Post from Truth Social

Six headlines before 7 a.m.

Nothing in this post is written in the subject's own voice. "Keep the lights on, sideline Big Brother: Trump EPA rollback could save ratepayers $310 billion" is a publication's headline construction — imperative clause, colon, quantified claim — followed by a bare URL. It is the sixth Just the News link in the same morning's run. One of the earlier five carried an added interjection ("Fantastic news for America!"); this one carries nothing added at all. The 10:32 UTC stamp is 6:32 a.m. Eastern, which is the overlap hour rather than a discriminating one: he posts then, and so does staff. What settles the reading is the absence of a single first-person clause — there is no sentence of his own for disorganization to show up in, which is what removes the usual authenticity evidence from play. Beyond that, the colon-and-dollar-figure headline is reproduced intact, and he appears in it in the third person ("Trump EPA rollback"), the way he appears in copy about him rather than copy by him.

The number in the headline is not the number in the link

The top line promises $310 billion to ratepayers. The URL beneath it ends roots-out-30-billion-waste-fraud-and-abuse — a different figure attached to a different subject, agency waste rather than utility bills. Nothing in the post reconciles them, and a reader who stops at the headline takes away a number an order of magnitude above the one in the address bar. The likeliest explanation is an edited headline over a stale slug, which makes this a hazard of forwarding other people's copy rather than an assertion he would have to defend. That is also why it is hard to pin: the claim is endorsed by selection, not stated.

Regulation rendered as an eye

"Big Brother" is the only rhetorical load the post carries, and it does the heavy work. An emissions or efficiency rule is recast as an instrument of watching, and the thing set against it is your electricity bill — "Keep the lights on." The pairing sorts the field into two parties, the ratepayer and the watcher, with the agency's technical function dropped out of the frame entirely. The grievance is real in shape but has no occupant: not one person is named anywhere in the post, so the fear appeal floats free of anyone who could be acted against, and the targeting read that "sideline" might otherwise invite has nothing to attach to.

Credit as the point of the exercise

The selection logic across the morning is consistent: courts vindicating him, arrests validating him, a UN speech realigning the world, and now an agency saving the public money — each headline an external voice attributing an outcome to his name. The motive served is status through third-party attribution, which is cheaper and more durable than first-person boasting. The narrative arc inside the headline is cleanup: "waste, fraud and abuse" on one side, savings on the other, with himself as the agent of the transition.

A pasted headline supplies no syntax of his own, so there is nothing here to weigh against earlier language samples; what the day does show is an orderly run — six distinct stories, each a separate story, arriving in sequence without repetition.

Fact Verification

I downloaded and parsed EPA's own Regulatory Impact Analysis for the final rule (utilities_ria_final_partial_repeal_of_cps_2026-09.pdf). The $310 billion appears in exactly one place: Table 5-4, labeled "General Equilibrium Social Costs," present value 2026 to 2047, in billion 2024 dollars. It is output from SAGE, EPA's computable general equilibrium model, and is defined in the RIA as "changes in real full consumption" across the whole economy, discounted using "the internal consumption discount rate in the model" (endogenous, household-specific rates), not a standard 3% or 7% rate. The RIA explicitly cautions that "SAGE does not account for the effects of changing environmental quality as a result of this final action."

EPA's own accounting produces four different numbers, and $310 billion is the largest and the one furthest removed from a utility bill. RIA Table 1-1: power-sector compliance cost savings, present value 2026 to 2047, $160 billion at a 3% discount rate and $95 billion at 7%. Section 1.2: real resource cost savings of $280 billion (3%) and $180 billion (7%). Table 6-1, which is EPA's formal costs-and-benefits summary for the rule, uses $160 billion / $95 billion and lists climate, health and water impacts as non-monetized. The $310 billion never enters EPA's official benefit-cost table.

