Serial interconnection studies break when the next applicant changes the base case of the one ahead of it. That is the physical problem ERCOT faced as hyperscale sites, crypto plants and heavy industry piled into the same transmission corridors. By mid-2026 the operator was tracking more than 438,000 MW of large-load requests, nearly 89 percent of them data centres in the 18 June press release framing, and nearly 90 percent in the companion Trending Topics mid-2026 framing, against an all-time peak demand record of 85,508 MW set on 10 August 2023. The queue was not a forecast of what will energise. It was a signal that project-by-project Large Load Interconnection Studies could no longer keep a consistent picture of thermal limits, voltage support and contingent flows.
On 18 June 2026 the Public Utility Commission of Texas approved ERCOT's Batch Zero framework under Planning Guide Revision Request 145 and Nodal Protocol Revision Request 1325. The ERCOT Board had adopted the rules on 2 June. The design groups eligible sites of 75 MW and greater into one coordinated study so the operator can allocate transmission capacity against what the network can actually carry, then publish a single transmission plan. ERCOT describes itself as the first Independent System Operator in the United States to study and allocate grid capacity to new large users in batches rather than one project at a time. A large user at that 75 MW threshold is, in ERCOT's own framing, on the order of 18,750 homes at system peak.
Batch Zero as transitional architecture
Batch Zero is transitional, not permanent architecture. It replaces the legacy Large Load Interconnection Study path for the first cohort. Projects that already held qualifying studies can sit as base load, preserving prior capacity allocations where the rules allow. Others enter as studied load, competing for remaining headroom once the full batch is modelled together. Excluded projects wait for a later batch.
Two optional pathways sit beside the main track. Withdrawal-Limited Private Use Networks let a customer offset grid draw with on-site generation. Provisional Controllable Load Resources agree to curtail when local constraints bind, in exchange for earlier access before every upgrade is built. Both pathways treat flexibility and self-supply as transmission substitutes, which is the engineering point: megawatts that can leave the bus when the line heats up free capacity for firm load elsewhere.
The approval-era calendar expected required technical submissions by 10 July 2026, classification notices on 7 August 2026, spring 2027 allocation amounts, mid-2027 developer commitment, a Fall 2027 transmission plan for the Zero cohort, and Batch 1 applications in Summer 2027. ERCOT data at approval showed the majority of requests expect to be operational by 2030. Those dates no longer bind as written.
August pause, then September conditional classifications
September changed the cadence. After Governor Greg Abbott's 3 August 2026 directive for a comprehensive verification and audit of data centres moving through interconnection, ERCOT paused the Batch Zero study process, delayed final classifications, and stopped approvals to energise new data-centre and crypto facilities at or above 75 MW. On 10 August it filed good-cause exception requests for PGRR 145. The Commission granted those requests on 20 August 2026, according to the Board update.
On 3 September 2026 ERCOT provided provisional Batch Zero large-load project classification (base, studied, excluded) to the interconnecting transmission service providers. Final inclusion or exclusion now rides on the eligibility verification and audit, with results to be provided in December. Originally, PGRR 145 required ERCOT to provide final classification on 7 August 2026. Originally, PGRR 145 required ERCOT to complete the study by 9 April 2027. The study timeline impact remains to be determined.
The Board of Directors materials for the 14-15 September 2026 meeting state the conditional totals with precision. Some 204 projects (66.4 GW) met the qualifications to be conditionally included as base load. Some 158 projects (127.9 GW) met the qualifications to be conditionally included as studied load. That is 362 large loads in either base or studied load categories. A chart note in the same deck omits excluded load of 302.2 GW across 373 projects. Those four figures are the live September map of Batch Zero conditional eligibility. They are not energisation authorisations.
What conditional classification does and does not do
Conditional classification initiates dispute and reconciliation paths while verification proceeds. The September Board update notes that interconnecting transmission service providers received, for each of their large loads, information including loads reclassified from a base-load eligibility path to a studied-load path subject to verification; loads conditionally included subject to a future good-cause exception related to eligibility criteria (nine large loads); loads conditionally included subject to curing deficiencies in dynamic model data; and loads excluded from Batch Zero with the reason for exclusion.
