Texas Is Preparing To Cut Off Power To Data Centers During Grid Emergencies

By Georg Rute of UtilityDive

Texas is preparing to cut off power to data centers during grid emergencies — a sign of just how strained the system has become.

Over the Fourth of July, deadly floods swept across central Texas, disrupting infrastructure and causing widespread outages. Meanwhile, the Electric Reliability Council of Texas (ERCOT) has already seen multiple price spikes and conservation alerts — not because there wasn’t enough power, but because we couldn’t move it where it was needed.

These aren’t isolated events. It’s not just a Texas problem.

Just days after the shutoff planning was announced, the U.S. Department of Energy warned that blackout risks across the country could rise 100-fold by 2030.

All of this points to a deeper vulnerability: We’re still running the grid with tools and assumptions built for a different era — one with fewer storms, slower load growth and no massive data centers. 

Texas’s new normal demands smarter, faster and more adaptive grid operations. Long-term infrastructure investments are critical, but they won’t arrive in time to manage the next three summers.

Texas has made real progress in building new generation capacity, especially in solar, storage and wind. But the wires that carry that power haven’t changed. More importantly, the way we operate the grid hasn’t evolved to match the demands of either changing weather patterns or electrical load growth.

Texas generates far more power than the next 2 or 3 states COMBINED.

And we’re rapidly adding much MORE power.

Also our grid is more resilient than ever.

Since I signed power reform laws no Texan has lost power from the ERCOT grid.

We are best positioned for the future. https://t.co/kRopH5cGDk

— Greg Abbott (@GregAbbott_TX) December 14, 2024

Now, surging demand from industrial expansion, electrification and AI data centers is doubling the strain. ERCOT’s own projections show that power demand in Texas may nearly double by 2030. And, other regions aren’t immune.

• The Midcontinent Independent System Operator recently green-lit a $22 billion transmission buildout to relieve rising congestion.
• The California Independent System Operator saw renewable curtailments surge nearly 30% last year.
• The PJM Interconnection anticipates 3% to 4% annual peak load growth through 2035 driven by data centers and expects up to 70 GW of demand over the next 15 years.
• Nationally, U.S. demand is projected to climb about 16% in five years — a pace not seen since the 1980s. 

That means more stress on an already-congested transmission system — one still being managed with decades-old assumptions about heat, wind and demand.

Those assumptions no longer hold. And in a hotter, stormier Texas, they’re becoming dangerous.

The case for operational intelligence

Utilities around the world are taking a different approach — one that doesn’t require waiting 10 years to build new lines. 

In Europe, software-based tools like hardware-free dynamic line ratings, or DLR, and hyperlocal weather forecasting are safely increasing the amount of power that can flow through existing lines. These tools don’t involve new hardware or major infrastructure. They use data — from satellites, LiDAR scans, and thousands of weather stations — to help operators see where and when extra capacity is available and plan accordingly.

I’ve helped implement this approach with national grid operators overseas. In Estonia and Finland, for example, we applied AI-driven DLR across 7,000 miles of transmission lines — many in hilly, forested regions like much of America. The result: up to 40% more capacity on lines that, by traditional standards, were considered maxed out.

The same physics apply here. A mild breeze — just four miles per hour — can cool power lines enough to boost capacity by 30%. But grid operators typically don’t have access to sufficiently accurate weather forecasts. As a result they assume the worst-case weather at all times, just in case. That means we’re leaving megawatts stranded every day, even during critical hours and emergencies.

Load flexibility shouldn’t be the only emergency tool

Demand-side management is essential. But we shouldn’t have to shut down critical infrastructure just to survive a summer heat wave. If we can increase visibility into grid conditions, forecast congestion earlier, and bring in more power from further away, we can avoid triggering firm load shed in the first place.

Shutting off industrial loads like data centers should be a last resort — not the default backup plan.

This isn’t a call to stop building new lines or power generators. We need them. But they won’t arrive in time to handle the surging load coming in the next few years. 

What we can do now is operate smarter with software-based operational intelligence — to reduce curtailment, ease congestion and lower consumer costs.

It’s not political. It’s practical. And it’s proven.

ERCOT has long served as a proving ground for U.S. grid innovation. But today, it’s also the canary in the coal mine. What Texas does next will shape how the rest of the country prepares for what’s coming.