10 Takeaways from the NERC Emerging Large Loads Technical Conference (February 2026)

By: Jiecheng "Jeff" Zhao

The NERC Emerging Large Loads Technical Conference made one thing clear: the grid is at a pivotal moment. The decisions made over the next 12–24 months on NERC Registration, NERC Standards development, interconnection reform, and modeling and performance requirements will shape the reliability of the North American bulk electric system for decades to come. The industry is aware of the urgency and necessity, accordingly stakeholders at the meeting. The question now is whether regulatory bodies and stakeholders can act with the speed and coordination that this moment in time demands.

The North American Electric Reliability Corporation (NERC) hosted its Emerging Large Load Technical Conference on February 24–25, 2026. The event brought together utilities, ISOs/RTOs, data center developers, and industry subject matter experts to address one of the most pressing challenges facing the North American grid: the rapid proliferation of large loads, driven primarily by artificial intelligence (AI) and data center growth, and the possible reliability implications that follow. The conference reiterated numerous critical considerations for large load interconnection that Elevate has been involved with and reported on with its industry partners.

Below are our top 10 takeaways from the event that define the industry’s current trajectory and near-term priorities.

1. NERC Launches Dual Workstreams for Large Load Registration and Standards

At the upcoming NERC Standards Committee meeting in March 2026, two parallel workstreams are expected to formally launch:

• Large Load Registration Effort

• Large Load Definitions and Standards Development

NERC has set an ambitious goal of completing registration, definition, and initial standards development efforts by the end of 2026—an accelerated timeline that underscores the urgency that the federal regulator and broader industry are treating the reliability risks posed by large load growth. The initiative will leverage NERC’s newly approved Modernization of Standards Processes and Procedures Task Force (MSPPTF) to rapidly mobilize subject matter experts.

All data centers are under consideration for registration under the Large Load Registration Effort. Final criteria for registration may encompass load type, size, and behavioral characteristics, with provisions for grandfathered sites. On size thresholds, analysis of over 400 NERC Level 2 Alert responses indicates that 50 MW and 75 MW are the most commonly used benchmarks—far more frequently than any other value (see figure below). For standards development, a results-based approach is expected, with BAL, PRC, MOD, and CIP standards among the most applicable.

2. A Level 3 NERC Alert on Large Loads Is Imminent

Both the registration and standards efforts will likely lead to mandatory requirements for large loads (and possibly a resurrection of the load serving entity). In the meantime, NERC is expected to issue a Level 3 Alert, its highest-tier notification, which mandates action by applicable entities, in the first week of May 2026. The alert will build on the Level 2 Alert issued in 2025 and is expected to provide further guidance on large load modeling and commissioning practices. This escalation highlights the risk profile that NERC is placing on this topic and emphasizes that further action beyond voluntary recommendations is needed.

3. Fault Ride-Through Is the Most Urgent Reliability Risk

Across utilities and ISOs/RTOs, fault ride-through (FRT) emerged as the single most urgent reliability concern. Recent real-world events illustrate the stakes:

• Ireland (May 2025): A single disturbance caused a 387 MW data center load drop—approximately 52% of total DC load.

• Spain/France: Emerging supply-demand imbalance risks linked to large load behavior such as ride-through capability and ramp rate.

• Northern Virginia (2024, 2025): Multiple events involving 500–1,700 MW load transfers, including a well-documented incident in which 1.5 GW of data center load was disconnected.

• ERCOT: ERCOT has reports on multiple events involving large loads reducing consumption or disconnecting during grid faults (see figure below (source)).

A critical finding: even co-located data centers at the same substation can exhibit dramatically different behavior during grid disturbances. This complicates grid planning and underscores why blanket application of Inverter-Based Resource (IBR) FRT requirements is not appropriate. NERC’s priority standard development effort is timely, as utilities such as ATC and Dominion Energy are already moving to enforce FRT requirements for large loads. ERCOT is scheduled to vote on NOGRR 282 (Large Electric Load ride-through requirements) in the April–May 2026 timeframe.

