Loading...
Loading...
Kentucky's electric grid is the most coal-dependent in the United States, with roughly 70% of electricity generation still coming from coal-fired plants as of 2024 — a figure that is declining but slowly, given the state's political and economic relationship with coal production and the Kentucky Public Service Commission's (KPSC) historically conservative approach to renewable mandates. Louisville Gas and Electric and Kentucky Utilities (LG&E and KU), operating companies of PPL Corporation since its 2011 acquisition of E.ON's U.S. assets, serve roughly 1.3 million customers across central and western Kentucky under KPSC regulation. The Tennessee Valley Authority provides power to some distribution cooperatives in eastern and southeastern Kentucky, creating a dual-regulatory landscape that complicates statewide grid planning. Coal's dominance in Kentucky's generation mix shapes the AI opportunity in a specific way: the most immediate AI applications are not renewable integration forecasting (as in Iowa or Kansas) but rather coal plant predictive maintenance — extending the operational life of assets that utilities expect to continue running for years longer than nationally average timelines — and preparing the grid infrastructure for the eventual transition that federal EPA regulations (specifically the 2024 Power Plant rules) will ultimately force. LG&E and KU's 2022 Integrated Resource Plan, approved by the KPSC, projects coal fleet retirement through the 2030s with replacement from natural gas, solar, and battery storage — a transition that requires AI capabilities the utility is actively building. LocalAISource connects Kentucky utilities, industrial customers, and co-ops with AI practitioners who understand KPSC regulatory economics and the specific operational challenges of large coal-fleet utilities.
Updated June 2026
LG&E operates the Mill Creek and Cane Run generating stations near Louisville; KU operates the Ghent, Brown, and Trimble County coal plants in north-central Kentucky. Collectively, these facilities represent billions of dollars in rate-base assets that PPL has committed to operating for at least another decade. The economics of coal plant AI are different from gas or nuclear: coal plants have high forced-outage rates compared to modern combined-cycle gas, driven by fuel handling equipment (crushers, conveyers, pulverizers), boiler tube failures, and turbine-generator aging. A single forced outage at a 500 MW coal unit can cost $500,000–$2 million in replacement power purchases in a single day, depending on MISO spot market conditions at the time. Machine learning vibration analysis on pulverizers and coal mill drives, AI-assisted boiler tube leak detection using acoustic emission sensors, and thermal anomaly detection on turbine-generator bearings are all deployed at LG&E and KU plants — programs developed in part through LG&E's membership in EPRI's (Electric Power Research Institute) Maintenance Technology and Analytics users group, which is headquartered in Charlotte but has significant Kentucky utility participation. The shortlist criterion for coal-plant AI vendors in Kentucky is not just ML capability but direct experience with subcritical and supercritical boiler configurations — the specific sensor geometries and failure modes of Kentucky's predominantly 1960s-1980s vintage generating fleet require models trained on similar equipment, not generic industrial anomaly detection algorithms.
LG&E and KU's transmission system spans 2,700 miles of high-voltage transmission across 77 Kentucky counties, interconnected with MISO's South-Central market in western Kentucky and with PJM in eastern and northern Kentucky. The KPSC's approach to grid modernization investment has been cautious — the commission has historically required demonstrable reliability or cost benefits before approving technology investments in rate recovery proceedings — which means AI projects at LG&E and KU need robust ROI documentation before reaching deployment scale. That said, KPSC has approved LG&E's Smart Grid program, which includes AMI deployment across Louisville Gas and Electric's service territory, and the data infrastructure from that AMI rollout provides the substrate for AI customer analytics and distribution automation programs that are now in production. SCADA anomaly detection for Kentucky's transmission system is a live investment area, particularly for the aging infrastructure in western Kentucky coal country where original equipment from LG&E's Big Sandy and Coleman plants interconnects with MISO at the Henderson, Kentucky 345 kV substation complex. The TVA-LG&E seam in southeastern Kentucky creates additional planning complexity: the boundary between PPL's regulated service territory and TVA's federal power authority jurisdiction runs through Pulaski, Wayne, and McCreary counties, and AI power-flow tools that don't account for TVA's separate scheduling and dispatch protocols will consistently misfore at the interchange points. Ask any transmission planning engineer at LG&E and they'll tell you that the TVA boundary is the single most commonly mishandled element in outside consulting firm grid models for Kentucky.
