Batch Zero, PGRR145, and NPRR1325: What the New Large Load Interconnection Framework Means for Developers, TSPs, and the ERCOT Grid

Why This Matters Now

The Texas grid is in the middle of the largest load-growth event in its history. Hyperscale data centers, AI training campuses, crypto mining facilities, and electrified industrial sites have pushed the active large-load interconnection queue well beyond what the historical ERCOT planning machinery was ever designed to absorb in a single cycle. Forecasts now place the 2032 load picture in the range of roughly 290 GW of cumulative load interest** against approximately **309 GW of available generation capacity and that generation total only reaches parity if planners are willing to count projects that have not yet completed full interconnection studies, and in some cases, projects that are technically classified as inactive.

That tight margin is the backdrop for two of the most consequential revision requests the ERCOT stakeholder process has produced in years: **PGRR145** (a Planning Guide Revision Request governing how large loads are studied and interconnected) and its companion protocol revision, NPRR1325. Together, they establish what the industry is now calling "**Batch Zero**" — the first coordinated, simultaneous study cycle for the wave of large loads currently sitting at the grid's front door.

For developers preparing site applications, for Transmission Service Providers (TSPs) sizing their next round of upgrades, and for load-serving entities planning around the new rules, the details inside Batch Zero will shape capital decisions for the next decade. This article walks through the framework, the engineering rationale behind the most-debated design choices, and three representative case studies that illustrate how the rules play out in practice.

The Core Problem: Why a "Batch" Approach At All

Historically, large loads in ERCOT followed one of two interconnection paths. If a load's in-service date was more than two years out, it could ride on an existing Regional Planning Group (RPG) transmission project review and use those studies for FAC-002 compliance. If the in-service date was inside two years, the load had to go through the **Large Load Interconnection Study (LLIS) process which examined that specific load against the existing system.

That model worked when large loads arrived one or two at a time. It breaks down when hundreds of gigawatts of requests arrive within a 36-month window, each load study assuming a different set of base conditions, each new request potentially invalidating an earlier one's reliability conclusions.

The Batch Zero concept fixes this by **studying all eligible loads simultaneously against a shared 2032 system case**. Rather than a serial first-come-first-served process where each new load is studied against an ever-shifting backdrop, Batch Zero freezes the modeling assumptions, applies a defined eligibility cutoff, and produces a single coherent allocation that can be defended on reliability grounds.

The trade-off is significant. Batching introduces a hard cutoff date that determines who is in and who waits for Batch 1. It also forces the question of how to rank loads whose studies overlap or conflict — a question the framework answers by ranking studies according to **completion date**, regardless of whether the study was conducted through the RPG pathway or through LLIS.

The Eligibility Picture: Load and Generation in 2032

Understanding the framework requires understanding the numbers it is built around.

On the load side, the modeled 2032 picture includes a base forecast of approximately 113 GW (covering organic growth, electric vehicles, rooftop PV, medium TSP loads, and crypto), plus the studied large-load categories that survive the validity check. After accounting for various overlapping categories — Permian Basin Plan loads, RPG-tracked large loads, validated LLIS loads, and additional load identified through stakeholder comments — the total potential modeled 2032 load reaches roughly 220 GW within Batch Zero scope, with another ~70 GW of additional requests visible in the broader queue.

On the generation side, ERCOT classifies projects against the Planning Guide Section 6.9 framework:

• 6.9(1) Operational — units actually energized today
• 6.9(5)(a) — Planned, signed SGIA
• 6.9(5)(b) — Planned, completed Full Interconnection Study (FIS)
• 6.9(5)(c)(i) — Planned, FIS started with steady-state and stability studies complete
• 6.9(5)(c)(ii) — Planned, FIS started with steady-state complete
• 6.9(5)(c)(iii) — Planned, FIS started
• 6.9(5)(d) — Inactive status with completed FIS stability study

The further down that list a project sits, the lower the probability it actually delivers MWs to the grid by 2032. Serving the current Batch Zero scope already requires reaching into the 6.9(5)(c)(iii) and later categories — projects that may or may not survive financing, siting, and construction. Expanding scope further would force modelers to count capacity from the 6.9(5)(d) inactive bucket, which is engineering speak for "we are pretending this will happen so the math works."

This is the central tension in the framework. The scope is already aggressive. Any expansion materially raises the risk of producing study outputs that look reliable on paper but cannot be physically delivered.

