The race to secure power for AI infrastructure is increasingly reshaping how developers evaluate data center opportunities. While available land and fiber connectivity remain important, power availability has emerged as the primary constraint across many major U.S. markets. As a result, developers are looking beyond traditional data center hubs and toward sites where large-scale infrastructure already exists.
The recently announced Paducah American Energy Hub in western Kentucky illustrates this trend. A consortium including Brookfield, NextEra Energy, Big Rivers Electric Power Corporation, Jackson Purchase Energy Cooperative, and Paducah Power System plans to redevelop the U.S. Department of Energy's Paducah Site into a large-scale data center and energy campus. The project has been described as a $100 billion private investment and, if fully developed, would rank among the largest AI-oriented infrastructure projects announced in the United States.
The scale of the proposal is significant. Project plans call for up to 1.8 GW of utility capacity supporting more than 1.2 GW of compute capacity by 2032. To support those loads, developers have outlined plans for up to 4.6 GW of dedicated generation resources, including natural gas-fired generation and battery energy storage systems.
What makes the project particularly noteworthy is not simply its size, but its location.
Across North America, many proposed AI developments are competing for increasingly scarce transmission capacity. Long interconnection queues, transmission congestion, and limited power availability have become recurring challenges for hyperscale developers. The Paducah site enters the market with a different profile. As a former federal energy facility, it already possesses many of the characteristics that large-scale infrastructure projects typically spend years attempting to assemble, including transmission access, water availability, transportation infrastructure, fiber connectivity, and a substantial land footprint.
The project also reflects a growing convergence between the power sector and the data center industry.
Historically, data centers have largely depended on utility-led resource planning to satisfy future electricity demand. Today's AI-driven facilities are creating load requirements that increasingly resemble those of large industrial customers. In response, developers are exploring dedicated generation strategies that provide greater certainty around power availability and project timelines.
The Paducah development appears to fit squarely within that model. Under the proposed structure, Brookfield would develop and operate the data center campus, while NextEra Energy would build and own the dedicated energy resources intended to support it. The proposed generation portfolio includes up to 2 GW of natural gas capacity and up to 2.6 GW of battery storage, underscoring the scale of resources that may be required to support future AI workloads.
The utility framework behind the project is also notable. Big Rivers Electric Power Corporation is expected to provide wholesale electric service, while Jackson Purchase Energy Cooperative would provide retail delivery. This level of coordination is becoming increasingly important as data center projects move into the gigawatt range and require participation from multiple infrastructure stakeholders.
More broadly, the project highlights an emerging characteristic of successful AI development sites: they are often locations where multiple infrastructure systems already intersect. Power generation, transmission, natural gas supply, water resources, telecommunications networks, and available land are all becoming critical elements in site selection decisions. While any one of these assets may be insufficient on its own, their combination can significantly improve a site's attractiveness for large-scale development.
Whether Paducah ultimately reaches its full planned scale remains to be seen. Large infrastructure projects face numerous execution risks, including permitting, construction timelines, customer commitments, and evolving market conditions. However, the announcement offers an important signal about where the industry is heading.
As AI-related electricity demand continues to grow, developers are increasingly prioritizing locations capable of supporting not just data centers, but entire integrated energy ecosystems. In that context, Paducah is less notable as an individual project and more significant as an example of how the next generation of AI infrastructure may be developed: around existing energy assets, dedicated power resources, and regions capable of accommodating large new loads.
For power, utility, and data center stakeholders, that may be the most important takeaway. The competition for AI infrastructure is increasingly becoming a competition for infrastructure readiness, and projects like Paducah provide an early indication of what that future could look like.
Why it matters
● The project would repurpose the former DOE Paducah Site in western Kentucky into a major data center and energy hub.
● The announced project value is $100 billion, privately funded.
● The campus is expected to support up to 1.8 GW of utility capacity by 2032.
● The project includes more than 1.2 GW of planned compute capacity.
● The energy plan includes up to 4.6 GW of dedicated generation resources.
● NextEra’s planned dedicated power buildout includes up to 2 GW of natural gas generation and up to 2.6 GW of battery energy storage systems.
● The project is designed so dedicated power resources help avoid passing added costs to existing residents and businesses.
What the map shows
A Kentucky regional infrastructure view centered on the proposed Paducah American Energy Hub in western Kentucky.
● DOE Paducah Site / Brookfield Data Center Campus / Paducah American Energy Hub
● operational data centers
● under-development data centers
● transmission natural gas pipelines
● electricity transmission lines
● electricity substations
● operational power plants
The map uses existing infrastructure layers for context. It does not imply that every nearby power plant, pipeline, substation, or data center is directly tied to the Paducah project.
A deeper dive with Energy DataLink
Using Rextag Energy DataLink, users can:
● review proposed data center sites against nearby transmission lines and substations
● map natural gas transmission pipelines around planned power generation areas
● compare power plants, substations, and gas infrastructure in one regional view
● evaluate nearby data center activity and broader power demand context
● identify owner and operator information for relevant infrastructure layers where licensed
● build site-screening views that combine power, gas, data center, utility, and geographic context
● export or share mapped infrastructure views for internal planning, market research, and investment review workflows