EPA's own projections also contradict a simple "lower bills" reading in the near term. RIA Table 1-3 projects retail electricity prices relative to keeping the 2024 standards at plus 0.7% in 2030, minus 5.8% in 2035, minus 1.1% in 2040, and minus 2.6% in 2045. Akshaya Jha, associate professor of economics and public policy at Carnegie Mellon, writing in The Conversation, puts the 2035 figure at roughly $8 a month for a household using 1,000 kWh. EPA also declined to monetize forgone climate and health benefits, citing Executive Order 14154; its own 2025 proposal analysis had estimated $130 billion in health costs from additional fine particle and ozone exposure over 2026 to 2047, and the 2024 rule it repealed had been scored at roughly $370 billion in net climate and health benefits. RIA Table 1-2 projects CO2 emissions increases of 20 million metric tons in 2030 rising to 533 million in 2040.

Note also that the linked Just the News article's own subheadline says the EPA "will save taxpayers $310 billion through 2047" while its headline says "ratepayers" — two different and both inaccurate framings of a general-equilibrium consumption measure. The repeal is under challenge: EDF, NRDC, the American Lung Association, the American Public Health Association, Clean Air Council and Clean Wisconsin filed suit in the D.C. Circuit on September 17, 2026.

Verdict rationale: a real, correctly quoted EPA projection, hedged appropriately with "could," but attached to the wrong beneficiary and the wrong cost category, and cut against by EPA's own near-term retail price path. | | "The EPA under Trump rooted out $30 billion in waste, fraud and abuse." | Mostly False | Two separate problems: the claim is not in the article at all, and the underlying characterization has failed in court and before EPA's own watchdog.

First, provenance. The $30 billion appears only in the URL slug ("trumps-epa-gets-action-roots-out-30-billion-waste-fraud-and-abuse"). I checked the Internet Archive CDX index for that URL: the earliest capture is 2026-09-22 03:52 UTC, and its embedded NewsArticle schema already carries the headline "Keep the lights on, sideline Big Brother: Trump EPA rollback could save ratepayers $310 billion," published 2026-09-21. No earlier article ever lived at that address. The slug is a content-management leftover from a discarded working headline. I read the full archived body text: it discusses the Carbon Pollution Standards repeal, the methane rule, MATS, the LNG permit pause and winter grid risk. It contains no reference to $30 billion, to waste, or to fraud.

Second, substance. EPA does make a roughly $30 billion claim, so the number is not invented. EPA's own May 2026 release describes "Administrator Zeldin's cancellation of roughly $30 billion in wasteful grants and contracts across the agency, including the $27 billion in GGRF money." Zeldin used $29 billion in a Just the News podcast interview, and Fox Business ran "EPA Administrator Lee Zeldin reveals $30B in agency budget savings." But "rooted out waste, fraud and abuse" describes detection of wrongdoing, and that is where the record breaks:

• On August 4, 2026, the full 10-judge D.C. Circuit sitting en banc held that EPA wrongly terminated $20 billion in Greenhouse Gas Reduction Fund grants "based solely on a policy disagreement." That $20 billion is the bulk of the $30 billion.
• District Judge Tanya Chutkan found that when the government was asked for evidence of fraud, EPA did not provide it and shifted to general concerns about oversight levels. In related proceedings DOJ was "unable to proffer... any information with regard to any kind of investigation or malfeasance."
• Denise Cheung, chief of the criminal division at the U.S. Attorney's Office for D.C., resigned rather than order the accounts frozen, citing insufficient evidence of criminal activity. A federal magistrate rejected a seizure warrant application.
• EPA's own Office of Inspector General, in a March 2026 report, found that the roughly $1.5 billion in Community Change Grants terminated in May 2025 were awarded properly, identifying no fraud and no waste.
• Zeldin's criminal referrals to DOJ and the IG have produced no prosecutions of grant recipients.
• Part of the GGRF money was rescinded legislatively by the One Big Beautiful Bill Act in July 2025 — an act of Congress, not an EPA fraud finding.

A countervailing point: an EPA OIG audit did find the Biden administration issued poorly structured grants with weak oversight, putting figures in the tens of billions "at risk," and a January 2026 Solar for All audit found monitoring deficiencies. Money at risk of waste through weak controls is a real finding, but it is a different assertion from waste, fraud and abuse having been identified and rooted out.

