On 3 August 2026 Governor Greg Abbott directed the Public Utility Commission of Texas and ERCOT to complete a comprehensive verification and audit of all data centres advancing through the interconnection process before allowing those loads to proceed. ERCOT's response was immediate and mechanical. The same day, the operator issued a Market Notice pausing the Batch Zero study process and delaying final classification of all large loads. It also paused approvals to energise data centres or virtual currency mining facilities that are 75 MW or greater. On 10 August 2026 ERCOT filed with the Commission a request for three good-cause exceptions to Planning Guide Revision Request 145, along with a status update on related matters. The analysis here reflects that mid-August architecture as documented in ERCOT's August Open Meeting materials, before the September conditional classifications.
Batch Zero had only just begun
Batch Zero is ERCOT's one-time transitional study of the reliability impacts of large loads that seek to interconnect and meet eligibility criteria under PGRR 145. Large loads for this purpose are 75 MW or greater. Medium loads are at least 25 MW and less than 75 MW. The Public Utility Commission had approved the Batch Zero framework on 18 June 2026 after Board adoption of PGRR 145 and NPRR 1325 on 2 June. The approval-era calendar had pointed to final classification notices around 7 August 2026. The Governor's directive landed inside that classification window and changed the cadence.
The pause is therefore not a repeal of batch interconnection. It is an insertion of eligibility verification and community-impact review before final classification and before energisation of the data-centre and crypto tranche at or above 75 MW.
Three good-cause exception requests
ERCOT's 10 August filing sought three good-cause exceptions for PGRR 145.
First, delay of the Batch Zero classification deadline. ERCOT asked to be excused from the 7 August 2026 classification deadline in Planning Guide section 9.3.1(2)(a) and instead to conditionally classify large loads by 31 August 2026, so that conditional classification could initiate the dispute and reconciliation process in the Planning Guide while eligibility verification proceeds. ERCOT was explicit about what it was not asking: it did not yet seek to move the 9 April 2027 deadline to deliver Batch Zero Interconnection Study results, because it did not yet know how the delay would affect the study timeline. Downstream, ERCOT warned that final classification delay would delay the Long-Term Load Forecast, which in turn could delay the Reliability Assessment and the December 2026 Capacity, Demand and Reserves Report. ERCOT recommended delaying Long-Term Load Forecast publication until final base-load classification is set, because incorporating verified Batch Zero base loads produces a more accurate forecast than the alternatives.
Second, inclusion of certain large loads in the Quarterly Stability Assessment before final classification. Classification as base load is a gate to inclusion in the Quarterly Stability Assessment. Because final classification was postponed, no large load could meet that requirement while verification was underway. Without relief, loads would drop from the 1 August and 1 November 2026 assessments and could not energise on projected dates solely because final classification was incomplete. ERCOT's August materials stated that about 6 large loads representing approximately 1,959 MW were expected in the 1 August Quarterly Stability Assessment with projected energisation in the first quarter of 2027, and that about 17 large loads representing approximately 6,874 MW were potentially eligible for the 1 November assessment with projected energisation in the second quarter of 2027. Inclusion in the assessment does not by itself authorise energisation, and it does not exempt a load from eventual final base-load classification.
Third, dynamic data review and cure period extension. ERCOT asked to notify large-load developers of deficiencies in dynamic data after the 7 August 2026 deadline and to give any developer notified late the same 24-day cure period the Planning Guide already provides. The operator reported that it had reviewed 290 dynamic models and that approximately 18 percent were acceptable on first review. The request does not change the substance of the review or the cure standard. It shifts when late notifications go out so developers are not treated inconsistently.
Energisation pause and what it does not cover
Separately from the three Planning Guide exceptions, ERCOT paused approvals to energise large-load data centres and crypto facilities until Batch Zero eligibility verification is complete. August materials quantify the near-term data-centre and crypto large-load impact through end of 2026 as 17 large loads totalling 6,608 MW of peak demand over five years. Those loads had finished ERCOT processes except for approval to energise.
The pause does not freeze every large load. ERCOT may still approve energisation for non-data-centre large loads. Medium loads are outside Batch Zero verification and continue to interconnect consistent with transmission and distribution service provider energisation processes. That distinction matters for boards reading headlines. The binding constraint is targeted at data-centre and crypto facilities at or above 75 MW, plus the Batch Zero study and classification clock for the wider large-load cohort.
Developer exception requests already in the door
On 23 July 2026 ERCOT had established, by Market Notice, a process for developers to seek a good-cause exception where they could not show strict compliance with an eligibility criterion, with interconnecting transmission or distribution providers to submit materials by 24 July 2026. By the August briefing, ERCOT had received 41 such requests on 23 July. It did not plan to advance exceptions to the qualifying-interconnection-study requirement (21 requests), exceptions to the dynamic-model deadline (2), submissions after 24 July or by the interconnecting large-load entity rather than the provider (3), or a full exception from the security requirement (1). It planned to consider as possible good-cause exceptions nine requests related to the financial security deadline and two related to the Form W Provisional Controllable Load Resource submission deadline. Any such exceptions were planned for submission no later than December 2026.
Two work streams to December
ERCOT's planned response to the 3 August directive splits into two distinct work streams.
Batch Zero Eligibility Verification will confirm that large loads conditionally included in Batch Zero that have yet to energise satisfy applicable PGRR 145 eligibility criteria. Information gathering runs through requests for information to interconnecting transmission and distribution service providers, with site verifications of certain eligible base loads that have upcoming requested energisation dates. Developers get notice of inconsistencies and a reasonable opportunity to cure. Large loads that do not pass verification, or that do not respond, will be disqualified from Batch Zero. ERCOT planned to file the Batch Zero Eligibility Verification Report with the Commission by 10 December 2026, ahead of a mid-December Open Meeting.
Community Impact Review will request information from data-centre and crypto facilities that have requested 25 MW or more and have yet to energise. That stream refreshes medium-load information (an April snapshot had 157 medium data-centre and crypto loads totalling approximately 8,766 MW) and sends community-impact requests for medium loads and for large loads with conditional Batch Zero classification. ERCOT planned to file a separate Community Impact Review Report by the same 10 December 2026 date.
Eligibility paths under PGRR 145 are not ornamental. Base-load and studied-load tracks carry different attestation, evidence, financial-security and site-verification requirements, including site control, long-lead equipment, notice to proceed, end-use customer confirmation, contractor contracts, contribution-in-aid-of-construction satisfaction and section 9.7 disclosures about substantially similar requests. The August pause makes those paper requirements a live gate before energisation for the data-centre and crypto tranche.
Engineering reading of the pause
Serial Large Load Interconnection Studies failed when each new campus invalidated the case ahead of it. Batch Zero was supposed to restore a consistent base case and a single transmission plan. The Governor's directive does not reject that physics lesson. It inserts a verification layer because eligibility claims, dynamic models and community impacts were not yet trusted enough to let classification and energisation run on the original August clock.
For counterparties, the near-term binding constraints are procedural. Watch the good-cause exception outcomes on classification timing, Quarterly Stability Assessment inclusion and dynamic-data cure. Watch whether a project is a data-centre or crypto facility at or above 75 MW, and therefore inside the energisation pause, or a non-data-centre large load that can still seek energisation. Watch RFI response quality, because non-response is a disqualification path. Watch the 10 December report dates. Until those reports land, speed-to-power for Texas data-centre and crypto loads at Batch Zero scale is an eligibility and audit problem, not an LMP problem.