To address the urgent challenges posed by the rapid growth of large, power-dense loads, particularly artificial intelligence data centers and crypto mining operations, the University of Texas at Austin recently convened a “Large Load Symposium.”
The growth of artificial intelligence data centers in Texas has the potential to meaningfully boost local economies. It also presents challenges in energy demand and grid reliability, the University of Texas Energy Institute noted.
With the addition of a substantial number of data centers and other large loads due to steadily increasing electrification of industrial and manufacturing processes, ERCOT is anticipating a consequential rise in power demand in Texas by 2030.
This has the potential to strain the energy system in Texas without strategic planning, especially in rural regions where the grid infrastructure is not as resilient as it is in the urban population centers, the University of Texas Energy Institute noted.
At the Large Load Symposium, there were sixty-five representatives from the Public Utility Commission of Texas (PUCT), ERCOT, Texas Reliability Entity (TRE), industry—including transmission service providers, load developers, technology providers, equipment manufacturers, generators—as well as regulatory and academic experts.
The event featured presentations from TRE and ERCOT, followed by two structured breakout sessions.
Participants discussed growing concerns around ERCOT's ability to manage near- and medium-term large load growth and were asked to (1) define the scale and nature of the challenge including voltage stability, thermal constraints, and resource adequacy, (2) assess existing capabilities and risk mitigation technologies, and (3) identify actionable market and policy solutions.
The dialogue that emerged highlighted the need for updated interconnection processes, clear performance standards, and better coordination across load developers, utilities, and system operators.
Key takeaways included:
• Loads are coming online faster than grid processes can accommodate.
• Voltage and frequency ride-through (VRT) capabilities are the most urgent technical hurdle. Load developers need certainty that their interconnection timelines will not be affected by updates to models or mitigation efforts.
• Lack of standardization in interconnection agreements, study assumptions, and modeling guidance is a core barrier.
• Stakeholders support technical conferences to inform standards and clearer guidance from regulatory bodies.
• Greater data transparency, communication and coordination, and incentives for flexibility are needed.
Several actionable solutions were identified including, among other things:
• Develop a standardized Large Load Interconnection Agreement (LLIA): ERCOT, PUC, and transmission service providers (TSPs) should jointly define modeling expectations, timelines, data requirements, and pass/fail criteria.
• Create a system-wide study milestone dashboard or reporting platform: Enable TSPs, ERCOT, and loads to track study progress and avoid duplicated efforts. If ERCOT cannot develop this in a timely fashion, it could be outsourced.
• Define interconnection protection rules for data/model updates: ERCOT should clarify that providing updated models will not automatically trigger a restudy or jeopardize energization scales or timelines.
• Launch a technical working group (ERCOT, TSPs, loads, original equipment manufacturers (OEMs)): Form a standing body to co-develop standards, validate models, and resolve technology feasibility questions. Iterate to move quickly.
• Standardize load modeling inputs and formats: ERCOT should release templates for dynamic profiles, battery energy storage system (BESS) configurations, ramp rates, and mitigation options.
• Create a third-party modeling data center or repository: Allow loads to submit data to a neutral entity for model review and validation outside the ERCOT formal process, i.e. “third party review.”
A report related to the Symposium is available on the University of Texas Energy Institute's website.
