Decision questionDoes the July 24 Batch Zero TSP-to-ERCOT submission deadline apply to Colchis, and if it does, which submissions remain open?
Cited public facts
load mw
1,000 MW
“1 GW Fully Executed DCA with AEP”
"1 GW of capacity supported by a signed Direct Connect Agreement with AEP to Anticipated power consumption in 2028 pending ERCOT approval" (Q3 2025 business update); "Colchis 2028 Target Energization 1,000 MW Gross Capacity … subject to ERCOT batch study process" (Q1 2026 business update).
A new site: "In October 2025, the Company purchased 53% of the equity in Colchis LLC ('Colchis'), a joint venture of a potential 1 GW site in Texas, the 'Colchis Site.'" (ownership since increased to 76%).
Reconfirmed in the Q1 2026 business update: "Colchis 2028 Target Energization (1) 1,000 MW Gross Capacity (1) … (1) Gross capacity and target energization subject to ERCOT batch study process."
Colchis is a ~1 GW site in Texas with a fully executed Direct Connect Agreement with AEP, land purchase options of up to 620 acres, and 2028 target energization "pending ERCOT approval", per Cipher's own SEC-filed business updates.
Cipher's Q1 2026 business update reconfirms Colchis at 1,000 MW gross capacity with 2028 target energization, footnoted "Gross capacity and target energization subject to ERCOT batch study process", Cipher's own qualifier that the ERCOT process, not the target, controls.
Open unknowns, not established by any reviewed public source
Batch Zero participation, Cipher's filings put Colchis "subject to ERCOT batch study process," but no public source shows whether the July 10 ILLE package was filed or accepted; whether it went in is now a fixed fact only Cipher's interconnection owner can confirm.
Eligibility path (standard study, PCLR election, or WLPUN/BYOG), not public; the July 24 TSP package contents depend on it.
Interconnecting-TSP role, the Direct Connect Agreement counterparty is AEP, but the interconnecting TSP responsible for the July 24 TSP-to-ERCOT qualification package is not confirmed in any public source.
Completed submissions, no public record shows which Batch Zero submissions, if any, are already in; no submission status is asserted here.
Crusoe Stargate 1 Abilene
Crusoe · RT-02
Decision questionDo the NOGRR282/NPRR1308 Large Computational Load ride-through requirements apply to Stargate 1 Abilene, and which Crusoe or Lancium owner holds each required fact and record?
Cited public facts
load mw
1,200 MWtotal power capacity
“a total power capacity of 1.2 gigawatts (GW)”
Crusoe reports total power capacity. This is prospect-filter evidence, not a declaration of aggregate peak Demand at the single site.
Crusoe reports total design capacity in its project-award release. This is prospect-filter evidence, not a declaration of aggregate peak Demand at the single site.
Crusoe's July 2026 release says the Abilene partnership model has Lancium acquire and own the land, secure the power interconnect, and manage energy orchestration, while Crusoe designs, builds, and operates the AI data center.
Lancium's official locations page names Stargate 1 Abilene, states that its 1.2 GW interconnect is fully ERCOT-approved.
Lancium locations: Stargate 1 Abilene · verified · 2026-07-28VERIFIED AS A LANCIUM-PUBLISHED STATEMENT; NOT INDEPENDENTLY VERIFIED AGAINST AN ERCOT PROJECT RECORD.
05
Crusoe's March 2026 release distinguishes a separate adjacent 900 MW campus from its existing Abilene infrastructure and projects about 2.1 GW of total capacity across the full Abilene site.
Open unknowns, not established by any reviewed public source
Applicability, total power capacity and approved interconnect capacity do not establish RT-02 applicability without the rule's project-specific facts.
Evidence, no ride-through capability evidence, exemption record, or computational-share documentation has been reviewed.
Crusoe/Lancium responsibility boundary, the public division of labor does not allocate the RT-02 facts, filings, evidence, or compliance ownership between the companies.
Single-site scope, the separate adjacent 900 MW campus and projected 2.1 GW full Abilene site do not establish the RT-02 single-site boundary for Stargate 1 or its aggregate peak Demand.
