PJM’s largest-ever data center load loss – nearly 4 GW vanishing in seconds on July 22 – has pushed the grid operator to draft mandatory ride-through standards for computational loads, a move that could reshape how hyperscalers connect to the bulk power system across the Mid-Atlantic and set precedent for ERCOT and CAISO.
Why a Single Afternoon Forced PJM’s Hand
The July 22 event was not an isolated anomaly. It was the third major Virginia data center trip in roughly 18 months, but the first to approach 4 GW – roughly the output of four large nuclear units – and the first to cut PJM’s total system load by nearly 4 percent in a single instant. Previous events in 2024 and early 2025 each shed about 1.5 GW. The pattern suggests a systemic vulnerability: when a transmission fault or voltage sag hits the dense Data Center Alley corridor along the Dulles corridor, dozens of facilities simultaneously disconnect to protect their own equipment, amplifying the disturbance across the interconnection.
PJM operators restored balance within nine minutes, comfortably inside the NERC BAL-002 30-minute recovery standard. But speed of recovery masks the risk. A 4 GW sudden loss forces spinning reserves to deploy instantly, stresses generator governors, and can trigger under-frequency load shedding if multiple contingencies overlap. The fact that three such events have now occurred in a region where data center demand is projected to grow by another 10-15 GW by 2030 – based on announced campus expansions – means the next event could coincide with tighter reserve margins or concurrent generator outages.
Virginia’s concentration is unique. Loudoun, Prince William, and Fairfax counties host the world’s highest density of hyperscale facilities, served by a transmission network originally designed for far lower, more predictable industrial loads. The July trip likely originated from a single transmission fault that cascaded into coordinated protective relaying across multiple campuses, each programmed to disconnect on voltage deviation rather than ride through. That behavior is legal under current interconnection agreements but creates a collective action problem: individual rationality (protecting servers) produces systemic fragility.
Ride-Through Standards Would Redefine Hyperscale Interconnection
PJM’s proposed ride-through requirement would compel new and potentially existing data center loads to remain connected during defined voltage and frequency excursions – similar to the low-voltage ride-through (LVRT) and frequency ride-through mandates that have applied to wind and solar plants since FERC Order 661-A and its successors. For a 100 MW campus, that means investing in uninterruptible power supply (UPS) capacity, on-site generation, or advanced power electronics capable of sustaining critical loads for the 0.15-0.5 second ride-through window typical of transmission faults. Retrofit costs for existing campuses could run $5-15 million per 100 MW depending on UPS architecture and whether backup generators are already permitted for grid-support mode.
The precedent is instructive. When ERCOT mandated LVRT for renewables after the 2011 and 2021 cold-weather events, compliance costs were absorbed into project economics within two procurement cycles. But data centers differ: their load profile is flat, their power quality requirements are stricter (IT equipment tolerates milliseconds of deviation), and their contractual leverage with utilities is greater. PJM’s stakeholder process will likely debate whether to apply the standard retrospectively – a move that would face legal challenge from hyperscalers arguing regulatory taking – or only to new interconnection queue entries, which would leave the existing 4 GW vulnerability unaddressed for years.
ERCOT and CAISO are watching closely. ERCOT’s 2023 “Large Load Integration Study” flagged similar risks from crypto-mining and hydrogen electrolyzer clusters in West Texas, while CAISO’s 2024 “Load Flexibility” proceeding examines whether data centers in the Bay Area and Central Valley should provide grid services in exchange for interconnection priority. A PJM rule finalized in 2026 could become the de facto national template, especially if FERC issues a policy statement endorsing ride-through for large computational loads under Section 206 of the Federal Power Act.
Who This Affects
- Utility transmission planners: Must model 4 GW+ contingent load loss as a credible N-1-1 scenario in 2025-2026 RTEP studies, likely requiring new dynamic reactive support (STATCOMs or synchronous condensers) at 500 kV substations serving Data Center Alley – capital costs on the order of $150-300 million per installation.
- Hyperscale developers (AWS, Microsoft, Google, Meta): Face a binary choice: negotiate customized ride-through exemptions with PJM and Dominion Energy, or redesign campus electrical architecture to meet a 0.15-second voltage sag ride-through – adding 6-12 months to permitting for campuses in the interconnection queue.
- Storage and hybrid developers: Gain a new revenue stack if PJM allows data center UPS systems to participate as demand response or frequency regulation resources; a 100 MW campus with 15 minutes of battery could earn $2-4 million annually in ancillary services at current PJM capacity and regulation prices.
- State utility commissions (Virginia SCC, Maryland PSC): Will adjudicate cost allocation for transmission upgrades driven by ride-through compliance – ratepayer advocates will argue hyperscalers should bear 100 percent of interconnection-specific reinforcements under existing tariff provisions.
What to Watch Next
- PJM’s Markets & Reliability Committee (MRC) vote on the ride-through problem statement – expected September 2025 – which will define scope (new vs. existing loads, voltage vs. frequency ride-through, measurement methodology).
- Dominion Energy’s 2025 Integrated Resource Plan filing, likely October 2025, for revised load forecasts incorporating ride-through compliance costs and any new transmission projects triggered by the July 22 event.
- FERC docket activity: watch for a Section 206 complaint or sua sponte investigation if PJM’s stakeholder process stalls past Q1 2026; a FERC order would accelerate implementation by 12-18 months.
- ERCOT’s “Large Load Ride-Through” working group deliverables due December 2025 – convergence with PJM’s approach would signal emerging national standard.
Bottom Line
The July 22 trip proved that voluntary reliability coordination among hyperscalers has failed; mandatory ride-through is now the only path to prevent a 4 GW load loss from becoming a 10 GW cascading outage as Data Center Alley doubles in size this decade.
Read the full report at Energy Central
Note: facts and figures attributed above to reflect that outlet's original reporting. Broader context, cross-sector connections, and forward-looking scenarios reflect independent analysis by our editorial team.
About this article: Drafted by Energy Ai with AI-assisted research and writing based on public reporting, then reviewed under our editorial process before publication.
Leave a Reply