Loading...
Loading...
Indiana is in the middle of the most aggressive coal fleet retirement in the Midwest, and the grid management challenge that transition creates is driving real AI investment at every major utility in the state. NIPSCO — Northern Indiana Public Service Company, an NiSource subsidiary serving the Gary-Hammond-South Bend industrial corridor — committed to exiting coal entirely by 2028 in its 2021 Integrated Resource Plan, replacing roughly 3,000 MW of coal with a portfolio of wind, solar, and battery storage. Duke Energy Indiana, which serves the Indianapolis metro and much of central and southern Indiana, filed a similar transition plan with the Indiana Utility Regulatory Commission (IURC) that phases out coal units at Cayuga, Gibson, and Edwardsport through the late 2020s. Indiana Michigan Power (an AEP subsidiary serving Fort Wayne and eastern Indiana) is navigating PJM capacity obligations while shedding its coal exposure under AEP's corporate-wide decarbonization commitments. Each of these transitions requires AI capabilities that Indiana utilities simply didn't need when they were dispatching flat-running coal units: ultra-short-term wind and solar forecasting, battery storage optimization, and load-flow modeling that accounts for the very different spatial footprint of distributed renewables versus centralized coal plants. The IURC has been an active participant in shaping these investments through rate-case cost recovery proceedings and its Grid Modernization Collaborative established in 2022. LocalAISource connects Indiana utilities, industrial customers, and co-ops with AI practitioners who understand the specific transition economics of MISO-South Central and PJM-Indiana markets.
Updated June 2026
NIPSCO's service territory in northwestern Indiana sits in one of the best wind resource areas east of the Mississippi — the Calumet region and the Indiana dunes corridor consistently deliver capacity factors above 40% for utility-scale wind. By the time NIPSCO's coal retirement schedule is complete, the utility's generation portfolio will shift from a dispatchable-on-demand coal/gas mix to a portfolio dominated by assets whose output is determined by weather, not operator commands. That transformation creates a load-resource balancing challenge that NIPSCO's existing EMS, built around predictable coal dispatch, is not designed to handle. Machine learning short-term wind forecasting — particularly the 4-hour and 24-hour horizons that drive MISO day-ahead and real-time market commitment decisions — is now a core planning investment for NIPSCO. The industrial load base in NIPSCO's territory amplifies the stakes: ArcelorMittal's Burns Harbor steel complex (now Cleveland-Cliffs), BP's Whiting refinery, and multiple specialty chemical manufacturers in Lake County represent several hundred megawatts of load that can shift consumption timing by hours in response to price signals — but only if NIPSCO's demand-response AI can communicate price forecasts reliably enough for industrial energy managers to act on them. NIPSCO's 2022 IRP identified AI-assisted demand response as one of the cheapest capacity resources in its portfolio, at an estimated cost of $25–$45 per kW-year compared to $75–$120 per kW-year for new peaking capacity. Making that economics work requires ML forecasting accurate enough that NIPSCO can commit to demand-response dispatch events 24 hours ahead without over-calling assets and straining industrial customer relationships.
Duke Energy Indiana operates a transmission system that spans most of central Indiana, interconnected with PJM at the Indiana-Ohio border and with MISO in southern Indiana — a dual-market position that creates similar but not identical challenges to Illinois' PJM-MISO seam. Duke Indiana's Edwardsport IGCC plant, a technically complex coal-gasification facility near Vincennes, has been the most expensive electric generating plant ever built in Indiana and a perennial source of regulatory controversy at the IURC. As Edwardsport is phased toward retirement, the AI optimization tools Duke has used to maximize efficiency from its complex operation are being redirected to the dispatch-optimization challenge of managing a large solar portfolio across central Indiana's relatively flat, agriculture-dominated terrain. Duke Energy's broader corporate AI program — developed at its Charlotte headquarters and deployed across operating companies — includes ML-based vegetation management prioritization for transmission ROW and an automated outage management system that uses ML to predict fault locations from smart-meter pinging patterns during outages, substantially reducing truck-roll time for distribution crews. In central Indiana, where Duke Energy Indiana serves major industrial customers at Eli Lilly's manufacturing campuses in Indianapolis and Lebanon, as well as the Honda plant in Greensburg and Subaru's Lafayette facility, AI power-quality monitoring has become a customer-retention tool. Precision pharmaceutical manufacturing and automotive assembly both have tight voltage and frequency tolerance requirements, and Duke's AI-assisted substation monitoring program flags power-quality anomalies for industrial accounts before they escalate to complaints.
