On August 3, 2026, Governor Greg Abbott directed the Public Utility Commission of Texas and ERCOT to verify and audit every data center advancing through ERCOT’s interconnection process before any more are approved. Seventeen days later, the commission made it operative at its August 20 open meeting, and ERCOT paused the Batch Zero study of large loads of 75 MW and more. The freeze has had no end date since — until now. ERCOT’s audit and community-impact reports go to the commission on December 10, and the December 17 open meeting is the earliest the audit can be treated as complete. For four months, the fastest queue in American power has been a paperwork queue. December 10 is when the market finds out how much of it survives.
The loads nearest the gate are the ones that make the freeze concrete. Seventeen data center and crypto loads carrying 6,608 MW of peak demand have finished every ERCOT process except approval to energize. They are not waiting on a study, a model, or a wire. They are waiting on a verification report. ERCOT has paused their approvals to energize until the audit is complete, and a load that fails verification, or does not respond, is disqualified from Batch Zero. That is the shape of the freeze: not a ban on building, but a gate that moved from engineering to evidence.
The pause widened beyond the grid in September. On September 21, Abbott directed the Texas Commission on Environmental Quality to stop issuing permits sought by data center projects until ERCOT completes its audit and the Texas Water Development Board finishes its municipal water impact evaluations, with a compliance report due October 19. A separate September 14 directive told the water board to enforce mandatory water-use reporting for data centers and coordinate with ERCOT on joint resource assessments.
The queue arithmetic
Start with the numbers, because they are routinely confused with each other, and the confusion is the story. ERCOT was considering approximately 474 GW of large-load requests to connect to the Texas grid as of June 2026, with roughly 90 percent of them from data centers. That figure is a queue of requests, not a demand forecast, and ERCOT does not treat it as one. Against it, ERCOT set a preliminary all-time peak demand record of 91,134 MW on July 22, 2026, in hour ending 18 — 5,581 MW above the previous record of 85,508 MW set in August 2023. The queue asked for more than five times the power the entire state has ever used at once. The queue is not a plan. It is a pile of requests, and the audit exists to sort the pile.
The September Batch Zero study did the first sort. ERCOT conditionally included 362 projects totaling about 194.3 GW — 204 as base load, about 66.4 GW, and 158 as studied load, about 128 GW — and excluded 373 projects totaling about 302 GW. The conditional classifications went to transmission and distribution service providers on September 3, in Market Notice M-A080326-04, after ERCOT had told the market on August 3 that it would miss the original August 7 classification deadline. Keep the two figures distinct: 474 GW is what was asked for, 194.3 GW is what the September study conditionally kept. The gap between them — roughly 280 GW — is the first thing the audit has already done, before a single report is filed.
Conditional is the operative word. A conditional classification is not a final one: the market notice says a large load that fails a stated condition may be disqualified, and ERCOT’s August 20 presentation to the commission says loads that do not pass verification, or do not respond, will be disqualified from Batch Zero. Final classifications are expected to follow the December reports, and the Batch Zero study itself — originally targeted to deliver results April 9, 2027 — will not meet that date. The study resumes after the audit, on a timeline ERCOT has not yet set.
The queue’s own history says the requests overstate the load. ERCOT’s 2025 long-term load forecast report found that sites reached, on average, 49.8 percent of the megawatts they had requested. And of the roughly 9 GW of large loads already approved to energize, observed peak consumption runs closer to 3.9 GW — about 40 percent of approved capacity, as reported from ERCOT data in October. A queue position is not a load. It is an option on a load, and the audit is the state repricing that option.
What the audit actually asks for explains why it takes four months. ERCOT’s verification request went to 465 loads, asking about water and energy use, ownership, on-site generation, cooling, and community impacts. As of ERCOT’s November 5 update to the commission, 433 had responded, covering about 177 GW, and a second round of developer-specific verification questions was under way. The gate is not a single form. It is a dossier, checked question by question.
Two reports, one meeting
December 10 is the date because two reports are due that day: the Batch Zero Eligibility Verification Report, which determines which large loads of 75 MW and above remain eligible for Batch Zero, and the Community Impact Review Report, which summarizes the water, power, ownership, and community-impact information collected from data centers. Both are slated for discussion at the commission’s December 17 open meeting. That meeting is the earliest date the audit can be treated as complete and the permit pause could lift, and ERCOT expects to file any good-cause exception requests — including any relating to Batch Zero timing — by then.
What happens after December 17 is the part that should not be assumed. A project that clears the audit in December still needs ERCOT to restart approvals to energize and the state to lift the permit freeze, and both of those decisions will fall to the incoming administration, which takes office in January. The December reports determine who keeps a place in line. They do not, by themselves, put power in anyone’s hands.