Study Validity: The Quiet Earthquake

One of the less-discussed but most operationally consequential elements of the framework is the study validity check. Before any large load can be classified as "base load" in the Batch Zero model — meaning it is assumed to be served reliably under all conditions and other loads must be planned around it — its underlying interconnection study has to be confirmed as still valid in the current system topology.

Here is where it gets interesting. Because LLIS studies and RPG-pathway studies have different assumptions about which other loads are present, changing the eligibility rules for RPG-tracked loads (which the framework does) ripples through the LLIS portfolio. A given LLIS study completed in February 2026 might have assumed certain RPG loads were not yet in the base case; if the rules now make those RPG loads automatically valid as of an earlier date, the LLIS study's reliability conclusions must be reassessed.

In practice, this means that a substantial number of LLIS loads that were considered to have valid studies under earlier draft language now require:

1. A new validity check against the updated automatic-validity assumptions
2. Potentially a full restudy if the validity check fails
3. Reclassification from "base load" to "studied load" if the restudy reveals reliability issues

The framework handles this by ordering studies by completion date and then walking down the list, checking each one against the cumulative set of validated loads ahead of it. Loads whose studies pass the check enter Batch Zero as base load. Loads whose studies fail enter as studied load meaning their MW allocation will be determined inside Batch Zero rather than treated as a fixed input.

For developers, this distinction matters enormously. Base load status carries a presumption that the requested MWs will be served. Studied load status means the MWs are subject to the outcome of the steady-state and stability analysis, and could be reduced — sometimes significantly — based on what the system can actually carry.

Provisional Controllable Load Resources (PCLRs): The Headroom Mechanism

Once the base-and-studied-load distinction is settled, the framework still has to answer a question that has frustrated large-load developers for years: what happens when transmission capacity to serve full requested load won't exist for several years, but the load itself wants to ramp as early as possible?

The answer is the Provisional Controllable Load Resource (PCLR) — a new resource category that splits a large load's consumption into two components:

• Low Power Consumption (LPC) — the firm portion, defined as the maximum load that can be reliably served under steady-state conditions given today's transmission capability. This is hard-limit firm service.
• Flexible portion — the difference between requested peak load and LPC, served via dispatchable demand under ERCOT's Security-Constrained Economic Dispatch (SCED) following base points like any other dispatchable resource.

Crucially, the full requested load is studied in Batch Zero, not just the firm portion. Transmission upgrades are identified that, over time, transition the flexible MWs to firm service. The PCLR mechanism is, in effect, a way to monetize transmission headroom as it appears rather than waiting for the entire upgrade plan to complete before allowing any consumption above LPC.

How PCLR Registration and Operation Work

A large load entity wanting to operate as a PCLR must:



1. Declare intent before Batch Zero starts. This is not a back-end opt-in; it must be on the table at the front of the process.
2. Submit a complete Declaration and supporting information by the defined deadline (currently late July of the cycle year).
3. Be studied in Batch Zero at full requested load for steady-state purposes (to determine LPC) and at full consumption for stability analysis.
4. Accept the Form W Part B election within the post-study window. This is the binding decision point.
5. Register as a Resource Entity (RE) with ERCOT and designate a Qualified Scheduling Entity (QSE).
6. Establish CLR parameters in RIOO and meter the facility for CLR participation.
7. Set up ICCP telemetry to the Operations control room.
8. Receive Approval to Energize from ERCOT.

Critically, steps 6–8 must occur before any non-firm load can energize. They can run in parallel with the Qualified Scheduling Entity Agreement (QSA), but they cannot be skipped or deferred.

Form W Part B: The Binding Election

After Batch Zero produces its results, each Interconnecting Large Load Entity (ILLE) receives an allocated LPC value (firm) and a Maximum Power Consumption (MPC) value (total ceiling including flexible). To proceed, the ILLE must file Form W Part B selecting one of four paths:

Failure to submit Form W Part B by the deadline is treated as Option 4 — automatic removal from the Refinement Study and loss of queue position. This is the single most expensive missed deadline in the framework.

Dynamic Bid Capping and Transmission Constraints

In real-time operations, PCLRs must follow SCED base points like any other dispatchable resource. But there is a wrinkle. Because the load is participating to consume *energy*, its energy bid curve must be capped under certain transmission-constrained conditions to ensure it actually backs down rather than continuing to consume through a binding constraint.

