Acentric Research is produced with AI tools under human review. It is affiliated with Acentric Capital, a private investment group whose members may hold positions in companies discussed. 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
Artificial intelligence (AI) and energy meet where software that learns from data changes how power is planned, produced or licensed. Electricity demand is rising from data centers, manufacturing and electrification, and the U.S. Department of Energy (DOE) states that the grid faces reliability challenges as it struggles to accommodate it. [1] AI is being tried as a way to make decisions faster, to run existing equipment better and to shorten the paperwork that holds up new plants.
What stands out is the spread in maturity. Reports describe AI that forecasts solar output or runs building heating as already in use, though the figures come from the organizations involved. [2][3] AI that drafts nuclear licence documents has been shown on one test case. [4] AI for fusion plasma control is described as research and simulation. [5] Much of the public record is what agencies and companies say they aim to do.
Current state
What AI does for energy. In power grids, AI is applied to planning, connecting new generation, operating the system and security; DOE’s Genesis Mission funding call states the aim of “20-100x faster decision-making” and at least a 10 percent improvement in electricity cost and reliability. [1] In renewables and buildings, a UK report describes AI-enabled solar forecasting that helps the national system operator save £30 million a year in balancing costs, and AI control of heating and cooling in large buildings. [2][3] In fusion, Google DeepMind describes AI that simulates and controls plasma, the hot ionized gas inside a fusion machine. [5] In fission, DOE reports an AI tool that drafted sections of a reactor licence application in one day. [4]
What energy does for AI. In the sources opened, energy mainly supplies the setting for AI work, which is real equipment with real operating limits, and public funding. DOE’s Genesis Mission opportunity names grid, fusion and nuclear energy as areas for AI work. [1] The reverse link, AI’s own demand for electricity, is a different subject and is not covered here (see the exclusions below).
Four Junctions.
- AI for power-grid operations and forecasting. AI used by grid operators and in federal programs to plan, operate and connect to the grid. One regional operator, PJM, announced a multiyear collaboration with Google and Tapestry on connecting new generation. See the Junction Brief. [6]
- AI for renewable generation and building efficiency. Forecasting and control tools described as already in use, with reported savings that come from the sellers and reporting organizations. See the Junction Brief.
- AI for fusion plasma control and design. A research partnership between Google DeepMind and Commonwealth Fusion Systems, which is building a tokamak fusion machine called SPARC. See the Junction Brief. [5]
- AI for fission reactor design, operation and licensing. A DOE test case on licence drafting, a laboratory agreement on AI for licensing, and a news report of an NRC official’s remarks on faster reviews. See the Junction Brief. [4][7][8]
What was left out. Two areas were considered and excluded. The energy used by AI hardware is a cost of the AI, not a gain in an energy system. AI-driven growth in electricity demand creates demand for power but does not change energy technology. A third area, AI for catalyst design, is flagged as a possible three-realm topic (AI, materials and energy) and is not covered here. [9]
Opportunities
Public funders have named the area as a priority. DOE’s Genesis Mission opportunity, issued in March 2026, sets stated goals for nuclear (at least a two-fold faster schedule and more than 50 percent lower operating cost), for the grid (20-100 times faster decisions) and for fusion (a digital platform that links plasma, materials and plant models). [1] These are goals set by the agency, not results. Where AI is already in use, reports cite gains. The UK report states that existing wind farms optimised by AI “could deliver 3-5% more energy output”, which is an estimate, not a measurement. [2] In a DOE demonstration on an open-source test case, AI produced a first-draft licence document in one day; DOE says a team typically needs four to six weeks. [4][10]
Challenges
- Engineering and evidence. The licence demonstration used preliminary, open-source documents and the output was reviewed by an expert as a first draft. [10] In fusion, DeepMind says it expects to validate and calibrate its simulator against past tokamak data. [5]
- Regulation. An official of the U.S. Nuclear Regulatory Commission (NRC) said a licensing review that once took four years had been done in nine months. The news report adds that this was not strictly due to AI, because a May 2025 executive order set an 18-month limit. [8]
- Physics. Fusion requires holding plasma at over 100 million degrees Celsius within a machine’s limits, which DeepMind calls a highly complex physics problem. [5]
- Capital and scale-up. DOE’s own call states that utilities have critical grid data but a low tolerance for risk, and that technology suppliers may lack test environments to validate systems. [1]
Data sources
- DOE funding opportunities and award lists, including the Genesis Mission opportunity, for federal programs. [1]
- USAspending.gov for federal grants and contracts. [11]
- NRC dockets and ADAMS for licensing records. The NRC site refused scripted access in this run and was not opened. [12]
- FERC and grid-operator interconnection queues, such as PJM’s service-request pages, for the status of connection requests. The FERC site refused scripted access and was not opened. [13][14]
- Company and laboratory announcements, labeled as company claims. No registry records AI performance in fusion or in renewable forecasting in the sources opened.
Changelog
- 2026-11-18: First published.
Sources
- 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; challenges 4, 5 and 16. science.osti.gov (accessed 2026-11-18)
- Energy Systems Catapult, “New report tracks AI’s role in delivering decarbonisation”, 2026-01-13, summarizing the ADViCE report “State of AI for Decarbonisation 2025”, funded by the UK Department for Energy Security and Net Zero. es.catapult.org.uk (accessed 2026-11-18)
- Rohini Srivastava, American Council for an Energy-Efficient Economy, “The Use of Artificial Intelligence in Building Energy Management Control Systems”, topic brief, November 2025. aceee.org (accessed 2026-11-18)
- US Department of Energy, Office of Nuclear Energy, “Department of Energy Unleashes AI to Reduce Reactor Licensing Timelines”, 2026-03-26. energy.gov (accessed 2026-11-18)
- Google DeepMind, “Bringing AI to the next generation of fusion energy”, 2025-10-16 (company statement). deepmind.google (accessed 2026-11-18)
- “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)
- Oak Ridge National Laboratory, “Oak Ridge National Laboratory, Atomic Canyon to accelerate nuclear licensing with AI”, 2025-07-22. ornl.gov (accessed 2026-11-18)
- Edward Graham, “AI has helped slash nuclear licensing review times, NRC official says”, Nextgov/FCW, 2026-06-25 (reports remarks by an NRC official). nextgov.com (accessed 2026-11-18)
- Acentric Research, “Landscape Report, November 2026”, exclusions and three-realm candidates. acentricresearch.com (accessed 2026-11-18)
- “The DOE’s plan for AI in NRC licensing”, Nuclear News (American Nuclear Society), 2026-04-02. ans.org(accessed 2026-11-18)
- USAspending.gov, home page. usaspending.gov (accessed 2026-11-18)
- US Nuclear Regulatory Commission, ADAMS. nrc.gov (listed, not opened: HTTP 403 to a scripted request on 2026-11-18)
- Federal Energy Regulatory Commission, home page. ferc.gov (listed, not opened: HTTP 403 to a scripted request on 2026-11-18)
- PJM Interconnection, “Service Requests”. pjm.com (accessed 2026-11-18)