A single fallen power line exposed a growing grid problem from AI data centers

BitcoinWorld

A single fallen power line exposed a growing grid problem from AI data centers

A single power line failure outside Washington, D.C., this week caused more than 3 gigawatts of data center load to vanish from the PJM grid in under a minute, triggering voltage spikes that made lights flicker across the region. The event, which took over 11 minutes to stabilize, has renewed concerns about the stability of the largest grid in the United States as the rapid expansion of AI data centers strains infrastructure designed for a different era.

What happened on the PJM grid

PJM Interconnection, which manages electricity for 67 million customers from New Jersey to Illinois, experienced a sudden disruption when a power line went down. Normally, the grid recovers in seconds. This time, the failure triggered a cascade: data centers in Northern Virginia — the world’s densest concentration of such facilities — detected the voltage fluctuation and simultaneously switched to backup power. In about 30 seconds, roughly 3.1 gigawatts of demand disappeared. The grid then saw an excess of 3.49 gigawatts before it began to rebalance, according to data collected by Ting Labs, a startup that monitors grid quality through residential IoT sensors.

While no blackout occurred, the event caused widespread light flickering and demonstrated how quickly concentrated data center loads can destabilize the grid. Ricardo de Azevedo, CTO of ON.Energy, described the incident as “the canary in the coal mine.” The disconnection was twice as large as a similar event in 2024, when 60 data centers simultaneously pulled 1.5 gigawatts from the same grid.

Why this matters for AI and energy infrastructure

Data centers currently account for about 6% of PJM’s load, according to Synapse Energy Economics. By 2040, that figure is projected to reach 24%. The grid requires near-perfect balance between supply and demand. When large loads disconnect in unison, voltage can sag or spike, triggering further protective disconnections. Ali Zain Banatwala, senior market models specialist at the Independent Electricity System Operator, told Bitcoin World that data centers need to “sequentially either disconnect or reconnect” rather than all acting at once. A more orderly process would allow grid operators to develop robust procedures in advance.

The problem is not limited to PJM. Grid operators across the country, including ERCOT in Texas, are beginning to require large loads like data centers to “ride through” disruptions rather than disconnecting. But the technology to enable this is still being deployed at scale.

How technology could fix the problem

Startups like ON.Energy are developing uninterruptible power supply systems for entire data center campuses, covering not just servers but also chillers and other equipment. The system hides the data center behind a bank of batteries and sophisticated power conversion equipment. From the grid’s perspective, the facility appears as a consistent, well-behaved load rather than a source of volatility. ON.Energy’s system can absorb power fluctuations by charging batteries during surges and dispatching power during dips, responding in milliseconds. The company is currently installing 3 gigawatts worth of its systems at four data center campuses, de Azevedo said.

This approach also allows data centers to ramp computing workloads up and down — including energy-intensive AI training — without bothering the grid. It represents a shift from building data centers that protect themselves at the grid’s expense to building them as cooperative grid participants.

Conclusion

The power line failure this week was a warning, not a catastrophe. But as AI data centers multiply and their energy demands grow, the margin for error shrinks. The simultaneous disconnection of 3.1 gigawatts of load was twice as large as a similar event just last year. Without changes to how data centers connect to and interact with the grid, such events will become more frequent and potentially more severe. The solutions exist — from sequential reconnection protocols to battery-backed ride-through systems — but they require coordination between grid operators, data center developers, and technology providers. The clock is ticking.

FAQs

Q1: What caused the lights to flicker across the Washington, D.C., region?
The flickering was caused by a voltage spike on the PJM grid after more than 3 gigawatts of data center load disconnected almost simultaneously when a power line failed. The sudden drop in demand left excess supply on the grid, causing voltage to surge.

Q2: How common are these data center disconnection events?
They are becoming more common. This week’s event was twice as large as a similar one in 2024, when 60 data centers disconnected 1.5 gigawatts. Experts say such events will increase as data centers account for a larger share of grid load.

Q3: What can be done to prevent future grid disruptions from data centers?
Solutions include requiring data centers to disconnect or reconnect sequentially rather than simultaneously, and deploying technologies like battery-backed uninterruptible power supplies that allow facilities to ride through grid fluctuations without disconnecting.

This post A single fallen power line exposed a growing grid problem from AI data centers first appeared on BitcoinWorld.