NERC Issues Alert on Data Centers Threatening Grid Stability

The North American Electric Reliability Corporation (NERC) has issued a rare Level 3 Essential Action Alert—its highest-urgency notification—warning that the explosive growth of data centers, particularly those powering AI training, cryptocurrency mining, and traditional computing loads, is introducing new risks to the bulk power system.

Issued on May 4, 2026, the alert highlights “customer-initiated large load reductions” in which more than 1,000 MW of computational load can drop offline in mere seconds during minor grid disturbances, such as voltage fluctuations. These rapid swings leave operators with “little or no room for real-time responses,” potentially triggering frequency instability or even cascading outages.

NERC, which oversees grid reliability across the United States, Canada, and parts of Mexico, noted that grid operators currently lack sufficient processes, procedures, or methods to model and manage these highly variable “computational loads.” The alert follows two prior warnings in the last nine months and is accompanied by a new Reliability Guideline on Risk Mitigation for Emerging Large Loads.

Registered entities—including transmission planners, operators, balancing authorities, and reliability coordinators—must implement seven essential actions covering detailed modeling (using tools like the PERC1 dynamic model), enhanced interconnection studies, rigorous commissioning tests, protection coordination, high-speed instrumentation, and improved real-time operations and communication with data center operators. Responses and risk-mitigation plans are due by August 3, 2026.

The surge in data center demand is driven by the AI boom. NERC and other analysts project data centers could account for the majority of new electricity demand growth over the next decade, with some forecasts showing U.S. data center consumption potentially reaching 6–12% of total national electricity use by the late 2020s.

Texas/ERCOT Takes a Proactive, Market-Driven Approach

While the NERC alert applies nationwide, Texas and its independent grid operator, the Electric Reliability Council of Texas (ERCOT), are handling the data center boom differently—and, in many respects, more aggressively—than other regions.

ERCOT operates an isolated, energy-only wholesale market under single-state oversight by the Public Utility Commission of Texas (PUCT). This structure gives it greater flexibility to adapt quickly compared with multi-state regional transmission organizations (RTOs) like PJM. Since 2022, ERCOT has run an interim Large Load Interconnection process for facilities of 75 MW or greater, treating them with many of the same scrutiny levels as new generators. It now uses batch studies to manage the massive queue—currently tracking roughly 410 GW of proposed large loads, about 87% of which are data centers.

In 2025, Texas lawmakers passed Senate Bill 6 (SB 6), which established a comprehensive framework for large-load planning, interconnection, cost allocation, and emergency curtailment. New rules require financial assurances to prevent stranded transmission costs, site-control proof, detailed forecasting, and the ability for large loads to be remotely curtailed or run on backup generation during grid emergencies. ERCOT is also formalizing dynamic modeling requirements (including PSCAD/EMT models) and exploring “controllable load resource” options that reward flexibility.

The result? Texas positions itself as a leader in safely integrating hyperscale data centers while protecting reliability. Its deregulated market and abundant land also encourage co-location with new generation (gas, renewables, or behind-the-meter resources), helping match supply and demand more nimbly than in constrained Eastern markets.

Challenges in Other Parts of the U.S.In contrast, the Eastern Interconnection—particularly PJM Interconnection, which serves the nation’s densest data-center corridor in Northern Virginia—faces acute strain. PJM has seen data center load double in recent years and has experienced multiple events in which dozens of facilities simultaneously tripped offline, shedding 1,000–1,500 MW almost instantly. Long interconnection queues (averaging 8+ years in some cases), transmission congestion, and skyrocketing capacity prices have forced PJM to pause certain generator queues and issue urgent white papers calling for fundamental market redesign.

Cost-allocation disputes are heating up: ratepayers in states like Maryland have pushed back against billions in transmission upgrades primarily benefiting Virginia data centers. Similar pressures are emerging in MISO and other regions, where rapid load growth is colliding with slower permitting and multi-state coordination hurdles.

California’s CAISO and the Western Interconnection are also seeing data center interest but face their own constraints around renewable integration and transmission bottlenecks. Nationwide, NERC’s alert underscores that without standardized modeling, better ride-through capabilities, and real-time coordination, even well-planned grids could face reliability events as computational loads continue to scale.

What Comes Next

NERC’s Level 3 alert and accompanying guideline signal a turning point: data centers are no longer treated as traditional “firm” industrial loads. They must now actively participate in grid stability through improved modeling, protection settings, telemetry, and flexibility. Utilities and hyperscalers alike are being asked to collaborate earlier and more transparently.Texas/ERCOT’s forward-leaning policies may serve as a national model, demonstrating that with the right regulatory and market incentives, the AI-driven energy boom can be managed without compromising reliability. Other regions will need to accelerate similar reforms—updating interconnection rules, revising planning assumptions, and potentially rethinking cost allocation—if the U.S. is to power the digital future without risking the lights going out.