AI Aims to Speed Up Grid Planning Models
Two research projects funded by the Department of Energy test whether AI can keep pace with grid models that generate thousands of scenarios.
Updated on Nov. 5, 2026 in Energy

The Department of Energy has awarded Genesis Mission funding to Pacific Northwest National Laboratory for two projects that apply artificial intelligence to electric grid planning. The work is research-stage: the two teams have six months to demonstrate their research goals.
Why it matters
Grid planning models are growing so large and computationally demanding that they generate more scenarios than human teams can efficiently review, a pressure that rises as the grid faces roughly twenty years of increasing complexity and demand. Applying AI to speed these models is one way planners may keep pace with that growth.
The source reports the modeling scales involved, not measured speedups: electromagnetic transient models resolve grid physics at microsecond resolution, and capacity expansion models can produce ten thousand scenarios. Neither project has yet published a result comparing AI-assisted runtimes or scenario selection against conventional methods.
The players
Pacific Northwest National Laboratory
A Department of Energy national laboratory that received the Genesis Mission funding and leads work on grid modeling and energy systems research.
Arizona State University
A public research university that leads the second grid planning project, which pairs AI scenario selection with capacity expansion modeling.
Marcelo Elizondo
A PNNL researcher leading the first project, which applies machine learning to approximate electromagnetic transient grid calculations.
Casey Burleyson
A PNNL researcher leading the laboratory's contribution to the second, Arizona State-led grid planning project.
The details
Electromagnetic transient models simulate the fast electrical behavior of the grid, calculating physics at microsecond intervals, which makes them computationally heavy. The first project, led by PNNL's Marcelo Elizondo, uses a machine learning algorithm that approximates these complex calculations. Capacity expansion models, which distill fuel costs, demand, policies, and technology into long-term grid pathways, produce large sets of possible futures. The second project, led by Arizona State University with PNNL participation, plans to build an AI model that selects the significant scenarios from that output.
Timeline
Next six months: project teams are to demonstrate their research goals.
The Tech Race
The projects sit within a federal effort to apply AI to national energy infrastructure, with a national laboratory and a university each taking a role. Their six-month milestone will show whether AI approximation of grid models can match the fidelity of conventional methods at practical speed.
The projects are research-stage, and the source gives no availability, pricing, or deployment timeline for AI-assisted planning tools. Any effect on everyday electricity service would come only after the methods are validated and adopted by grid planners.
The takeaway
Grid planning is moving toward AI approximation as models outgrow what human teams can review. Watch whether the two teams demonstrate their research goals within the next six months, and whether the AI-assisted runs match conventional model output.
What happens next
Within the next six months, the two project teams are expected to demonstrate their research goals, which will show whether the AI-assisted approaches perform as planned.
Further reading
For broader context on how grid planning is changing, explore the Energy section.
Source note: This article includes information reported by HPCwire.