4. Existing Load Models Are Inadequate — Next-Generation EMT Models Need to Evolve

There is broad consensus that conventional phasor domain modeling tools fall short for representing AI and data center load behavior. The traditional Composite Load Model (CMLD) was never designed to represent data centers, and is insufficient for studying modern large load behavior. More advanced electromagnetic transient (EMT) models are often required to represent the details of large load behavior, including:

• Over/under voltage and frequency response

• Ramping rates and transient fault behavior

• Sub-synchronous oscillation (SSO) risks posed by AI variability

• Ramping behavior and its impacts on nearby synchronous turbine-generator fatigue and IBR-based controller interactions

Model verification and model validation during commissioning are essential, as is post-commissioning monitoring via digital fault recorders (DFRs) and phasor measurement units (PMUs). These capabilities are often available in modern digital relays but must be enabled and configured appropriately. One particularly challenging scenario involves co-located data centers coming online with no load information available at commissioning—a gap that calls for creative regulatory and technical solutions.

5. AI-Driven Load Variability Is Reshaping Balancing and Stability Paradigms

According to NERC’s Level 2 Alert response data, 70% of in-service large loads by end of 2027 are projected to be data centers and 42% of those will be AI-specific facilities. Unlike traditional data centers, which operate at roughly 90% load factor with relatively stable demand, AI data centers exhibit large, rapid power fluctuations during training cycles. These fluctuations have material risks for active power-frequency control, voltage control, balancing and ramping, inter-area oscillations, and other essential reliability services (ERS), which therefore demand new approaches to grid management and stability analysis.

6. Interconnection Reform Is No Longer Optional

The current interconnection landscape is poorly equipped to handle the volume and complexity of large load requests. NERC’s Level 2 Alert responses also revealed a stark picture: 86.8% of entities receiving large load requests lack clear performance requirements, and 91.6% have no formal commissioning procedures or practices (see figure below). Recommended reforms include:

• Defining operating envelopes early in the process, at the load Point of Interconnection (POI)

• Standardizing interconnection requirements across transmission providers, ISOs, or other regulatory entities

• Mandating modeling and data collection requirements, model quality checks, and site information collection for model verification

• Conducting post-commissioning model validation and documentation reviews

• Coordinating protection system settings and providing sufficient information transparency of the customer electrical characteristics and behavior

• Increasing transparency of analysis methodology and study results to allow developers to better understand system upgrades and cost allocation

• Using application fees and credit deposits to deter speculative project filings

7. Data Centers Are Transitioning from Passive Load to Quasi-Resource

One of the more consequential shifts discussed at the conference is that many data centers are no longer content to remain passive grid consumers. With co-located generation assets, including gas turbines, solar, and battery storage, facilities are developing operations and blackstart capabilities, and actively seeking to share reliability responsibility with utilities. QTS, one major data center developer, for example, cited 1.5 GW of co-located generation and expressed interest in a proactive reliability partnership. This evolution raises a fundamental policy question the industry must answer: what constitutes a “fair share” of reliability obligation for large load owners and operators?

8. Resource Adequacy Faces Mounting Pressure

Between 30 and 130 GW of generation capacity is expected to retire by 2030, with gas replacement posing the greatest challenge for both resource adequacy and maintaining sufficient inertial support. In addition, the residential rate impact of this transition is politically sensitive and cannot be ignored. Meanwhile, large loads are intensifying resource adequacy shortfalls and supply chain risks at precisely the wrong moment. Federal responses, including efforts to revive gas generation and expand nuclear capacity, reflect the urgency of the situation.

9. The Speed-vs.-Reliability and Affordability Tension Will Define the Next 3–5 Years

Transmission and generation expansion remain the primary bottlenecks to data center growth. Ideally, data centers would move from application to operation in as little as two years, in order to keep pace with GPU and TPU technology cycles. But speed cannot come at the expense of reliability and affordability for the typical residential ratepayer. This tension will be felt most in specific data center hot spots but is a national policy issue. Geographically, approximately 50% of near-term data center demand is concentrated in Virginia, Ohio, Illinois, Texas, and California. NPCC and SERC regions show the lowest near-term expectation for large load interconnection, according to the NERC Level 2 Alert response data.

10. Communication and Transparency Are Imperative

A clear cultural theme throughout the conference was that the current pace of large load growth cannot be managed through siloed, reactive processes. Informal early engagement between utilities and developers, transparent data sharing, and clearly defined performance requirements and study results were cited repeatedly as essential to building the trust and coordination a reliable grid will require. Communication is a critical reliability tool that must be required and leveraged for this all to come together effectively.

Reach out to the Elevate team of subject matter experts (info@elevate.energy) to talk further about how we can help you navigate this complex and rapidly evolving regulatory landscape around large loads. We have strong technical expertise in technical interconnection requirements development, NERC/FERC regulatory processes, modeling and studies, and other engineering aspects.