Galaxy Helios
Galaxy Digital Inc. (NASDAQ: GLXY) · RT-02
Decision questionDo the NOGRR282/NPRR1308 Large Computational Load ride-through requirements apply to Helios, and which project facts and records resolve the screen?
Cited public facts
load mw
1,630 MWapproved power capacity
“the campus’s total approved power capacity has expanded to 1.63 gigawatts (“GW”)”
Galaxy reports total approved power capacity. This is prospect-filter evidence, not a declaration of aggregate peak Demand at the single site.
Galaxy says it completed ERCOT Large Load Interconnection Studies for Helios and secured approval for an additional 830 MW; it also identifies an AEP Texas electric-service agreement and WETT as the transmission interconnection provider.
Open unknowns, not established by any reviewed public source
Applicability, approved power capacity and completed LLIS work do not establish RT-02 applicability without the rule's project-specific facts.
Aggregate peak Demand, the public 1.63 GW approved-power figure is not a declaration of aggregate peak Demand at the single site.
Computational share of site Demand, not stated in the reviewed public sources.
Exemption or grandfathering, no reviewed public source confirms either.
Evidence, no ride-through capability evidence, exemption record, or computational-share documentation has been reviewed.
Rowan Temple
Rowan Digital Infrastructure · BZ-01, BZ-04
Decision questionWhich of the three Batch Zero dates bind Temple, the July 24 TSP-to-ERCOT submission, the conditional August 1 Q1-2027 QSA prerequisite gate, and the currently unknown ERCOT classification date, and what must be true for each to matter? August 7 remains the adopted deadline ERCOT says it will miss; August 20 is only a PUCT process checkpoint.
Cited public facts
load mw
300 MW
“construction is underway at its 300-megawatt (MW) Project Temple site”
Construction is underway at Rowan's 300 MW Project Temple site (roughly 700 acres, Temple, TX; Rowan's largest Texas project), with operations expected in 2027, stated year-only, and power secured through a partnership with Oncor. Rowan is a Quinbrook portfolio company.
Open unknowns, not established by any reviewed public source
Whether Temple is being submitted through Batch Zero at all, the July 10 ILLE filing window has closed, so whether the package went in is now a fixed fact only Rowan can confirm; public information does not show it.
Eligibility path (standard study, PCLR election, or WLPUN/BYOG), not public; the supporting materials due with the July 24 TSP package depend on it.
Whether Initial Energization in January to March 2027 is targeted , "operations 2027" is year-only, so the August 1 QSA prerequisite gate cannot be resolved from public information.
The interconnecting TSP's identity for §9 purposes, the Oncor power partnership is public, but the interconnecting-TSP role for the July 24 TSP-to-ERCOT package is Rowan's to confirm with Oncor.
ERCOT study status and completed submissions, not public; no submission or owner is asserted here.
Sabey SDC Austin
Sabey Data Centers · RT-02
Decision questionDo the NOGRR282/NPRR1308 Large Computational Load ride-through requirements apply to SDC Austin, and which project facts and records resolve the screen?
Cited public facts
load mw
84 MWaggregate power capacity
“84 MW of aggregate power delivered to the campus”
Sabey reports aggregate campus power capacity. This is prospect-filter evidence, not a declaration of expected aggregate peak Demand at the single site.
Open unknowns, not established by any reviewed public source
Applicability, published campus power and critical IT capacities do not establish RT-02 applicability without the rule's project-specific facts.
Aggregate peak Demand, the public 84 MW capacity figure is not a declaration of expected aggregate peak Demand at the single site.
Computational share of site Demand, not stated in the reviewed public source.
Exemption or grandfathering, no reviewed public source confirms either.
Project state and evidence, no study, agreement, ride-through capability, exemption, or computational-share record has been reviewed.
Satoshi Dragon Goby
Satoshi Energy · RT-02
Decision questionDo the NOGRR282/NPRR1308 Large Computational Load ride-through requirements apply to Dragon Goby, and which site-boundary, Demand, exemption, and evidence facts resolve the screen?