The dynamic bid capping logic works as follows:

Step 1: After each SCED run, the system generates a list of transmission constraints whose shadow price reached at least **90% of the maximum shadow price** observed in that run. This identifies the constraints that were materially binding.

Step 2: Before the next SCED execution, that list is compared against the active constraints in the current run. Constraints that appear in both lists are flagged as "triggering" constraints for bid capping.

Step 3: For each PCLR whose **shift factor is more negative than −0.02** on a triggering constraint, the PCLR's energy bid curve is capped at an **Adjusted Bid Cap (ABC)** value before the SCED step executes. The cap is calculated as:

> **ABC = Step 1 System Lambda − max(MaxShadowPrice × ShiftFactor) − $0.01/MWh**

The mechanism ensures that when a PCLR's continued consumption is contributing to a binding constraint, the price at which it is willing to buy energy is mechanically pulled below the marginal system price plus its contribution to the constraint, plus a $0.01 tiebreaker. The net effect is that SCED will dispatch the PCLR down rather than allow the constraint to violate.

It is worth noting that **no standard market mitigation applies to energy bid curves under this mechanism**. The bid cap is the mitigation. The logic remains grey-boxed in the protocol language pending implementation of NPRR1188 and the corresponding system changes for bid capping, but the design intent is clear.

Bring Your Own Generation (BYOG): Co-Located Load and Generation

 Why BYOG Is Attractive

For the load developer, BYOG delivers timing certainty. The load can energize on the generator's schedule rather than waiting for backbone transmission upgrades that may take five to seven years from study to commissioning. Curtailment risk from grid constraints is also reduced because the facility is partly self-supplied.

For the generator, BYOG converts what would otherwise be a merchant exposure into a contracted demand from day one. That improves project bankability dramatically and gives generation developers a clear role in resolving the load interconnection bottleneck rather than competing for the same scarce transmission.

For the system, BYOG accelerates capacity build precisely in the locations where it is needed most, reduces congestion on the rest of the network, and improves overall resource adequacy by linking new demand to new supply.

The Three-Study Framework

The engineering challenge with BYOG is that a co-located load-and-generation facility looks different to different ERCOT study processes. A coordinated framework requires three study workstreams running in parallel but with distinct scopes:

1. Generation Interconnection Study (Planning Guide Section 5).

Evaluates the generator independently under standard GINR assumptions to determine export capability and interconnection requirements. Co-located load is included in this study when it materially affects system performance — particularly for stability, voltage, and operational impacts. Output: maximum injection limits, stability constraints, interconnection requirements.

2. Batch Study (Load Interconnection — Planning Guide Section 9 / PGRR145).

Evaluates the large load *as if it were grid-supplied only* — that is, with the co-located generation explicitly not counted. The purpose is to determine the maximum import (withdrawal) capability the load could draw from the ERCOT grid under steady-state conditions. Output: withdrawal limits and initial load allocation, subject to stability limits from the combined study.

3. Transmission Planning Study

Models load and generation together as a single integrated facility and assesses full system impact under standard reliability criteria including N-1 (single contingency) and G-1 (generator outage) conditions. Identifies required network upgrades and validates **Self-Limited Withdrawal exit (SLF-exit)** conditions — the configuration in which the BYOG facility no longer needs to self-limit because adequate transmission exists. Output: required transmission upgrades and validation of long-term post-BYOG operating conditions.

Final Allowable Load

Final allowable load = min(Requested Load, Stability Limit)

Example: 1,000 MW requested, 600 MW on-site → 500 MW final (100 MW grid + 400 MW on-site)

Operational Phases

Interim phase: self-limited withdrawal

Post-transmission: standard ERCOT operation

Sequential BYOG Timeline Table

Studying Forward to 2032

Approach 1: Rollover

• Years 2027–2031 studied, 2032 deferred
• Effort: 100%, Timeline: 6 months

Approach 2: Bookend

• Fully study 2028, 2030, 2032, interpolate 2029, 2031
• Effort: 140%, Timeline: 8.5 months

Approach 3: Study All

• Fully study 2028–2032
• Effort: 200%, Timeline: 12 months

Paragraph:

2032 dominates due to location effects, queue depth, cross-zonal interactions, and limited parallelization.

Adjusted Timelines

Case Study

Case Study 1: Hyperscale Data Center Campus in the ERCOT North Zone

Facility profile

A hyperscale data center developer plans a multi-phase campus with a target peak demand of **800 MW by 2032**, ramping from initial energization at 200 MW in 2027, to 400 MW by 2029, 600 MW by 2031, and full 800 MW by 2032. The campus is located in the ERCOT North zone in an area with moderate existing transmission capacity but no committed backbone upgrades within the developer's required timeline.

