Author: Mareeswari Kadhiresapandian
ERCOT’s 2025 Long-Term Hourly Peak Demand and Energy Forecast (LTDEF) offers a comprehensive view of future system needs by integrating six key forecast components: Economic Base Load Forecast, Electric Vehicle Forecast (EV), Behind the Meter Rooftop Photovoltaic Forecast (PV), Large-Flexible Load Forecast (LFL), Large Load Contracts, and Large Load Officer Letters. The LTDEF uses the waterfall method to combine each forecast to create the ERCOT Net Forecast.
Waterfall Methodology – The method used to create the ERCOT net forecast is a waterfall approach that sequentially combines individual components of the forecast. As shown in the following formula, the LTDEF net forecast is the sum of base load, EV load, LFL forecasts, Contracted loads, and Officer Letter load ramps, less the Behind-the-Meter Rooftop PV forecast.
Net Forecast = Base Economic Forecast + EV Forecast + LFL Forecast + Adjusted Contracts + Adjusted Officer Letters – PV Forecast
Base Economic Load Forecast – This forecast is based on a set of econometric models describing the hourly load in the region as a function of the number of premises in various customer classes (e.g., residential, business, and industrial), weather variables (e.g., various temperature values), and calendar variables (e.g., day of week and holidays) to create the base load forecast.
ERCOT consists of eight distinct weathers. Each weather zone has two or three weather stations that provide data for the assigned weather zone. To reflect the unique weather and load characteristics of each zone, separate load forecasting models were developed for each of the weather zones.
- The key driver of the forecasted growth of demand and energy is the number of premises like Residential, business, and industrial premise counts, also on Housing Stock, Population and Non-Farm Employment.
- All premise models were developed using historical data from January 2018 through May 2024. An autoregressive model (AR1) was used for all premise forecasts.
Electric Vehicle Forecast – The current number of vehicles in each class and the relative rate of EV adoption is used to calculate the number of forecasted EVs.
- ZIP code level vehicle projections are converted to substation level projections based on non-coincident peak load of each substation.
- EV load profiles for each substation are generated by season/day.
Table 1 shows the forecasted maximum EV demand, by year.
Table 1: EV Peak Charging
Behind the Meter Rooftop Photovoltaic (PV) Forecast – The Behind the Meter Rooftop Photovoltaic Forecast was generated by customer class (Residential or Business) at the weather zone level.
- The modelling dataset contains historical weather and calendar data from January 2012 to July 2024.
- The assumed installed capacity in 2031 was 8,027 MWh and the forecasted solar maximum generation was 6,049 MW.
Table 2: Rooftop PV Scenarios
Large Flexible Forecast – Large Flexible Loads are categorized as:
- Co-located, meaning that the load is behind existing generation
- Non-co-located, meaning that the load is not behind existing generation
The Large Loads can generally be categorized in the following types of end uses:
- Hydrogen: Hydrogen in this context refers to hydrolysis plants that use electricity to turn water into hydrogen and oxygen.
- Data Centers: Data Centers are facilities designed for cloud storage and computing. They can also be designed for artificial intelligence training.
- Crypto: Crypto refers to certain data centers mines cryptocurrency that uses Graphics Processing Units to solve Blockchain equations.
- Oil and Gas: Oil and Gas refers to oil and natural gas exploration and recovery operations.
- Industrial: Industrial in this context refers to large manufacturing plants and other facilities that do not fall into one of the categories above.
There is approximately 3,700 MW of LFLs on the ERCOT system.
- TSP Provided Large Load Forecast: All Contracts and Officer Letter Large Load additions were based on the ramp schedules and MW size that the TSPs provided.
- ERCOT Adjusted Load Forecast: Assumes 180 Day Delay to ramp schedules of Contracts and Officer Letter Large Load additions with Data Center Large Load additions reduced to 49.8% then Officer Letter Large Load additions reduced to 55.4%.
