AI for power-grid operations and forecasting: Junction Brief – November 2026

Acentric Research is produced with AI tools under human review. The authors are not financial professionals, and nothing here is investment advice or a recommendation to buy or sell any security. AI analysis can be wrong; check every claim against its cited source.

This Brief is an initial scan, not a full investigation. The method and scores behind it are in AI × Energy in the Landscape Report.

Overview

A power grid has to match supply and demand at every moment, plan where new lines and plants connect, and stay within physical limits. The U.S. Department of Energy (DOE) states that the grid faces reliability challenges and infrastructure limitations as it struggles to accommodate large increases in electricity demand from data centers, manufacturing and electrification. [1] This Brief covers artificial intelligence (AI) used to plan, forecast and operate the grid and to connect new generation to it.

What stands out is that the public record is mostly about intentions. A federal funding call states aims for faster decisions, and one large grid operator has announced a collaboration to speed up connecting new generation. No source opened for this Brief reports a measured result from AI in grid operations, such as a shorter connection time; vendors and operators may hold results this Brief did not open. The sources opened date from 2018 to 2026, and the trade report of the grid operator’s announcement is from April 2025; later developments were not searched.

Current state

A federal program. DOE’s Genesis Mission funding opportunity, issued in March 2026, includes a challenge titled “Scaling the Grid to Power the American Economy”. It states that AI, using deep and reinforcement learning on newly integrated big data sources, “will reduce uncertainty, improve insights, and speed processes in grid planning, interconnection, operations, and security”, and that the effort “aims to enable 20-100x faster decision-making and at least 10 percent improvement in electricity cost and reliability”. One focus area asks for a foundational power-system model, meaning a large AI model of the grid that could support planning and operations. These are the agency’s goals for funded work. [1]

A grid operator and a technology project. PJM Interconnection is the grid operator for a service area of 13 states and Washington, D.C. In April 2025 a trade publication reported that PJM had announced a multiyear collaboration with Google and Tapestry, with the stated aim of streamlining the planning process for connecting new generation resources and significantly cutting the time to review applications. Under it, Tapestry is to develop AI-enhanced tools and models for PJM’s interconnection process, powered by Google Cloud and Google DeepMind. The report adds that PJM’s own work to automate interconnection planning began in 2023. [2] Tapestry describes itself as “X’s moonshot for the electric grid”, building “an AI-powered, unified platform for every job across the grid”. [3]

Why connection time matters. The same trade report describes the pressure behind the PJM effort. It states that PJM’s separate initiative to fast-track shovel-ready generation projects drew 94 applications totaling 26.6 gigawatts, and that PJM’s capacity auction in July 2024 saw prices rise from $29 to $270 per megawatt-day. It lists long interconnection queues among the causes. The report says the AI collaboration is in addition to PJM’s own effort to automate interconnection planning. [2]

Earlier federal work. In September 2018 DOE announced a $5.8 million funding opportunity for research on using big data, AI and machine learning on sensor data from the grid. It says phasor measurement units, which are GPS-timed sensors, had been deployed at over 2,500 locations in the U.S. bulk power system. [4]

How far it has come. Based on the sources opened, this is early deployment and planned work. The PJM collaboration is described as an announced agreement and its aim; the report does not state a result. The Genesis targets are goals. This Brief did not find an independent measurement of AI improving grid operation.

Opportunities

The possible gain is speed in the steps that gate new supply. Tapestry states that it “takes too long to connect new sources of energy”, and puts the figure at 6 years on its own page. [3] The trade report cites forecast load growth driven by data centers, and lists long interconnection queues among the factors behind a jump in capacity prices. [2] If AI tools shorten study times, the effect would show in queue records, which are public. DOE’s call also speaks of developing “validated, deployable AI solutions for the grid” with grid operators, and requires a demonstration component of at least 20 percent in its grid-modeling focus area. [1]

Challenges

  • Engineering and integration. DOE states that technology suppliers “may lack access to robust operational technology test environments” to integrate and validate their systems. [1]
  • Institutions. The same document states that the utility sector has critical grid data but “a low risk tolerance, limited ability to develop new technology, and regional focus”. [1] PJM’s collaboration is one operator in one region. [2]
  • Regulation. Interconnection is governed by rules and reviews that software does not change by itself; the sources opened do not describe how AI-produced studies would be accepted. [2]
  • Evidence. No source opened measures a gain from AI. The Genesis targets are the agency’s goals for funded work. [1]

Data sources

  • DOE funding opportunities and award lists, such as the Genesis Mission opportunity, for named programs and awards. [1]
  • USAspending.gov for federal grants and cooperative agreements. [5]
  • PJM’s published service-request pages for interconnection queue status. They are a starting point for queue status. [6]
  • FERC and other operators’ interconnection queues for queue times. The FERC site refused scripted access in this run and was not opened. [7]
  • Operator and company announcements, as company claims. No registry records AI use in grid operations.

Changelog

  • 2026-11-18: First published.

Sources

  1. US Department of Energy, “The Genesis Mission: Transforming Science and Energy with AI”, Notice of Request for Application DE-FOA-0003612, issued 2026-03-17, Amendment 000003 of 2026-04-20; challenge 16. science.osti.gov (accessed 2026-11-18)
  2. “PJM-Google AI partnership aims to speed up generator interconnection”, POWER Engineering, 2025-04-10 (trade publication reporting a PJM announcement). power-eng.com (accessed 2026-11-18)
  3. X, “Tapestry” (company page, company claims). x.company (accessed 2026-11-18)
  4. US Department of Energy, “Department of Energy Announces Investment to Improve Resilience and Reliability of the Nation’s Energy Infrastructure”, 2018-09-25. energy.gov (accessed 2026-11-18)
  5. USAspending.gov, home page. usaspending.gov (accessed 2026-11-18)
  6. PJM Interconnection, “Service Requests”. pjm.com (accessed 2026-11-18)
  7. Federal Energy Regulatory Commission, home page. ferc.gov (listed, not opened: HTTP 403 to a scripted request on 2026-11-18)