Loading...
Loading...
New Hampshire's electric utility market is small by national standards but operationally complex in ways that matter for AI implementation — and the complexity stems from three interacting factors that are unique to this state. First, Eversource Energy's New Hampshire subsidiary serves approximately 530,000 customers in the southern and western portions of the state, while Liberty Utilities serves roughly 42,000 customers in the northern and seacoast areas, creating a fragmented service territory where two utilities with very different technology investment capacity share the same ISO-NE balancing authority. Second, the Seabrook Station nuclear power plant in Seabrook — operated by NextEra Energy Resources and providing roughly 1,244 MW to the ISO-NE New England Pool — supplies a meaningful fraction of New Hampshire's in-state generation and the entire New England grid's baseload reliability during winter peaks. Seabrook's operational continuity is a grid stability question for all of New England, not just New Hampshire, which makes AI-assisted operations monitoring at the plant a regional reliability concern. Third, the New Hampshire Public Utilities Commission has been operating under significant political pressure on electricity rates — New Hampshire consistently ranks among the top five most expensive states for residential electricity — and the PUC's rate case review process is unusually adversarial, with the Office of Consumer Advocate and Attorney General's office both actively challenging utility capital expenditures. That regulatory environment shapes what AI gets approved for rate recovery and what gets funded from discretionary budgets. The Merrimack Valley load center — encompassing Manchester, Nashua, and the Route 3 industrial corridor — is the most economically active part of the state, with BAE Systems' electronic warfare and defense electronics manufacturing in Nashua, the Fidelity Investments operations campus in Merrimack, and DEKA Research's engineering complex in Manchester all generating large, sophisticated industrial and commercial loads that differ substantially from the residential winter-heating demand in the North Country. LocalAISource connects New Hampshire utilities and their industrial customers with AI professionals who understand ISO-NE interconnection constraints, NH PUC regulatory dynamics, and the specific demand pattern challenges of a compact, high-price-sensitivity service territory.
Updated June 2026
Seabrook Station's 1,244 MW of nuclear baseload output clears into ISO-NE's Forward Capacity Market and is allocated to load serving entities across New England through the capacity market settlement process. For Eversource New Hampshire, accurate load forecasting isn't just an operational convenience — it's a financial obligation, because Eversource NH's Forward Capacity Market position is determined by its forecasted peak load contribution to the New England system. Load forecast errors that cause Eversource NH to over- or under-procure capacity result in real financial settlements, and the ISO-NE zone for New Hampshire (Internal New England, ISONE-NE zone) has been among the tighter zones for reserve margin in recent winters. The specific load forecasting challenge in Eversource New Hampshire's territory is winter heating load in older housing stock. New Hampshire has one of the highest rates of fuel oil heating dependence in the country — roughly 40% of homes still use fuel oil as primary heat — and when oil prices spiked in 2022 following the Ukraine conflict, a measurable segment of Eversource NH customers switched to supplemental electric heating, creating unexpected winter demand growth that the utility's traditional degree-day forecasting models didn't capture. ML models that incorporate fuel oil price differentials as a feature input have been shown to predict oil-to-electric heating fuel switching at the census-tract level with sufficient accuracy to improve Eversource NH's winter peak forecast by 2–3 percentage points MAPE — a meaningful improvement in a zone where 1% forecast error represents 8–10 MW of reserve margin exposure. Eversource NH's load research team in Manchester has presented this methodology at the New England Energy Statistics working group, which meets quarterly with ISO-NE representatives to review regional demand forecast accuracy.
The December 2008 ice storm — which left 400,000 New Hampshire customers without power for up to two weeks, causing Governor Lynch to call it the worst natural disaster in state history — catalyzed a decade of grid hardening investment by Eversource NH and its predecessor PSNH. That investment was funded through a series of PUC-approved rate cases that built a database of infrastructure condition data far richer than most utilities of comparable size maintain. AI predictive maintenance models trained on Eversource NH's post-2008 infrastructure database — including pole inspection records, circuit recloser maintenance logs, and overhead line replacement histories — have a higher-quality training dataset than the same models would have at a utility that hadn't undergone systematic hardening documentation. Eversource NH's distribution SCADA system monitors approximately 130 automated switching points across the state's 14 operating districts, with the highest density of automation in the Manchester-Nashua corridor where BAE Systems, Fidelity, and the dense commercial strip along Daniel Webster Highway create the highest customer economic-impact-per-outage-minute ratio. AI-assisted fault location and isolation tools that use the switching point telemetry to automatically identify the faulted segment and reconfigure around it have reduced sustained outage duration on automated feeders by 35–45% in Eversource NH's most recent reliability performance reports to the PUC. Liberty Utilities' smaller northern service territory, which includes the seacoast communities of Exeter and Hampton, has less automation density but is exploring AI predictive maintenance through the Algonquin Power and Utilities parent company's enterprise technology program.