Cited public facts
load mw
475 MWproject card size label
“Size 475MW”
Satoshi Energy's project card does not define "Size." This is prospect-filter evidence, not a declaration of requested, contracted, studied, allocated, or aggregate peak Demand.
Satoshi Energy's Dragon Goby project card lists the market as ERCOT, status as "Development," timing as "COD Q4 2026," and size as "475MW." The COD label is not an exact Initial Energization date, milestone schedule, or ramp, and the size label's measurement basis is undefined.
Satoshi Energy describes development and marketing of construction-ready sites and offers colocated-development services generally. The Dragon Goby card does not state that this project is colocated.
Satoshi Energy Corp. filed comments in PUCT Project 58481 on April 16, 2026. Its 476 MW energized-with-renewables, 5 GW ERCOT pipeline, nearly 2 GW completed-study, and zero-CIAC statements are company-wide and do not name Dragon Goby, so none is attributed to this project.
Open unknowns, not established by any reviewed public source
Applicability, the undefined 475 MW size label does not establish RT-02 applicability without a location, single-site boundary, aggregate peak Demand, computational share, and exemption facts.
Project identity and request, the owner, operator, customer, and ILLE legal entities; location; site boundary; whether the request is new or expanded; request date; meaning of the 475 MW size label; requested peak Demand; contracted Demand; aggregate peak Demand; and computational MW or fraction are not public.
Interconnection path, the TSP, DSP, POI, substation, request or LLIS number, study phase, status, results, Intermediate Agreement, Interconnection Agreement, and whether Dragon Goby is within Satoshi's company-wide pipeline, completed-study, or zero-CIAC sets are not established by the reviewed sources.
Batch Zero, no public source establishes filing, classification, route, allocation, deficiency status, Forms W or X, Affiliate Attestation, PCLR status, WLPUN status, a prior dynamic-stability study, or a Section 9.2.1.4 eligibility route.
Demand and energization records, the exact meaning and milestones of the Q4 2026 COD label, exact Initial Energization, ramp, current LIF or LCP, site-control proof, duplicate-project and backup-generation disclosures, QSA, NTP, telemetry, and energization evidence have not been reviewed.
Capital, construction, and generation, fees, security, CIAC, long-lead commitments, colocated-generation status, export, and netting are not public project facts. The general colocated-development service does not establish colocation at Dragon Goby.
Ride-through, no NOGRR282 exemption or frequency- and voltage-ride-through evidence is public. No DWG, dynamic, steady-state, stability, or short-circuit evidence has been reviewed.
Skybox PowerCampus Dallas
Skybox Datacenters · RT-02
Decision questionDo the NOGRR282/NPRR1308 Large Computational Load ride-through requirements (PUCT-approved 2026-07-09, effective 2026-08-01) apply to PowerCampus Dallas, and which project facts resolve the screen?
Cited public facts
load mw
300 MWavailable power capacity
“PowerCampus® Dallas 300 MW … 115-Acre Site”
Skybox's own listing: "supported by a private, onsite substation with 300 MW of available power. Located in Oncor's deregulated power territory." Listed available power, not a metered peak-demand declaration, see the aggregation unknown below.
PUCT approved NPRR1308 on 2026-07-09; it takes effect 2026-08-01. NPRR1308 defines a Large Load as aggregate peak Demand >= 75 MW at a single site and a Large Computational Load as one whose computational load is >= 50% of site Demand.
Open unknowns, not established by any reviewed public source
Computational share of site demand, public information does not show whether the planned load mix meets the >= 50% Large Computational Load threshold (NPRR1308 definition).
Aggregation, the 300 MW figure is Skybox's listed available power, not a declaration of aggregate peak Demand at the single site; whether energized demand will meet the definition's >= 75 MW aggregate-peak-Demand test as stated is for Skybox to confirm.
Exemption / grandfathering, no public record shows whether a NOGRR282 ride-through exemption or grandfathering has been confirmed for this campus.
Evidence, no ride-through capability evidence, exemption record, or computational-share documentation has been reviewed; this is a public-information first pass.
