ERCOT Delays Interconnection Study Amid 474 GW Power Request
Texas grid operator halts Batch Zero transmission review as data centers flood the queue
The Electric Reliability Council of Texas (ERCOT) has delayed its landmark Batch Zero transmission planning study, halting the state's new process for connecting large-load projects. The suspension comes as state regulators face an unprecedented 474 gigawatts of power requests, the vast majority of which are driven by artificial intelligence and hyperscale data center projects.
Key details
In response to a directive from Texas Governor Greg Abbott, ERCOT has hit the brakes on its Batch Zero review process to audit all data centers in its interconnection queue. The state's queue has ballooned to approximately 474 gigawatts (GW) of requests, a figure that is more than five times the historic peak electricity demand ever recorded on the ERCOT grid.
According to state regulators, approximately 90 percent of these new power requests are for data center facilities, showcasing the massive scale of the AI infrastructure boom in Texas. Batch Zero was designed to be the inaugural transmission planning study under ERCOT's new large-load interconnection process, which was approved by the Public Utility Commission of Texas (PUCT) to handle loads of 75 megawatts or more. The delay forces data center developers and power generation companies to completely reassess their construction and operational timelines.
Why this matters
The suspension represents a major bottleneck for the expansion of physical AI infrastructure in the United States. Texas has been a premier destination for high-density AI data centers due to its independent grid, abundant land, and historically low energy costs. However, the sheer volume of power requests—nearly five times the grid's record peak—has forced regulators to prioritize ratepayer protection and grid security over rapid deployment.
Context
The move follows a series of state-level actions across the United States targeting the massive resource footprint of AI. Regulators in states like Florida, North Carolina, and New Jersey have recently enacted or proposed laws to prevent residential ratepayers from subsidizing the massive grid upgrades required by hyperscale facilities. In Texas, the sheer scale of the 474 GW queue has brought the tension between rapid AI scaling and electrical grid reliability to a critical head.
Risks and open questions
The primary risk is the potential for significant delays in the deployment of next-generation AI clusters and training facilities. Developers who were relying on the Batch Zero process to secure timely grid connections may now face years of waiting or be forced to seek expensive behind-the-meter generation options, such as natural gas turbines or on-site fuel cells. Additionally, it remains unclear how state auditors will distinguish between speculative "interconnection squatters" and viable projects with committed capital.
What happens next
State regulators and ERCOT officials will conduct a comprehensive audit of the active queue to filter out speculative power requests and establish stricter financial and operational requirements for developers. In the meantime, the suspension of the Batch Zero study is expected to freeze new large-scale data center approvals in the state, potentially prompting developers to shift investments to other regional markets or accelerate off-grid, self-powering solutions.
Source: Utility Dive Published on AI Usage Global, author: AUG Bot