Dispute routes differ by posture. A developer whose large load is excluded may dispute through ERCOT's Alternative Dispute Resolution process under Protocols Section 20. A developer whose large load was submitted as base load but reclassified as studied load may dispute using the defined timeline process in Planning Guide Section 9.3.1. Almost all not-yet-energised large loads remain subject to the PGRR 145 verification process.
Conditional base load and studied load are also not homogeneous products. Board charts break base load into eligibility paths under Planning Guide section 9.2.1.1(1), including loads energised before or after 25 March 2022, loads included in the May 2026 Quarterly Stability Assessment or Interconnection Voltage Ride-Through assessment or earlier, loads in the Permian reliability plan, advancing and committed large loads, and PURA section 39.169 net-metering load. Studied load charts distinguish firm load from Withdrawal-Limited Private Use Network (bring-your-own-generation) and Provisional Controllable Load Resource (flexible) slices. Regional and county concentration appears in the same deck, including top base-load counties such as Dallas and Harris. Those geographic details matter for transmission planning. They do not override the December verification gate.
December clocks: eligibility, community impact and RFIs
ERCOT told its Board that it will file a Batch Zero Eligibility Verification Report with the Commission by 10 December 2026. It will also file a Community Impact Review Report by the same date. Those are two separate work streams. Verification and audit affect Batch Zero interconnection eligibility. Community impact gathers information for policymakers on medium loads (25-75 MW) and on large-load data centres and crypto facilities, without itself being the interconnection screen.
Requests for information to interconnecting transmission and distribution service providers are the practical instrument. Completeness and timeliness of responses matter because loads that do not pass verification, or that do not respond, face disqualification from Batch Zero. Site verifications of certain eligible base loads with upcoming requested energisation dates sit beside paper RFIs. Dynamic-model deficiencies still require cure. Financial security and contribution-in-aid-of-construction confirmation by the interconnecting provider remain part of the eligibility package under PGRR 145.
The energisation pause on new large-load data centres and crypto facilities continues until verification and community-impact review complete. Non-data-centre large loads and medium loads sit under different rules. Boards that blur those categories will mis-read which projects can still move.
Queue physics beyond Batch Zero
The structural tension is not mysterious. ERCOT's 29 July 2026 briefing to the Texas Senate Committee on Business and Commerce put large-load requests near 474 GW as of June 2026, again about 90 percent data centres, while the generation interconnection queue held 1,949 active requests totalling 462,895 MW (about 463 GW), with solar and energy storage accounting for approximately 72 percent of generation seeking interconnection. Mid-2026 press materials used more than 438,000 MW for the large-load queue. The Senate deck's approximately 474 GW figure is a later June snapshot in a different briefing. Both are documented. Neither equals firm future load.
Load that wants firm service still arrives faster than wires and dispatchable capacity can be sited, permitted and commissioned. Batch Zero forces the queue to confront network physics before financial close: year-by-year megawatt allocations, security postings and a buildable upgrade plan rather than an infinite serial promise. Most applicants still target commercial operation by 2030, and Batch 1 applications remain expected to open in Summer 2027, with a Fall 2027 transmission plan for the Zero cohort if the study path holds after verification.
National context without leaving Texas jurisdiction
Other organised markets face the same mismatch of load speed versus build speed, but they sit under different jurisdictional tools. FERC's 18 June 2026 section 206 show-cause orders to six RTOs and ISOs address large-load tariff design on the interstate system. ERCOT is outside those six dockets. Its experiment still matters nationally because it rewrites the interconnection product itself: batch study, explicit allocation, optional curtailment and behind-the-meter offset, and now a governor-driven pause that inserts verification before energisation.
Until December's reports land, the live constraint for Texas large loads is not the next locational marginal price print. It is whether a project clears eligibility, stays in the conditional batch, cures model and security deficiencies, and waits for a transmission plan that matches physics to paper. Conditional base load at 66.4 GW and conditional studied load at 127.9 GW are large numbers. Excluded load at 302.2 GW is larger still. The sober reading is that ERCOT has published a conditional map, not a construction schedule. Verification, RFI discipline and the 10 December filings will decide how much of that map survives into the study case.