The demand for power on the Texas grid is projected to increase so significantly that it will require the equivalent of 30 new nuclear power plants by 2030 to satisfy this growth.
This estimate comes from the Electric Reliability Council of Texas (ERCOT), which oversees the state’s power grid. The anticipated surge in demand is largely attributed to the influx of new data centers driven by artificial intelligence, raising concerns about whether Texas’s infrastructure can keep pace with this rapid expansion and the associated costs. Emerging from the pandemic, electricity demand in Texas was already rising faster than in any other state. This trend is now being accelerated by the state’s ambition to become a leading hub for data centers, potentially on a national or even global scale. Some individual projects are requesting up to 1 gigawatt of power—sufficient to supply about 250,000 Texas homes—which introduces new challenges for maintaining grid stability, according to Agee Springer, ERCOT’s senior manager of grid interconnections.
Springer highlighted that the introduction of large industrial loads, such as data centers, poses reliability risks to the ERCOT system. “We’ve never been in a situation where such substantial industrial demands could impact grid reliability, and now we are entering that reality,” he stated during a panel discussion at Infocast’s ERCOT Market Summit in Austin. Risk of Grid Strain ERCOT has received requests totaling 99 gigawatts for new grid connections from major power consumers, including data centers, bitcoin miners, and hydrogen producers, as reported in an internal presentation. This is a significant increase from 40.8 gigawatts just last March. The state anticipates that peak power demand will increase by 75% by 2030, surpassing the current record of 85.5 gigawatts. This forecast has been adjusted to account for the new data centers, alongside the rising demand from an electrifying economy. ERCOT, along with lawmakers and regulators, is working to find ways to integrate data centers without overburdening the grid, which could lead to an increased risk of blackouts during extreme conditions. “There can’t be more demand than there is supply,” stressed Beth Garza, a senior fellow at the R Street Institute think tank.
Infrastructure challenges
A pressing concern is whether the necessary infrastructure can be constructed quickly enough, given current supply chain issues that lead to prolonged waits for essential equipment like turbines and transformers. An additional challenge is determining who will finance this expansion. Data centers remain profitable until electricity prices reach approximately $2,000 per megawatt-hour, while ERCOT’s price cap is set at $5,000, as pointed out by Resmi Surendran, vice president of regulatory policy at Shell Energy North America. It’s uncertain whether data centers will be willing to adapt, but Surendran noted that the ability to adjust power usage based on price signals—known as demand response—could help alleviate some of the grid’s challenges. However, the flexibility of data centers and other large electricity consumers brings another complication: it shifts the costs of transmission projects onto households and smaller businesses. This stems from a program initiated about two decades ago that allows Texas’s chemical plants, refineries, and even Bitcoin miners to significantly reduce or eliminate their contributions to grid upgrades by scaling back their power consumption during peak demand periods in the summer. Known as four coincident peaks (4CP), this approach was designed to mitigate stress on the grid during the hottest months when electricity is in high demand.
Texas Senator Charles Schwertner criticized this cost-sharing methodology, stating, “This is just not right,” as reliability risks have evolved. He noted that a current Texas Senate bill aims to reform this system. “We need to ensure that the methodology is fair to all ratepayers,” Schwertner asserted.