Loading...
Loading...
Connecticut's electricity system operates under two overlapping authorities that shape everything from procurement timelines to AI tool requirements: ISO New England manages the regional grid and competitive wholesale market for all of New England, while the Connecticut Public Utilities Regulatory Authority (PURA) and CT DEEP regulate distribution utilities and energy policy at the state level. Eversource Energy serves most of Connecticut's 1.4 million electric customers through its Connecticut Light and Power subsidiary, and United Illuminating — now an Avangrid subsidiary — serves New Haven and Bridgeport. Both utilities answer to PURA for their distribution operations and sell into ISO-NE's energy and capacity markets for their supply arrangements. The state's generation fleet is dominated by Dominion Energy's Millstone Nuclear Power Station in Waterford, a two-unit PWR plant that produces roughly 2,100 MW — about 50% of Connecticut's in-state generation — and has a long track record of ISO-NE capacity market participation. Connecticut's offshore wind ambitions are legally binding: PURA has overseen multiple offshore wind procurement solicitations, and Avangrid's Park City Wind and SouthCoast Wind (formerly Mayflower Wind) projects, combined with Eversource's partnership in Revolution Wind, are in various stages of development along the Connecticut and Rhode Island coast. As those projects come online, ISO-NE's real-time dispatch system will need to manage the variable output of multi-gigawatt offshore wind portfolios on a grid that already has minimal interconnection diversity — Connecticut has one AC transmission interconnection to New York and is otherwise dependent on New England's internal transmission network.
Updated June 2026
Millstone Unit 2 and Unit 3 — both pressurized water reactors operated by Dominion Energy Nuclear Connecticut — provide roughly half of Connecticut's electricity generation and a significant fraction of New England's low-carbon baseload. The plant's two-unit site configuration means that a forced outage at either unit can shift ISO-NE's real-time dispatch substantially, drawing power from natural gas peakers across New England and affecting LMP prices in Connecticut's load zone. Dominion's nuclear operations management at Millstone follows the same NRC quality assurance framework (10 CFR Part 50, Appendix B) that governs all US commercial nuclear plants, and AI predictive maintenance tools must satisfy those requirements to be used in maintenance-informed decisions. Dominion Energy Nuclear operates a fleet of 15 reactors and has enterprise-level AI predictive maintenance programs that Millstone participates in through Dominion's centralized nuclear technology organization — meaning plant-level AI decisions at Millstone are influenced by Dominion's corporate nuclear technology group in Richmond, Virginia, not just by plant management in Waterford. The recent license renewal extension for Millstone Unit 2 (through 2035) and Unit 3 (through 2045) means the plant has a long operating horizon ahead, and the economics of AI maintenance tools that extend the interval between forced outages are compelling when Millstone's capacity market revenue and ISO-NE market participation are factored in.
Connecticut's offshore wind procurement — driven by Public Act 19-71 and subsequent PURA proceedings — has resulted in executed power purchase agreements for approximately 2,000 MW of offshore wind capacity in the waters south of Long Island and Rhode Island. When Revolution Wind (704 MW, Eversource/Ørsted partnership), Park City Wind (804 MW, Avangrid), and associated projects come online, they will collectively represent a variable generation addition to ISO-NE's grid equal to roughly 20% of New England's peak demand. ISO-NE's market and operational systems were designed for a generation mix dominated by gas, nuclear, and hydro; adding large-block offshore wind changes the statistical properties of the generation forecast problem in ways that ISO-NE is actively managing. The New England Wind Integration Study, which ISO-NE published in cooperation with state agencies and utilities, identified specific forecasting and dispatch challenges for high-offshore-wind scenarios. AI load and generation forecasting tools deployed by Connecticut utilities need to be compatible with ISO-NE's forecast data sharing protocols and the market bidding requirements of ISO-NE's Forward Capacity Market, which prices capacity three years in advance and needs accurate output forecasts to set clearing prices. PURA has been an active participant in ISO-NE policy forums on offshore wind integration, and Connecticut utilities that want to deploy AI tools for wind management need to coordinate their approaches with ISO-NE's own technology roadmap.
