Domain-informed AI Couldhelp Electricity Grids Plan for Rising Demand and Changing

Artificial intelligence has become a major new pressure on the electricity grid at the same time as it is emerging as one of the most promising tools for making that grid resilient, according to a new publication by United Nations University scientists. The policy question, the authors argue, is no longer whether to deploy AI for grid resilience, but under what guardrails. Domain-informed AI could help electricity grids plan for rising demand and changing climate.

September 7, 2026 Domain-informed AI could help electricity grids plan for rising demand and changing climate by United Nations University edited by Sadie Harley, reviewed by Andrew Zinin Sadie Harley Scientific Editor Meet our editorial team Behind our editorial process Andrew Zinin Chief Editor Meet our editorial team Behind our editorial process Editors' notes This article has been reviewed according to Science X's editorial process and policies. Editors have highlighted the following attributes while ensuring the content's credibility: fact-checked trusted source proofread The GIST Add as preferred source The policy question, the authors argue, is no longer whether to deploy AI for grid resilience, but under what guardrails. The policy brief, " Employing Domain-Informed AI for Energy Planning and Decarbonization under Uncertainty," published by the United Nations University Institute for Water, Environment and Health (UNU-INWEH), warns that governments are approving electricity infrastructure with 15–20-year lifespans on the basis of historical weather records that are unlikely to hold in the coming decades.

Written for national regulators and climate-finance institutions, the publication is a step forward in setting out safeguards to support accelerated decarbonization targets through 2030. At the center of the energy transition, the policy brief describes a feedback loop. Renewable energy is a critical instrument of climate change mitigation, yet it is more susceptible to adverse climate conditions than fossil fuel-based generation, so the technologies deployed to slow climate change become less predictable as it advances.

Warming is simultaneously pushing demand upward through more air conditioning and more groundwater pumping in drier conditions. The renewable sector is projected to nearly triple in size by 2030, according to the International Energy Agency's Renewables 2024, but that expansion is being planned against shifting rather than stable weather. Demand is rising faster than planners assumed.

The authors warn that the growth trajectory of power-intensive end uses, including data centers and electric vehicles, already exceeds anticipated capacity expansion needs. Unregulated AI data centers consume vast amounts of electricity and directly strain the grids they are built on.

"We are asking the grid to absorb more renewable energy and more demand at the same time, and we are making those decisions with data that describes a climate we no longer live in," said Dr. Renee Obringer, research fellow of Urban and Interdependent Infrastructure Systems at UNU-INWEH and lead author of the publication.

"The uncomfortable part is that AI sits on both sides of the ledger. It is one of the reasons demand is climbing, and it is also the fastest route we have to planning for what is coming." The policy brief identifies the failure of long-term capital planning to account for forward-looking climate risk as a primary barrier to grid resilience and frames it as a fiscal exposure rather than a technical one.

Deploying capital based on stationary historical data for assets that will run 15–20 years accumulates stranded assets as physical climate risks materialize. For climate-finance institutions, that makes forward-looking risk assessment a condition of sound capital allocation.

Research on climate impacts has advanced, the brief concludes, but it is not consistently reaching the regulatory agencies responsible for capacity planning, which still rely on historical weather data to anticipate demand spikes. Domain-informed AI, the publication argues, is one of many necessary solutions rather than a remedy on its own.

These algorithms are more transparent about how they are trained and are tailored to a specific application area, such as energy systems, which sets them apart from large language models and general-purpose deep learning. Built for the problem, they often outperform general-purpose alternatives, and because their reasoning is legible, planners can fold them into existing processes.

Their value lies in what current tools cannot do. The general circulation models underpinning modern climate impact assessment are hard to integrate, hard to downscale to the spatial scales infrastructure decisions turn on, and slow to yield the variables energy systems modelers need.

Domain-informed AI models, by contrast, already deliver highly accurate short- and long-term renewable forecasts, and AI climate emulators are progressing rapidly in accuracy. The authors call on regulators to institutionalize adaptive governance frameworks, bring energy systems modelers together with climate scientists in integrated modeling across water, transport, and information and communications technology networks, and mandate transparent, domain-informed AI models as an enforceable standard for capacity planning approval.

To ensure safety and public trust, regulators should classify opaque deep learning systems as severe operational hazards and enforce strict guardrails against algorithmic training bias, while policymakers should require that AI data centers operate primarily on verifiable renewable energy. "AI is being offered to governments as an answer to the energy transition while quietly becoming one of its largest new burdens, and both of those things are true at once," said Professor Kaveh Madani, director of UNU-INWEH and a co-author of the brief.

"Rejecting these tools would slow decarbonization. Adopting them without transparency rules or limits on their own energy use would simply exchange one risk for another.

The task is not to choose between the two, but to set the conditions under which AI is allowed near critical infrastructure." More information Renee Obringer et al, Employing Domain-Informed AI for Energy Planning and Decarbonization under Uncertainty, (2026).