Initial study outcomes

The LLIS was completed in February 2026 and passed initial validity review. However, after the study validity check against subsequent RPG-tracked loads that became automatically valid under the new framework, the original LLIS conclusions required reassessment. The reassessment determined that the full 800 MW could not be reliably served as firm load under 2032 conditions without significant transmission upgrades. The Batch Zero study determined a firm allocation (LPC) of **475 MW** and a stability-validated maximum (MPC) of **800 MW**, leaving 325 MW of flexible capacity dependent on PCLR participation.

Developer decision

The developer evaluated the three Form W Part B options seriously. Option 1 (accept PCLR as awarded) preserved the path to 800 MW but introduced exposure to SCED dispatch and dynamic bid capping for the flexible 325 MW. Option 3 (withdraw PCLR, retain firm) would have capped the campus at 475 MW permanently — sufficient for early phases but insufficient for the contracted AI training workload by 2031. The developer selected **Option 1** and registered as a PCLR.

Operational results

During 2027–2028, the campus operated comfortably within its 475 MW LPC for early-phase workloads. As the load ramped past 475 MW in 2029, the PCLR mechanism began actively dispatching the flexible portion. SCED dispatched the campus to its full requested level approximately **84% of operating hours**. During the remaining 16%, dynamic bid capping was active — primarily during summer peak hours and during planned transmission outages affecting the North zone. The campus held a sufficient operational reserve in its workload scheduling to redirect non-time-sensitive AI training jobs to other regions during capping events, limiting the operational impact.

Transmission catch-up

Backbone transmission upgrades identified in the Batch Zero Refinement Study completed in stages between 2030 and 2032. By the end of 2032, the campus had been fully transitioned to firm 800 MW service.

Key engineering takeaways

• The PCLR mechanism delivered the developer's strategic objective (path to full 800 MW) at the cost of accepting transmission-driven operational variability for the flexible portion.
• The shift factor screening (≤ −0.02 threshold) worked as designed: bid capping events were concentrated on the specific North-zone constraints where the campus materially contributed to the binding flow, not on every system-wide constraint.
• Form W Part B was the critical decision point. The 60-day window was tight given internal stakeholder approvals; developers in similar situations should pre-align their decision criteria before Batch Zero results are issued.
• The post-validity-check reclassification from full firm to partial firm + PCLR was a significant change from the original LLIS assumption. Engineering teams should not assume that an early LLIS pass guarantees firm allocation in Batch Zero.
  

Case Study 2: Industrial Decarbonization Site with Co-Located Solar + Storage (BYOG)

Facility profile

A specialty chemicals producer plans to electrify a large industrial site in the ERCOT South zone, replacing existing gas-fired process heat with electrified equivalents. Peak demand at full electrification is **600 MW by 2030**. The site is located in an area where transmission capacity to serve 600 MW of new firm load is at least 4–5 years from completion, with the available pathway being a planned 345 kV line extension that has not yet entered the active RTP queue.

The developer chose to pursue a **BYOG configuration**: 600 MW of process load co-located with **450 MW of utility-scale solar PV and 200 MW / 800 MWh of battery storage**, sized to cover daytime production with battery shifting to support evening operations.

Three-study results

The Generation Interconnection Study evaluated the 450 MW solar and 200 MW storage independently and determined a maximum injection limit of 380 MW at the POI, constrained by local 138 kV network capacity. Standard stability requirements were validated.

The Batch Study evaluated the 600 MW load as grid-supplied only and determined a withdrawal limit of 180 MW under 2030 conditions, reflecting limited transmission headroom from the existing network until the 345 kV extension completes.

The Transmission Planning Study modeled the integrated facility (600 MW load + 450 MW solar + 200 MW storage) and assessed N-1 and G-1 conditions. Stability analysis identified a maximum stable total load of 620 MW, which was non-binding (above the requested 600 MW). The study confirmed the BYOG configuration is reliable under all standard contingencies including the loss of the largest single generation unit on site.

