$100B AI Campus Takes Shape in Kentucky

Brookfield and NextEra plan a $100B Kentucky AI campus with 1.2 GW of compute and up to 4.6 GW of dedicated on-site power.

26 Aug 2026

Brookfield and NextEra Energy plan to invest $100bn in an artificial intelligence data centre campus at the former Paducah Gaseous Diffusion Plant in western Kentucky. Announced on July 29, the project would provide more than 1.2GW of computing capacity on the former US Department of Energy site.

Power would come from as much as 4.6GW of dedicated on-site infrastructure, including up to 2GW of natural gas generation and 2.6GW of battery storage. NextEra would own the generation assets, while Brookfield would operate the campus. The companies said the privately financed power system was designed to avoid passing infrastructure costs to existing electricity customers.

John Ketchum, NextEra’s chairman, president and chief executive, said the data centre “will bring its own power, pay for its own power infrastructure and create good-paying jobs for local workers.” The commitment comes as rapid growth in electricity demand from AI data centres raises questions across the US about grid capacity and how the cost of new generation and transmission should be shared.

Regional utilities Big Rivers Electric, Jackson Purchase Energy Cooperative and Paducah Power System are also involved, providing local grid expertise. Beyond additional jobs and tax revenue, the development would put a former federal industrial site back into large-scale commercial use.

The project reflects a broader effort by data centre developers to secure power directly as electricity demand from AI computing grows faster than conventional utility planning cycles. In Paducah, the proposed combination of dedicated generation and storage could limit reliance on existing regional infrastructure, although the campus will still require a large build-out of new energy assets.

Whether similar privately financed power models become common will depend on project economics, regulatory requirements and the pace of AI demand. For utilities and policymakers, the division of infrastructure costs between data centres and existing customers is likely to remain a central issue as larger computing projects seek connections to US power systems.