AI vendors on the final list gain state-backed deployments at grid, power and coal operators, and a route into national AI demonstration programs and funding.
China’s National Energy Administration has shortlisted 93 projects for its "AI+" energy pilot program and opened them to public scrutiny. A notice from the NEA’s General Affairs Department (综合司), dated November 3 and carried in full by the 21st Century Business Herald, says the projects passed expert review and a second check before being placed on the draft AI+ energy pilot project list (“人工智能+”能源试点项目清单). The only project the notice names is one for intelligent decision-making in the maintenance of main substation equipment (变电主设备检修智能决策).
The shortlist is the step before formal pilot status. The public notice period runs from November 5 to November 11, and any organization or individual can file a written objection in that window, signed with a real name and contact details, to an NEA email address. Anonymous objections, objections without specific facts and improper demands in the objector’s own interest will not be accepted, according to the notice, which Jiemian News also carried.
The list itself sits in image attachments, so the names of the energy companies and AI suppliers behind most projects are not yet visible in text. Two reposts on Sina add detail. One Sina repost, of a report by the trade journal Coal Industry, says six coal projects made the list. Another Sina repost, from the solar trade site Solarbe, names a project for the intelligent operation of the Yalong River basin’s 10-gigawatt-scale hydro, wind and solar base. Neither detail has been confirmed against the attachment.
The shortlist ends a year-long selection built around 51 scenarios
The selection began in November 2025, when the NEA asked provinces and central energy enterprises to propose high-value uses of AI. That notice, published by China Energy News Network, asked for scenarios that were "few but refined," tied to long-standing problems in the industry and measured by quantifiable targets for cost, carbon or safety. It excluded uses that are already mature, citing knowledge question-and-answer tools as an example, and capped each province and each central enterprise at five proposals.
In May the NEA published the result: the first AI+ energy high-value scenario list (“人工智能+”能源高价值场景清单), with 51 scenarios such as the intelligent generation and assessment of grid planning options. The NEA’s notice of that list, document No. 50 of 2026, invited energy companies with a listed scenario to team up with AI technology suppliers and submit pilot construction plans. Each recommending body, a provincial energy authority or a central energy group, could put forward up to 10 projects through an online filing system, and had until July 30 to submit its stamped list of recommendations. Each submission had to name its scenario exactly as on the list, and applications were barred from including state or commercial secrets.
Xinhua reported that the 51 scenarios were released at a national meeting on AI+ energy, where 25 energy companies signed an initiative. The scenarios fall into eight application groups, including grid planning and dispatch, virtual power plants, vehicle-to-grid links and renewable power forecasting.
The program ties energy data to AI suppliers through an open-bidding model
The pilot program runs on "揭榜挂帅," a mechanism in which the NEA publishes the problems energy companies want solved and technology suppliers compete to take them on. Applicant energy companies must commit to opening their scenarios and related data to the suppliers they team up with, and in return get a route into national programs. The November 2025 notice said the NEA would support recommending qualifying pilots for the national AI demonstration program, that they should make use of state funding and policy support, and that strong performers would get priority when applying for national demonstration tasks and major science and technology projects.
The NEA also said results should be turned into national or industry standards, so that other energy companies avoid paying to build the same tools again. Pilots that fall behind can lose their status under a system of dynamic monitoring and quantified assessment.
Energy demand from computing is part of the case for the program. Bian Guangqi, deputy director of the NEA’s energy conservation and science and technology equipment department, said in May that computing facilities were moving from the 100-megawatt scale to the gigawatt scale. China’s computing centers used 170 billion kilowatt-hours of electricity in 2025, according to figures cited by Xinhua. The same report said dozens of specialized large models were already in use in the energy industry.
Final approval and the names behind the projects come next
After the objection window closes on November 11, the NEA can confirm the list, and projects without upheld objections become pilots. The published final list, with readable project owners and technology partners, will show which AI developers, cloud providers and grid software vendors have won places in state-backed deployments. How the 93 projects divide between grids, power generation, renewables and coal will show where the NEA expects AI to pay off first, and the size of any national funding attached to them is the figure to watch.