Eversource Connecticut's distribution system — serving customers from Hartford to New Haven to the Shoreline — operates in a Northeast climate where ice storms, nor'easters, and tropical storm remnants create the highest-cost outage events in the utility's annual operating budget. The October 2011 nor'easter (Alfred), Superstorm Sandy in 2012, and Tropical Storm Isaias in 2020 each caused multi-day outages affecting hundreds of thousands of Connecticut customers and generated PURA investigations into Eversource's storm response. Isaias specifically resulted in a PURA performance management investigation and subsequent rate case proceedings where storm restoration performance became a metric tied to earnings. AI predictive maintenance that reduces the number of distribution line segments in a degraded pre-failure state before a major storm — by identifying equipment anomalies during routine patrols — directly reduces the number of storm-related failures and the restoration time those failures drive. Eversource's Smart Energy Solutions Center in Berlin, Connecticut coordinates the utility's advanced metering infrastructure program and grid modernization investments, and AI tools for distribution asset management are reviewed through that center's technology evaluation process. United Illuminating's smaller service territory in New Haven and Bridgeport — which includes some of the most densely populated and historically underinvested distribution infrastructure in the state — has a different failure mode profile: cable-heavy urban systems where underground cable failure is the primary outage driver, and AI thermal monitoring tools for underground cable systems are the relevant application, not overhead conductor inspection.
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
ISO-NE's Forward Capacity Market clears capacity prices for delivery three years forward — Millstone, Eversource's distribution-connected demand response resources, and offshore wind projects all participate. Utilities that use AI forecasting to more accurately project their peak load contribution to ISO-NE's demand forecast can optimize their capacity market positions, either buying or selling capacity more efficiently. The difference between a slightly high and slightly low load forecast can translate to millions in capacity obligation costs. PURA's oversight of Eversource and UI requires utilities to demonstrate that their capacity market strategies are cost-effective for ratepayers — AI forecasting that demonstrably reduces over-procurement of capacity has a regulatory cost-justification path.
New England's winter reliability risk — where natural gas-fired generation can be constrained by pipeline capacity during polar vortex events — is concentrated in Connecticut because the state has limited fuel oil backup capacity outside of Millstone's nuclear output. ISO-NE has declared winter reliability concerns in multiple recent years, and Connecticut utilities have been directed by PURA to maintain demand response programs that can reduce load during ISO-NE emergency alerts. AI demand response platforms that can aggregate controllable loads — commercial building HVAC, industrial process heat, EV charging — and respond to ISO-NE's 30-minute or 10-minute emergency dispatch signals are a documented reliability investment. Eversource's ConnectedSolutions program, which pays customers to reduce load during ISO-NE scarcity events, is the current implementation; AI optimization of which customers to dispatch and in what sequence is an improvement over current rule-based selection methods.
PURA's performance-based regulation framework ties a portion of Eversource's earnings to metrics including outage duration (SAIDI), outage frequency (SAIFI), and customer satisfaction. AI predictive maintenance that reduces unplanned outages improves SAIDI and SAIFI performance, directly affecting Eversource's earnings under the PBR mechanism. Eversource has an operational and financial incentive to invest in AI maintenance tools beyond generic efficiency — the PURA earnings linkage makes the ROI calculation more direct than it is for utilities without performance-based rate structures. United Illuminating under Avangrid faces similar incentives; Avangrid's parent company Iberdrola has enterprise-level grid AI programs that UI can draw on, which is a different procurement pathway than independent vendor engagement.
Millstone clears in ISO-NE's Forward Capacity Market at prices that have historically been $3–$8/kW-month, generating $100M–$200M annually in capacity revenue on top of energy market sales. That revenue creates a financial context where avoiding a single forced outage event — which might cost Millstone $15–40M in lost capacity and energy market revenue plus repair costs — justifies AI predictive maintenance investments in the $1M–$5M range with payback periods of 1–3 years. Dominion Energy Nuclear's actuarial data on forced outage causes, combined with Millstone's specific equipment age and maintenance history, is the input dataset that makes such ROI calculations credible rather than speculative.
CT DEEP (Department of Energy and Environmental Protection) is the state agency responsible for energy policy and the Integrated Resource Plan that governs Connecticut's long-term generation mix, while PURA regulates utility rates and service quality. AI tools used for long-term planning — renewable procurement optimization, carbon reduction pathway modeling, EV load integration forecasting — are more likely to originate in CT DEEP proceedings, where the framing is policy outcomes rather than utility cost recovery. Eversource and United Illuminating both engage in CT DEEP's IRP process and have filed AI-informed resource analyses in recent proceedings. Vendors who want to participate in Connecticut's energy AI market need to understand both regulatory tracks: PURA for distribution utility operations, CT DEEP for generation and capacity policy.
List your energy & utilities AI practice and connect with local businesses.
Get Listed