejects duplicate identifiers rather than trusting that a string starting with 'http' is a usable link.

How it was caught: The first run of the nightly review job.

What we said: Fifteen footnotes on the two argument pages linked to the wrong sources. The handbook and the site had each been numbering sources independently from 31 onward, so a point citing note 32 sent readers to an unrelated source 32 on the sources page.

What it says now: The handbook and the site share one numbering system. A note number in the handbook is now the same source number on the site, and the nightly review checks that they still agree.

How it was caught: Jon asked where the site delivers its promise to say who funded each source, which exposed the fork.

What we said: A funding note had been copied from another source and not rewritten. It said the consultancy E3 had also produced a study paid for by the Data Center Coalition, extending that conflict of interest to a Virginia legislative report E3 never worked on. The citation also pointed at a Department of Taxation document while claiming to be a legislative audit.

What it says now: The note describes only what that document's own evidence supports, and the citation points at the audit it names. Implying a conflict of interest that does not exist is the worst error this site can make.

How it was caught: Our own audit of every funding label.

What we said: The most-cited number on the pro side was wrong. We said the industry's own commissioned study found 5.5 million supported jobs and 4.5 additional jobs for every direct job.

What it says now: The study reports 4.7 million supported jobs and an employment multiplier of 7.5, meaning 6.5 additional jobs per direct job. We overstated the total by 800,000 jobs while understating the multiplier.

How it was caught: Reading the report itself instead of our own summary of it.

What we said: Counts that were typed by hand had drifted: the site advertised a 39-page handbook when the PDF was 42 pages, said 'Seven arguments' when there were eight, and 'Fourteen arguments' when there were sixteen.

What it says now: Every count is derived from the handbook at build time, so it cannot drift again.

How it was caught: A question from Jon about whether the PDF was being kept current.