US Grid Operators Struggle With AI Data Center Power Surge
US regulators and utilities are implementing audits and new tariffs to manage a surge of speculative AI data center power requests that threaten grid stability.
The United States is facing a critical bottleneck in electrical grid capacity as artificial intelligence data centers drive a massive surge in power requests. Research from Wood Mackenzie indicates that only about 28% of the ,1066 gigawatts requested for data centers are likely to materialize, with the remainder consisting of speculative phantom projects. This inflation of demand has overwhelmed utilities and grid operators, leading to increased bottlenecks and potential rate hikes for residential consumers.
In response, Greg Abbott, the Governor of Texas, ordered a mandatory audit of all data centers seeking grid access on August 3, effectively pausing approvals. The Electric Reliability Council of Texas is currently tracking 474 gigawatts of connection requests, with data centers accounting for 90% of that total. Other utilities, such as Exelon Corp., have slashed their demand pipelines to prioritize credible projects over speculative ones.
To prevent households from subsidizing corporate infrastructure, states including Virginia and Georgia have implemented special large-load tariffs requiring data centers to pay for dedicated grid upgrades, a move Colorado is now considering through Xcel Energy. Simultaneously, federal regulators, acting on a request from U.S. Energy Secretary Chris Wright, have ordered grid operators to accelerate connections for large power users to maintain a strategic AI lead over China. This federal push faces local opposition over concerns regarding water usage, noise, and power draw.