Transmission System Providers-Provided Large Load Additions
- The proposed Large Load requests are broken down into two categories: Contracts and Officer Letters.
- Contracts are defined as prospective loads with a signed agreement from all parties and a financial commitment in place.
- Officer Letters are defined as a request without a signed agreement but the TSP attests to the viability of the request in a letter between an officer of the TSP and an officer of ERCOT.
- TSPs provided Contracts and Officer Letters via a Request for Information (RFI).
- ERCOT applied forecast adjustment factors to these projections based on recent actual observations of Large Load projects.
Figure 1: TSP’s Provided Large Load by Type
ERCOT’s Adjustment to Large Load Forecast – ERCOT has produced two forecasts based on all the TSP-provided large loads.
- The first includes the Large Load projections as ERCOT received from the TSPs. This forecast can be found on the ERCOT Load Forecasting website under “TSP Provided Large Load Forecast”.
- The second is an adjusted forecast based on observation of behavior and characteristics of these loads, including average project delay, load profile by type, and average project realization.
- The first adjustment was based on the average project delay of 180 days from the original project requested energization date for projects with in-service dates in 2022 through 2024.
- The next adjustment was applied to Data Centers. ERCOT studied requested MWs versus the peak consumption by Data Center site for Data Centers with in-service dates in 2022 through 2024. The average peak consumption per site was 49.8% of the requested MW. This factor was applied to all non-crypto Data Center Load additions.
Figure 2: ERCOT’s adjusted summer peak forecast
- The final adjustment used the percentage of previously filed Officer Letter projects with in-service dates in 2024 that have energized (55.4%). This percentage is based on percentage of loads energized, not a ramp rate or current MWs consumed.
As shown in Figure 3, historical annual energy for the calendar years 2014-2024 grew at an average annual growth rate (AAGR). of 3.1%. The forecasted AAGR for energy from 2025-2031 is 13.6%.
Figure 3: ERCOT Annual Energy Forecast
Other Forecast Adjustments
A portion of the load in the city of Lubbock was moved into the ERCOT Region in 2021, and the entire load was moved into ERCOT by the end of 2023. An hourly forecast was created for Lubbock based on Lubbock Power and Light’s (LP&L) peak forecast of its own growth. This separate forecast for Lubbock was added to the ERCOT forecast from LP&P’s projected integration date onward. LP&P’s forecasted load was added to the North weather zone.
Additional Rayburn Country Electric Cooperative (RCEC) load was included in the East weather zone. This load was initially added to the East weather zone in January 2020. A forecast was created based on data included from RCEC’s PUCT filing.
Hourly Demand Models
The long-term trend in hourly demand was modelled by estimating a relationship for each of the eight ERCOT weather zones between the dependent variable (hourly demand) and the following:
- Month
- Day of Week
- Hour
- Weather Variables
- Temperature including various lagged values
- Temperature squared including various lagged values
- Temperature cubed including various lagged values
- Interactions
- Day of Week and Temperature variables
- Hour and Day of Week
- Hour and Temperature variables
- Month and Temperature variables
- Number of premises
All the variables listed above are used to identify the best candidates for inclusion in the forecast models and to provide details on the types of variables that were evaluated in the creation of the models. Not every variable listed above was included in each model. Unique models were created for each weather zone to account for the different load characteristics of each zone.
ERCOT’s 2025 LTDEF provides a detailed and data-driven roadmap for understanding the region’s evolving power demand. By combining diverse forecast inputs and applying proven methodologies, ERCOT delivers a comprehensive view of future grid needs—supporting better planning, reliability, and long-term investment. As the grid faces increasing complexity from emerging technologies and large flexible loads, trusted insight is more critical than ever.
At ZEG, we specialize in supporting transmission developers and energy stakeholders with deep expertise in interconnection strategy, load forecasting, and grid modeling. If you’re looking to navigate the implications of LTDEF or optimize your projects within ERCOT, contact our team for tailored advisory services that move your development forward.