STACK DFW02
STACK Infrastructure · RT-02
Decision questionDo the NOGRR282/NPRR1308 Large Computational Load ride-through requirements apply to DFW02, and which project facts and records resolve the screen?
Cited public facts
load mw
500 MWadvertised campus power capacity
“Secure up to 500MW at STACK's DFW02 campus”
STACK advertises available campus power capacity. This is prospect-filter evidence, not a declaration of expected aggregate peak Demand at the single site.
Open unknowns, not established by any reviewed public source
Applicability, advertised campus capacity does not establish RT-02 applicability without the rule's project-specific facts.
Aggregate peak Demand, the public 500 MW figure is not a declaration of expected aggregate peak Demand at the single site.
Computational share of site Demand, not stated in the reviewed public source.
Exemption or grandfathering, no reviewed public source confirms either.
Project state and evidence, no current phase, executed agreement, milestone, ride-through capability, exemption, or computational-share record has been reviewed.
Tract Caldwell Valley
Tract · RT-02
Decision questionDo the NOGRR282/NPRR1308 Large Computational Load ride-through requirements apply to Caldwell Valley, and which site-boundary, Demand, exemption, and evidence facts resolve the screen?
Cited public facts
load mw
4,000 MWproject page capacity label
“4000 MW Capacity”
Tract's project page labels Caldwell Valley with 4,000 MW of capacity. This is prospect-filter evidence, not a declaration of requested, contracted, studied, allocated, or aggregate peak Demand.
Tract labels Caldwell Valley "Owned & In Development," lists 2,973 acres, states site energization "in 2028," and describes multiple campus options on 345-kV transmission lines. The year-only energization label is not an exact Initial Energization date or a ramp schedule.
Tract says the full roughly 3,000-acre park can support more than 4 GW of data-center capacity; Tract, Bluebonnet Electric Cooperative, and LCRA are planning an onsite substation capable of scaling to 1.6 GW; and the first 250 MW of capacity is contracted for delivery in Q1 2028. These are distinct public measurement labels, not Demand declarations or proof of the parties' ERCOT interconnection roles.
The same announcement says Tract initiated plans for onsite generation that would become grid connected. It does not establish current generation, WLPUN status, export, netting, or an ERCOT interconnection.
Tract Holding Company I LLC filed comments in PUCT Project 58481 on April 13, 2026. Its 2.8 GW proposed, 6.2 GW ultimate, four-project, three-TSP, and advanced-LLIS statements are company-wide and do not name Caldwell Valley, so none is attributed to this project.
Open unknowns, not established by any reviewed public source
Applicability, capacity, substation-scale, acreage, and delivery labels do not establish RT-02 applicability without a defined single-site boundary, aggregate peak Demand, computational share, and exemption facts.
Project identity and request, the owner or ILLE legal entity, whether the request is new or expanded, its request date, requested peak Demand, contracted Demand, aggregate peak Demand, and the campus division of the public 4 GW capacity label are not public.
Interconnection path, the TSP, DSP, POI, request or LLIS number, study phase, status, results, Intermediate Agreement, Interconnection Agreement, and whether Caldwell Valley is among Tract's four company-wide LLIS projects are not established by the reviewed sources. Bluebonnet's and LCRA's ERCOT regulatory roles cannot be inferred from the announcement.
Batch Zero, no public source establishes filing, classification, route, allocation, deficiency status, Forms W or X, Affiliate Attestation, PCLR or WLPUN status, a prior dynamic-stability study, or a Section 9.2.1.4 eligibility route.
Demand and energization records, exact Initial Energization, ramp, current LIF or LCP, site-control proof, duplicate-project and backup-generation disclosures, QSA, NTP, telemetry, and energization evidence have not been reviewed.
Capital and construction, fees, security, CIAC, long-lead commitments, the legal nature and counterparties of the first 250 MW delivery label, and current onsite-generation, WLPUN, export, or netting status are not public project facts.
Ride-through, the computational MW or fraction, NOGRR282 exemption, and frequency- and voltage-ride-through evidence are not public. No DWG, dynamic, steady-state, stability, or short-circuit evidence has been reviewed.