Indiana has 38 rural electric cooperatives, most served by Hoosier Energy or Indiana Michigan Power at the wholesale level, whose metering and customer-service infrastructure lags significantly behind investor-owned utilities. The IURC's Grid Modernization Collaborative has focused primarily on IOU investments, but the Indiana Statewide Association of Rural Electric Cooperatives (ISAREC) has been facilitating a shared-services approach to AMI data analytics and AI customer service tools that allows smaller co-ops to access capabilities they couldn't afford individually. AI automated meter reading — replacing legacy drive-by AMR with interval AMI that feeds ML anomaly detection — is the highest near-term ROI application for co-ops serving the agricultural southeastern quadrant of Indiana, where meter tampering and service theft around cannabis cultivation operations (legal since 2023 for medical, significant gray-market activity before that) has been a documented non-technical loss problem. For IOU customer service at scale, Duke Energy Indiana has deployed AI-powered chatbot and IVR systems that handle routine outage reporting, payment arrangement requests, and energy efficiency recommendations without live agent involvement. The challenge specific to Indiana is the state's unusual mix of IOU and cooperative service territories — a customer in rural Madison County might be served by either Duke Energy Indiana or Northeastern REMC depending on their specific address, and AI customer-service routing tools have to account for this fragmented territory map rather than assuming clean service-territory boundaries.
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
NIPSCO's approved resource plan includes roughly 1,600 MW of battery storage to be added by 2028, co-located primarily with utility-scale solar projects in Lake, Porter, and LaPorte counties. Battery dispatch optimization is being managed through AI-assisted EMS that models day-ahead MISO market prices, real-time renewable output forecasts, and distribution congestion patterns to determine optimal charge/discharge schedules. NIPSCO contracted with PowerFlex (an EDF subsidiary) for some of its behind-the-meter storage optimization and is working with ABB on EMS modernization for its substation automation program.
The Indiana Utility Regulatory Commission's Grid Modernization Collaborative, convened in 2022 with participation from NIPSCO, Duke Energy Indiana, Indiana Michigan Power, and Vectren (now CenterPoint Energy Indiana), is an ongoing multi-stakeholder process for developing cost-recovery and technology standards for grid modernization investments. Utilities present proposed AI and automation programs to the Collaborative for stakeholder input before filing formal rate-case recovery requests with the IURC. This process has streamlined cost recovery for AMI analytics, FLISR automation, and demand-response AI programs that previously required individual IURC proceedings, reducing the regulatory timeline for deploying approved technologies.
Indiana Michigan Power (IMP) serves Fort Wayne and eastern Indiana under PJM, but its generation portfolio and some transmission arrangements interact with MISO's market in ways that create basis-risk forecasting challenges similar to those in northern Illinois. IMP's Fort Wayne load centers are close to the AEP transmission system that spans PJM-West, meaning locational marginal prices at key delivery points can diverge significantly from PJM-wide averages during grid constraints. AI LMP forecasting for IMP requires node-specific models that capture the East Central Area Reliability coordination zone dynamics — a level of market granularity that requires MISO and PJM data access and domain expertise in AEP's transmission system topology.
Yes. The concentration of automotive assembly plants — Subaru in Lafayette, Honda in Greensburg, Toyota in Princeton, and dozens of Tier 1 and Tier 2 suppliers across central and southern Indiana — creates industrial load clusters with distinctive AI demand-response profiles. Automotive assembly runs on takt-time-based production schedules that create predictable ramp patterns at shift changes. AI demand-response models tuned to automotive production schedules can identify sub-plant flexibility windows (paint shop, body shop, compressed air systems) that can be dispatched without disrupting production flow, and Indiana utilities have documented this application in IURC filings as a cost-effective peak-shaving resource.
Indiana co-op AI engagements are typically scoped as shared-service projects through ISAREC or Hoosier Energy's member services program, with per-co-op costs ranging from $20K–$80K for AMI analytics and AI customer-service tools on a hosted platform. Independent co-op projects run higher — $75K–$200K for a full AMI data analytics implementation. For investor-owned utilities, SCADA anomaly detection pilots at defined substation scopes typically cost $75K–$200K, with enterprise-level EMS AI integration in the $500K–$1.5M range. These costs are generally recoverable through IURC-approved grid modernization riders, which substantially reduces the net cost to the utility's rate base.
Browse verified professionals across Indiana.