US Grid Operators Seek Deadline Relief as Data Center Loads Outrun Transmission Upgrades
Regional operators want more time to comply with FERC's transmission requirements, as data center demand forecasts outstrip what the grid can deliver.
US regional grid operators have asked federal regulators for an extension on a deadline to upgrade their transmission infrastructure, according to industry reports, a request that lands as data center developers confront mounting power availability constraints. FERC directed all six major regional operators outside of Texas to establish programs in late 2021. That work is now colliding with a wave of new electricity demand that the existing network was never built to carry.2
The timing is uncomfortable. Data center power demand could reach 9% to 17% of US electricity supply by 2030, or up to 790 terawatt-hours, compared with around 4% as of mid-May 2026, according to the Electric Power Research Institute. That forecast assumes the grid can actually deliver the electrons. Operators asking for more time to upgrade transmission suggest it cannot, at least not on the schedule developers have in mind.1
Some companies are already voting with their feet. Data centers are being told to generate power onsite rather than wait for grid connections, a shift Bloomberg reported and one that fits a pattern visible across US power markets. Developers report that grid upgrade costs, equipment availability, and permitting timelines have widened from a single bottleneck into a broader set of power-related constraints, according to a mid-year survey published by Utility Dive on August 31, 2026.7
Texas offers a live experiment. ERCOT, the state's main grid operator, was set to vote on Tuesday (2026-06-02) on rules requiring data centers to meet certain standards to avoid "cascading" outages — a concern US grid managers have raised about large loads tripping offline suddenly. Whatever ERCOT decided is now being watched closely by operators in other jurisdictions facing the same problem: how to integrate massive, volatile loads without destabilising the system.3
The storage industry sees an opening. The US added a record 57.6 GWh of new battery capacity in 2025, according to the Solar Energy Industries Association, bringing total deployed capacity to 166.1 GWh. SEIA projects annual deployments will reach 110 GWh by 2030, with a significant share driven by data center demand. Fluence CEO Julian Nebreda said the company is engaged in over 30 GWh of data center-related projects globally, with a meaningful portion in the US.1
But batteries face their own grid problem. Storage firms report lengthy interconnection queues and a supply chain heavily dependent on China, hurdles that limit how quickly they can scale to meet data center demand. Tesla saw $430 million in revenue last year from selling storage systems to Elon Musk's xAI, and Calibrant Energy has agreed to provide a 31 MW/62 MWh battery system at an Aligned data center campus in the Pacific Northwest, according to Reuters reporting. Those deals are real. Replicating them at the pace the demand forecast implies is a separate question.1
Off-grid gas has been floated as a fix. It will not be enough, according to analysis from Canary Media, which noted that BloombergNEF's forecast for data center demand is nearly double what the firm projected in December 2025 and more than the grid is expected to accommodate. Onsite generation solves the interconnection queue problem for individual facilities. It does not solve the aggregate supply problem if every developer pursues the same strategy.6
PJM Interconnection offered a preview of the strain. America's largest grid, serving 67 million people across 13 states and Washington, DC, saw demand surge to roughly 163 gigawatts during a heat wave in the week of June 29, 2026, pushing the system toward an all-time record. Data centers were cited as a contributing force. Power prices tripled on the grid during that period. That is what the collision of extreme weather and high-baseload demand looks like in practice.5
The problem is not uniquely American. Chinese grid operators have resisted plans to boost renewable power for AI data centers, concerned that peak demand at those facilities is difficult to forecast and could raise risks for power firms. The bottleneck has shifted from generation to the grid itself, a constraint that does not resolve simply by building more solar farms or gas plants.4
In the US, the FERC deadline extension request, if granted, delays the transmission upgrades that would relieve some of this pressure. It does not eliminate the need for them. Developers weighing site selection, power purchase agreements, and onsite generation investments are making decisions now based on assumptions about grid availability that may not hold.2
The next signal comes from ERCOT's rulemaking process and whether other regional operators follow Texas's approach to data center interconnection standards. FERC's response to the extension request will indicate how much flexibility regulators are willing to grant. Neither outcome changes the underlying arithmetic: demand is growing faster than the wires.3,2