The other December decision
The audit is not the only December decision in Texas, and the other one may matter more over the long run. Senate Bill 6, passed in 2025, requires the commission to amend its rules by December 31, 2026, to ensure that wholesale transmission charges appropriately assign the cost of transmission investment. The vehicle is Project No. 58000, in which the commission proposed rules in July: replacing the four coincident peak (4CP) methodology — which allocates transmission costs on the basis of four summer peak hours — with a twelve coincident peak (12CP) methodology that measures peaks across all twelve months, and creating a new contractual take-or-pay transmission rate for customers above 75 MW, under which costs are allocated on contracted capacity rather than actual peak usage, for two decades, with the obligation kicking in when the transmission is ready rather than when the customer is.
The money at stake is the reason the rule exists. ERCOT’s 2024 Regional Transmission Plan identified approximately $33 billion of transmission investment, including the Permian Basin Reliability Plan and the 765kV Strategic Transmission Expansion Plan. Oncor, in a November analysis of the Project 58000 reforms, modeled a scenario with roughly 60 GW of new large-load demand and estimated residential transmission cost responsibility would fall by about $35 per customer per year under the proposed framework, against about $15 under a 35 GW scenario. Oncor is an interested party — it owns much of the wire in question — but the direction of the arithmetic is the point: under the current allocation, the cost of wires built for new load lands substantially on existing customers.
That is the stranded-cost problem Senate Bill 6 was written to solve, and it is the mirror image of the queue problem. A 474 GW queue of requests, each of which may or may not become a load, is also a queue of transmission projects that may or may not be needed. The September 18 interconnection rule — adopted under the same bill, effective October 8 — already makes queue positions expensive: new loads of 75 MW or more post $50,000 per MW of financial security before ERCOT will study them, plus a $100,000 study fee and size-based study charges. The December cost-allocation rule would make the load itself expensive to serve if it arrives and cheap to abandon if it does not — or, more precisely, it makes abandonment expensive, which is the point. Take-or-pay for twenty years is a commitment device, and commitment is what a queue of options has been missing.
What the freeze does to the market
Here is the test that separates this event from an ordinary permitting delay: ERCOT is an energy-only market with no capacity market, and its product, for a decade, has been speed to power. A developer could request interconnection, get studied, and energize on a timeline measured in months rather than years, in a market where the price of power is set hour by hour and there is no capacity payment to smooth the risk. The queue was the market’s forward curve — the visible pipeline of future demand, priced into generation investment, transmission planning, and land deals across West Texas. The freeze did not change the price of power. It changed what the queue means. A place in line is no longer a forecast. It is a claim, and the claim is now being audited.
The congestion story this desk covered on November 14 is the reason the freeze bites where it does. West Texas swung $184 in a single day — from -$1.19/MWh at hour ending 11 to $183.08 at hour ending 23 on November 3 — because one 138kV line in the Stanton corridor bound in both directions, exporting solar at midday and importing power at night. The Permian Basin reliability plan’s answer is steel: ERCOT’s first 765kV lines, approved by the commission in April 2025. The freeze does not build that steel. It pauses the load side of the equation while the wires catch up — and while the commission decides, in Project 58000, who pays for them.
For developers, the practical math has already changed, and BloombergNEF’s estimate gives it a scale: the audit could delay 49.8 GW of data center load, nearly 20 percent of the U.S. pipeline it tracks. A project that posted $50,000 per MW in security to be studied, and now faces a twenty-year take-or-pay transmission obligation if it proceeds, is a project whose economics were written under one set of rules and are being repriced under another. Some will proceed. Some will not. The December reports will show which, project by project, and the market will read them as a list of survivors.
What it means for buyers
For a buyer of power in Texas, the freeze rearranges the risk rather than removing it. A data center’s energization date was already the most consequential number in any West Texas power agreement; it is now a regulatory date, set by an audit report and a commission meeting rather than by a construction schedule. Power purchase agreements and retail contracts priced on energization dates carry schedule risk that no hedge covers, because the hedge settles on price and the delay is in time.
For everyone else in ERCOT, the December decisions set the terms on which the next decade of load growth arrives. If the audit disqualifies a large share of the queue, the transmission plan built on 474 GW of requests gets smaller, and the $33 billion figure gets re-examined. If the cost-allocation rule lands as proposed, new large loads pay for the wires built for them, and the residential share of transmission cost falls.
What to watch
December 10: the two reports — which loads passed verification, which were disqualified, and what the community-impact review found on water, ownership, and local effects. Read the disqualification list alongside the 194.3 GW that survived September: the gap between the two is the audit’s real output.
December 17: the commission’s open meeting — the first action on which facilities keep a place in the Batch Zero line, and the earliest the permit pause could lift. Watch whether ERCOT files good-cause exceptions on Batch Zero timing, and what they say about the study’s new schedule.
December 31: the statutory deadline for the Project 58000 rule amendments. The commission proposed the 4CP-to-12CP shift and the take-or-pay framework in July; whether the final rule keeps both, and in what form, determines who pays for the $33 billion build.
Texas built the fastest interconnection queue in American power by treating a request as a plan. For four months it has treated a request as a claim to be verified. December 10 is the day the verification comes back — and the day the market learns whether the thaw is a resumption or a reckoning.