U.S. utilities are reporting a surge in interconnection requests from proposed data centers, many tied to artificial intelligence. The scale is unprecedented—often surpassing existing demand in entire service territories. Oncor in Texas received 552 large-customer requests by mid-2025, up 30% in a single quarter, with proposed data centers seeking nearly six times its current peak demand.
American Electric Power faces pending requests equal to five times its current system size. Utilities call them “phantom data centers”: many will never be built, but all are forcing utilities to rethink grid expansion.
Why It Matters
Even if only part of this demand materializes, it would mark the sharpest increase in U.S. electricity use in decades. Utilities are weighing how much new transmission and generation to add, knowing that overbuilding could saddle customers with costs if AI demand slows. While Texas and Northern Virginia remain the most visible examples, queues are expanding nationwide, meaning anyone tracking U.S. power markets should pay attention.
Meeting this demand will hinge on new generation. Natural gas pipelines and renewable projects are positioned to play a central role in filling the gap, making proximity to fuel supply (gas infrastructure) and clean energy resources (wind, solar, storage) as important as grid expansion. The data-center surge is not just about power consumption—it is reshaping how and where utilities plan future energy infrastructure.
What the Map Shows
In Rextag DataLink, the map highlights the overlap between Data Centers (filtered and colored by Total Power MW) and the Fiber Optic Network. Data centers represent the projected load growth, while fiber routes indicate the digital connectivity backbones that hyperscale operators depend on to run AI and cloud services.
Together, these layers show why regions like Dallas–Fort Worth, Houston, and Northern Virginia are under pressure: they combine both the demand for electricity and the connectivity backbone that makes large-scale data centers viable. This pairing explains not just where power will be needed, but why operators are choosing these locations.
A Deeper Dive with DataLink
● Filter Data Centers by Total Power (MW) to compare the scale of requests across metro areas.
● Use the Fiber Optic Network overlay to trace where connectivity supports proposed sites.
● Overlay Power Plants (All Types), Natural Gas Pipelines, or Renewable Energy projects (wind, solar, storage) to understand which regions are best positioned to supply this demand.
● Zoom into regions such as Dallas–Fort Worth, Houston, and Northern Virginia, then compare against other fast-growing metros across the U.S.
By exploring these layers together, users can see why certain regions are attracting speculative interest, how infrastructure overlaps with demand, and where expansion of both power and connectivity may accelerate first.
