The Algorithmic Defense: Why Northeast Utilities Are Importing Wildfire AI

SAN FRANCISCO and HAMPTON, N.H. – October 21, 2026 — For decades, the primary existential threat to New England's electrical grid fell from the sky in frozen, heavy sheets. Regional utilities built their entire operational playbooks around ice storms, nor'easters, and the grinding weight of 50-plus inches of annual snowfall. But the geography of environmental risk is shifting rapidly, and with it, the strategic posture of the companies tasked with keeping the lights on.

Today's announcement that Unitil Corporation has signed a three-year partnership with San Francisco-based software firm Rhizome signals a profound realignment in how non-traditional risk regions are managing their infrastructure. Unitil, which serves over 215,000 customers across Maine, Massachusetts, and New Hampshire, will deploy Rhizome's artificial intelligence platforms—gridADAPT and gridFIRM—to model multi-hazard, climate-driven risks.

On its surface, this is a technology procurement story. But beneath the headlines, this partnership reveals a much more complex narrative about the modern utility sector: the urgent need to import Western-style wildfire mitigation tactics to the East Coast, and the growing reliance on algorithms to justify massive capital expenditures to increasingly skeptical state regulators.

The New Front Line of Climate Risk

To understand the intent behind Unitil's maneuver, one must look at the shifting meteorological baseline in the Northeast. New England is not historically known for its wildfire risk; it is a region characterized by dense, damp, broadleaf forests. Yet, the data tells a rapidly evolving story. New Hampshire alone has recorded more than 504 wildfires since 2021, culminating in the state breaking a 130-year record for its driest summer in 2025.

Utilities like Unitil are now caught in a compound-hazard trap. They must maintain traditional winter storm hardening while simultaneously retrofitting their systems for the kind of prolonged drought and fire exposure typically associated with California or the Pacific Northwest. Managing two substantial and growing risk factors at once requires a level of predictive capability that traditional, backward-looking utility planning models simply cannot provide.

"We are excited to partner with Rhizome to better understand the climate risks facing our systems and prioritize investments that strengthen resilience," said Sara Sankowich, Unitil's Vice President of Shared Services and Sustainability, in the press release. "This work will help us make informed, cost-effective decisions that benefit our customers for years to come."

Sankowich's emphasis on "cost-effective decisions" is the critical underlying signal here. The adoption of AI is not merely an operational upgrade; it is a financial defense mechanism.

Algorithmic Rate Cases and Regulatory Scrutiny

In the utility business, the phrase "prioritize resilience investments" is often a polite proxy for "convincing the public utility commission to approve our capital expenditure." Both the Massachusetts Department of Public Utilities (DPU) and the New Hampshire Public Utilities Commission (NHPUC) have historically maintained rigorous oversight of utility spending.

With resilience spending increasingly scrutinized at the state level, utilities are facing a severe affordability squeeze. They are burdened by aging infrastructure, load growth demands from electrification, supply chain constraints, and compounding weather threats. When a utility goes to a regulatory body to request rate increases to fund grid hardening, they can no longer rely on broad warnings about climate change. They need granular, defensible, asset-level data.

This is the precise friction point where Rhizome's software enters the equation. By integrating climate models, historical operational data, and machine learning, platforms like gridADAPT and gridFIRM translate meteorological probabilities into financial cost-benefit analyses.

"New England has become a region where utilities are thinking holistically about the system risks they face today and tomorrow," said Rhizome co-founder and CEO Mishal Thadani. "Today's partnership announcement with Unitil is the first step in first taking a uniform approach to risk quantification across all hazards, then improving resilience and investment planning to keep their customers' lights on and bills low."

Thadani's strategy is clear: position the algorithm as the ultimate arbiter of prudence. If a utility can approach a regulator with an AI-generated model demonstrating that spending $10 million on vegetation management and line undergrounding in a specific New Hampshire zip code will avoid $50 million in outage and liability costs over the next decade, the rate case practically writes itself. It shifts the burden of proof, turning subjective infrastructure debates into objective, data-backed mandates.

The Black Box of Capital Expenditure

However, the introduction of machine learning into ratepayer economics is not without controversy. Consumer advocacy groups, such as the New Hampshire Office of the Consumer Advocate and the Massachusetts Attorney General's Office Energy and Ratepayer Advocacy Division, are inherently wary of any mechanism that might accelerate capital expenditure approvals without guaranteed consumer benefits.

As one independent regulatory analyst noted privately, there is a persistent fear that AI risk models could function as a "black box" for utilities, generating unassailable, highly technical justifications for baseline infrastructure spending that ultimately drives up consumer electricity rates. If the algorithm dictates that a massive capital upgrade is the only way to mitigate a newly identified wildfire risk, consumer advocates are left with few tools to challenge the underlying assumptions of the proprietary software.

The success of the Rhizome-Unitil partnership will therefore be judged not just by whether the software accurately predicts where a tree might fall on a power line during a drought, but by whether its recommendations genuinely reduce the cost burden on ratepayers. Directing limited capital to where it reduces the most risk per dollar is a sound philosophy, but the discipline matters most when every single resilience dollar is ultimately recovered from the public.

The Rapid Ascendancy of Climate-Grid Tech

The broader context of this partnership is the explosive growth of the climate-grid technology sector. Rhizome, launched only in 2023, has rapidly assembled a formidable portfolio of clients. Today's announcement follows a December 2025 partnership with National Grid covering Massachusetts, New York, and the UK, as well as a March 2026 deployment with Horizon Energy Network in New Zealand. Within New England, Rhizome has already secured contracts with Vermont Electric Power Company (VELCO) and Vermont Electric Cooperative (VEC).

This rapid market penetration underscores a systemic panic within the global utility sector. The traditional models for managing grid reliability are failing under the weight of unprecedented climate volatility. Federal funding initiatives, combined with intense shareholder pressure regarding environmental and governance (ESG) metrics, have flooded the zone with capital, but utilities desperately need the analytical frameworks to deploy that capital effectively.

By moving aggressively into non-traditional wildfire regions like New England, Rhizome is proving that climate risk is no longer a localized phenomenon confined to the American West. The compound threats of drought, fire, and ice require a unified, algorithmic approach to infrastructure defense, permanently altering the dynamic between utility operators, the regulators who watch them, and the ratepayers who fund them.