Final allowable load determination

• Requested load: 600 MW
• Stability limit: 620 MW (non-binding)
• Withdrawal limit from grid: 180 MW
• Maximum from on-site generation: 450 MW (solar peak) + 200 MW (battery discharge) — but limited by load + storage charging needs
  

Final allowable load

Operational results

The facility operated under the BYOG agreement with self-limited withdrawal from 2028 through 2031. During daylight hours with high solar output, the facility met its full 600 MW load primarily from on-site generation, with grid imports averaging 80–140 MW. During evening hours, battery discharge covered the post-sunset shoulder, transitioning to higher grid reliance overnight when import demand sometimes approached the 180 MW cap. The chemical production schedule was adjusted to shift the most energy-intensive process steps to daylight hours when on-site generation was abundant.

During three separate events in 2029–2030, on-site generation was unavailable (planned solar maintenance combined with battery state-of-charge depletion). The facility's self-limiting controls reduced load to 180 MW within the required response time (subseconds for fast curtailment, with operational adjustment to safe-state production within minutes). No grid reliability events resulted from BYOG operation.

The 345 kV line extension completed in late 2031. After validation of SLF-exit conditions, the BYOG agreement dissolved in Q2 2032 and the facility transitioned to standard ERCOT operation. The solar and storage continued to operate but as independent market resources rather than as co-located self-supply.

Key engineering takeaways

• BYOG enabled the chemical producer to electrify on a timeline (2028 operational) that would have been impossible under firm-load-only assumptions (~2033 earliest given transmission completion).
• The three-study framework worked exactly as intended: the Batch Study correctly identified the limited grid contribution, the Generation study established export limits, and the integrated Transmission Planning Study identified the upgrades needed for the post-BYOG configuration.
• Load scheduling against solar availability required substantial process control investment. Facilities considering BYOG should account for the operational technology investment, not just the electrical interconnection cost.
• Battery sizing (200 MW / 800 MWh = 4-hour duration) was sufficient to cover evening shoulders but not full overnight operation. Facilities with 24/7 high-utilization workloads (such as data centers) would require larger storage or different supply configurations.
  

Case Study 3: Crypto Mining Operation Failing the Study Validity Check

Facility profile

A crypto mining operator planned a **350 MW expansion** of an existing site in the ERCOT West zone, co-located with existing wind generation capacity from a third-party owner. The original LLIS was completed in October 2025 — before PGRR115 implementation in December 2025 — under the assumption that an adjacent RPG-tracked transmission project would be available by the in-service date, and that several other large loads in the area would not yet be operational.

Validity check outcome

Under the Batch Zero study validity ranking by study completion date, the operator's LLIS was checked against subsequently validated loads and against the updated RPG eligibility framework. Two issues emerged.

First, the original LLIS had assumed that two large data center loads in the same West zone

substation district were *not* yet in the base case. Under the new automatic validity rules, those data centers became validated before the mining operation's study completion date, meaning they should have been in the base case for the original study but were not.

Second, the adjacent RPG transmission project that the LLIS had relied on for FAC-002 compliance had its expected in-service date pushed by approximately 18 months due to permitting issues, which the original LLIS had not contemplated.

The validity check **failed**, and a restudy was triggered.

Restudy results

The restudy determined that with the two data centers in the base case and without the delayed RPG transmission upgrade available within the relevant timeframe, the West zone substation district could not reliably serve the full 350 MW of mining expansion as firm load. The restudy concluded

• Firm allocation (LPC) feasible: **120 MW**
• Maximum studied under stability conditions: **350 MW**
• Difference (flexible / contingent): **230 MW** subject to either PCLR or BYOG treatment

Developer decision pathway

Option A

Accept 120 MW firm allocation, withdraw remaining 230 MW from Batch Zero, re-enter with future expansion in a subsequent batch. Selected against — would have limited near-term hash rate growth.

Option B

Accept PCLR designation for the full 350 MW, with 120 MW firm and 230 MW flexible. Strong consideration — preserved path to full capacity but introduced exposure to dynamic bid capping during West zone congestion events. Hash rate operations are extremely price-sensitive, and the operator's economic model showed that frequent bid capping would substantially reduce profitability.

Option C

Pursue BYOG configuration by contracting with the adjacent wind generator to formalize a co-located arrangement. The wind asset (nominal capacity 280 MW, average capacity factor ~38%) could supply a meaningful portion of the mining load on an availability basis. The Batch Study withdrawal limit remained at 120 MW, but the BYOG framework allowed the operator to draw the remainder from the wind asset when available and curtail mining hash rate during low-wind periods.

Detailed Frequently Asked Questions

Eligibility and Scope

PCLR Operations

BYOG and Co-Located Facilities

Timeline and Process