LG&E and KU's service territory includes some of the most physically complex electric infrastructure in the country: river-crossing transmission structures on the Ohio and Kentucky Rivers, aging distribution systems in Appalachian eastern Kentucky (served primarily by co-ops like Big Sandy REMC and Appalachian Power in parts of the state), and coal plant switchyards that have been in continuous operation since the mid-20th century. Drone inspection with AI-assisted defect detection is now used for LG&E's Ohio River crossing structures — locations where the combination of extreme tower heights, barge-traffic airspace restrictions, and weather-window limitations make traditional helicopter patrol difficult. Computer vision inspection of coal plant cooling towers — a specialized application that uses thermal imaging to identify delamination and structural cracking in reinforced concrete hyperbolic towers — is an active deployment at several Kentucky generating stations. Kentucky's Appalachian coalfield region in the east presents a different inspection challenge: steep terrain, narrow hollow roads, and legacy distribution infrastructure serving communities that have seen decades of population decline. AI predictive infrastructure scoring for rural distribution — identifying poles and transformers statistically likely to fail before the next inspection cycle, based on age, maintenance history, and weather-event exposure data — allows LG&E's co-op wholesale partners (like Kentucky Mountain REMC and Cumberland Valley Electric) to target limited capital maintenance budgets more efficiently than equal-weighted replacement schedules.
Connecting AI systems to existing business infrastructure and workflows
Workflow automation using AI, including Make.com-style automation and RPA
Predictive models, data analysis, and ML pipeline development
Image recognition, object detection, video analysis, and visual inspection systems
The Kentucky Public Service Commission has historically required utilities to demonstrate concrete reliability improvements or cost savings before approving technology investments for rate recovery. LG&E and KU submit grid modernization plans with detailed ROI analyses — including forecasted SAIDI/SAIFI improvements, fuel cost savings from predictive maintenance, and O&M cost reductions — as part of their rate case filings. The commission's 2022 approval of LG&E's Smart Grid rider included explicit performance metrics that the utility must meet to retain cost recovery, making AI projects with measurable outcomes significantly more approvable than speculative capability investments.
Kentucky's coal units are aging assets with high forced-outage rates that directly translate to MISO replacement-power purchases — at current prices, a forced outage at a 600 MW LG&E coal unit can cost $600,000–$2 million in a single day. Renewable assets (wind and solar) have lower forced-outage rates and lower replacement-power cost exposure when they do fail. Coal plant AI maintenance ROI is therefore computable and compelling in a way that a solar panel string-inverter monitoring program simply isn't at the same dollar scale. LG&E and KU's EPRI participation gives them access to fleet-benchmarked AI maintenance standards without reinventing the wheel.
TVA serves several Kentucky distribution cooperatives under long-term wholesale power contracts, including Big Sandy REMC and portions of the southeast. TVA operates under a federal charter and its own planning standards rather than KPSC jurisdiction, which means co-ops served by TVA wholesale power are subject to different reliability reporting and infrastructure investment frameworks than LG&E and KU's regulated service territory. AI load forecasting and grid planning tools used by TVA-served co-ops in Kentucky need to incorporate TVA's rate structures, demand-response programs (TVA's eScore and Valley EnergyRight programs), and TVA's generation dispatch logic, which differs materially from MISO market dispatch.
Kentucky has 26 rural electric cooperatives and several significant municipal utilities (including Louisville Gas and Electric's parent and the city-owned utilities in Bardstown and Bowling Green). Co-op AI projects are typically accessed through the Kentucky Association of Electric Cooperatives (KAEC) technology programs or through Big Rivers Electric Corporation's member services. Managed-platform AMI analytics programs for a mid-size co-op run $25K–$90K annually. Distribution infrastructure inspection using drones with AI defect classification runs $0.50–$2.00 per structure inspected depending on terrain and access difficulty — western Kentucky's flat farmland is at the low end; eastern Kentucky's hollow roads and steep terrain are at the high end.
Yes. Kentucky's coal plants receive fuel by river barge on the Ohio, Green, and Big Sandy Rivers, by rail, and by truck — three distinct supply chains each with its own AI optimization opportunity. AI-assisted coal inventory management, which forecasts fuel consumption based on ML load forecasts and unit commitment plans and optimizes barge and rail delivery scheduling to minimize demurrage costs, has been deployed at several Kentucky generating stations. Stockpile monitoring using drone-based volumetric surveying (LiDAR combined with AI volume estimation) allows coal plant operators to track actual fuel inventory weekly rather than relying on load-cell estimates that drift between calibration cycles — a practical improvement with direct O&M cost implications.
Get discovered by Kentucky businesses looking for AI expertise.
Get Listed