Verdict rationale: a real EPA talking-point figure, but it counts canceled grants and contracts (a policy choice, the largest slice of which an appellate court held unlawful) rather than substantiated waste, fraud or abuse — and the claim does not appear in the cited article at all. | | "Sidelining EPA regulation is necessary to "keep the lights on" — i.e. the rules threatened electricity reliability." | Mostly False | The independent reliability authority does not name EPA rules as a driver, and EPA's own analysis contains no reliability modeling, though the rulemaking preamble does make a forward-looking reliability argument that rests on other people's demand-growth findings.

I downloaded NERC's 2025–2026 Winter Reliability Assessment (the assessment the article itself cites for the coming winter) and searched the extracted text. It contains zero occurrences of "EPA," "Environmental Protection Agency," "greenhouse gas," or "Carbon Pollution." What it does name: "Two trends affecting resource adequacy across the BPS for the upcoming winter are rising electricity demand forecasts and a continued shift in the resource mix characterized by the retirement of thermal generators and growth in battery resources." Aggregate peak demand is up 20 GW (2.5%) year over year against a 9.4 GW resource increase; regional drivers listed are data centers, residential and transportation electrification, commercial and industrial growth; and the acute winter risk is constrained gas pipelines and fuel depletion in extended cold. NERC's January 2026 Long-Term Reliability Assessment tells the same story: 18 GW of new data center capacity, summer peak demand growth revised up to 224 GW, 105 to 115 GW of retirements, and margins falling below the reference level starting in 2029.

EPA's own Regulatory Impact Analysis performs no reliability analysis. The word "reliab" appears three times in the entire document, all incidental: two in describing IPM's internal constraints and its use of NERC's assessment as a data input, one in a sentence calling a gas market model "reliable and efficient." And EPA's near-term projections are negligible: for 2030 the repeal yields CO2 up 20 million metric tons, coal production up 4%, and retail electricity prices 0.7% higher, not lower. Nothing in EPA's modeling supports an effect on this winter, which is the frame the article uses when it invokes Winter Storm Uri and the 2014 polar vortex.

The 2024 rule was also not operating on any near-term timeline. Its compliance dates were 2030 (a 40% gas co-firing rate for coal units retiring by 2039) and 2032 (90% capture CCS for long-term units). D.C. Circuit litigation was held in abeyance from February 2025 onward at EPA's own request, so nothing was being enforced. EPA's 2024 reliability fact sheet documents built-in safeguards: a reliability assurance mechanism letting a regional administrator extend a unit's operation by up to a year, and state flexibility to opt units out on age or reliability grounds. EIA's Annual Energy Outlook projects roughly 70 GW of coal retirements even with the greenhouse gas rules gone, against 100 to 125 GW with them — retirement continues either way.

In fairness to the other side, the rule preamble signed by Zeldin does argue reliability, and it is not baseless. It cites NERC's revision of summer peak growth from 0.65% to 1.67% CAGR, ERCOT's 8.9% and PJM's 3.1% forecasts, the One Big Beautiful Bill Act's wind and solar subsidy phaseout, and DOE's July 2025 Resource Adequacy Report claim that outage risk could rise "a hundred fold" by 2030. Note what that list is made of: the reliability pressure is attributed to load growth, retirement timing and tax-credit changes, exactly as the first-pass analysis held. And the DOE 100x figure has been contested on methodology by GridLab (DOE assumed 104 GW of 2030 retirements where EIA data showed about half that, and ignored data-center load flexibility), by Advanced Energy United, the American Clean Power Association and ACORE in a rehearing request, and by state attorneys general, with critics noting DOE used a deterministic rather than probabilistic framework.

Verdict rationale: there is a real, contested policy argument that CCS mandates would compound retirement pressure in the 2030s. There is no support for the headline's implication that the rules were threatening the lights, and the reliability warnings actually on the record are driven by demand growth and retirement timing, not by the repealed rule. |

Overall Veracity: 30%