New Hampshire's electricity prices — averaging 24–27 cents/kWh for residential customers, among the highest in the continental U.S. — create both strong customer demand for AI energy management tools and strong regulatory sensitivity to utility technology investments that appear to add cost without measurable customer benefit. The NH PUC's Office of Consumer Advocate has been specifically critical of utility AMI investments that produced expected smart meter benefits without documented rate impact reduction, and that history means AI tools deployed on AMI data need to demonstrate ratepayer benefit through mechanisms the OCA will accept — primarily through demand response value, non-technical loss detection revenue recovery, or documented avoided capital deferral. Eversource NH's AMI deployment covered approximately 90% of its service territory by 2022, and the interval data is supporting behavioral energy efficiency programs — personalized home energy reports, time-of-use rate pilot enrollment targeting — that the PUC approved as part of Eversource's Energy Efficiency program portfolio. The New Hampshire Electric Cooperative, which serves rural areas not covered by Eversource or Liberty, is a member-owned cooperative with its own AMI deployment and is piloting AI demand response tools through the NRECA CoopTech program in its North Country territory. Commercial customers in the Merrimack Valley industrial corridor — particularly the defense electronics manufacturers who run three-shift operations and generate consistent 24-hour demand profiles — have deployed AI demand management tools that specifically target Eversource NH's demand charge structure, where the peak demand charge component represents 40–50% of total electric bills for large commercial accounts.
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
Seabrook's 1,244 MW clears through ISO-NE's Forward Capacity Market, and NextEra's capacity obligation management for Seabrook involves AI-assisted outage scheduling optimization — minimizing the overlap of Seabrook refueling outages with ISO-NE high-demand windows. For New Hampshire load-serving entities, Seabrook's capacity market clearing price directly affects the capacity component of retail electricity rates, and the PUC's rate case review scrutinizes ISO-NE capacity procurement decisions closely. AI tools that improve Eversource NH's capacity market position management — by more accurately forecasting peak load contribution and bidding demand response resources efficiently — translate directly to rate impact reduction that the OCA values.
For Eversource NH's scale — approximately 530,000 customers, 130 automated switching points — a full AI-assisted FLISR deployment covering the Merrimack Valley operating districts costs $600K–$1.2M in integration and configuration, leveraging Eversource's enterprise SCADA platform from its Massachusetts operations. The incremental cost for NH-specific implementation is lower than a greenfield deployment because Eversource's enterprise AI platform — developed through its NSTAR/PSNH/WMECo system integration — is already in production at comparable service territories. Annual maintenance and model update costs run $80K–$150K. The PUC has accepted FLISR reliability improvements as a rate base justification when documented against NH-specific SAIDI improvement targets.
BAE Systems' Nashua facilities run continuous production with classified operations that create predictable but non-deferrable load profiles — demand response is limited by production security requirements. The highest-value AI application for BAE's facilities is power quality monitoring: AI anomaly detection on power quality data that catches harmonic distortion and voltage sag events before they damage precision manufacturing equipment. Fidelity's Merrimack campus, which runs large data center loads, is better suited for demand charge optimization AI — Fidelity has deployed building management AI that reduces coincident peak contribution by pre-cooling the facility during the overnight hours before expected peak days.
Roughly 40% of New Hampshire homes use fuel oil as primary heat, and during oil price spikes they shift to supplemental electric resistance heat — creating winter demand surges that don't follow temperature curves as predictably as a state where electric heat is the primary heating fuel. AI models that incorporate oil price forecasts and New Hampshire DOE home heating oil survey data as features have produced 2–3 percentage point MAPE improvements on Eversource NH's January-February peak forecasts compared to degree-day-only models. This is not a feature any off-the-shelf utility load forecasting platform provides out of the box; it requires custom feature engineering from New England energy market data sources.
Eversource NH's capital investments recover through rate base on a traditional cost-of-service basis, with the PUC reviewing capital additions in rate cases filed approximately every 3–4 years. AI tools that are capitalized — integrated into SCADA infrastructure or AMI management systems — enter rate base and earn a return; AI tools purchased as operating expense SaaS subscriptions do not. Eversource NH has structured several AI implementations as capital additions specifically to earn a return on the investment through the PUC-approved rate base, which changes the procurement economics compared to states where utilities expense AI tools. The current Eversource NH rate case, filed in late 2024, includes AI-assisted grid management costs as an explicit capital addition.
Get listed on LocalAISource starting at $49/mo.