US AI Action Plan: Deregulation and fossil fuels

BIANCA GIACOBONE | The Trump administration released its AI Action Plan last week. Described as a blueprint to “win the AI race,” it marks the most articulated expression of its AI policy so far. 

On the same day, President Trump signed three related executive orders, focusing on the AI policy priorities outlined in the plan: promoting the export of the American AI technology stack; accelerating federal permitting of data center infrastructure; and preventing the federal government’s use of so-called “woke” AI. 

For energy and data centers, the plan is an ode to deregulation and a confirmation of the administration’s preferential treatment of fossil fuels over renewables.

“Build, Baby, Build!” it reads, vowing to reject “radical climate dogma and bureaucratic red tape” and streamline the construction of qualifying projects. These include big data centers with over 100 megawatts of load and $500 million in committed capital expenditures; supporting energy infrastructure, such as transmission lines and gas pipelines; and gas, coal, nuclear, geothermal, or other “dispatchable baseload energy sources” supporting a data center. Note the missing renewables, even combined with storage.

To achieve this goal, the plan, as outlined in one of the executive orders, is to sidestep the National Environmental Policy Act’s environmental impact studies by creating “categorical exclusions” for the qualifying projects. (The Trump administration, with help from the courts and the congressional GOP, is working to unravel NEPA more broadly as well.) It also aims to expand the FAST-41 program, which streamlines federal permitting processes for large infrastructure projects, to include data centers and related energy infrastructure. 

And it’s making federal land available for AI infrastructure. Just a few hours after the plan was released, the Department of Energy announced it had selected its first federal sites: at Idaho National Laboratory, Oak Ridge Reservation in Tennessee, Paducah Gaseous Diffusion Plant in Kentucky, and Savannah River Site, in Georgia. 

The plan does acknowledge the urgent need to upgrade and stabilize the grid, but — in an echo of the Department of Energy’s reliability study issued earlier this month — it suggests doing so by delaying the decommissioning of power plants and embracing new technologies like geothermal, nuclear fission, and nuclear fusion. Again, no mention of renewables, even as it calls for prioritizing “the interconnection of reliable, dispatchable power sources as quickly as possible.” 

Meanwhile, a directive circulated earlier this month requires all wind and solar projects on federal lands to be reviewed by Interior Secretary Doug Burgum. This is widely seen as yet another attempt to hamstring the clean energy industry, rather than leverage the resources to meet load growth. 

This attitude towards clean energy puts the U.S. AI strategy in direct contrast with China’s. 

Just a couple of days after the Trump administration’s release of its action plan, China came forward with its own. This weekend, at the opening ceremony for the World Artificial Intelligence Conference in Shanghai, Premier Li Qiang unveiled an “Action Plan on Global Governance of Artificial Intelligence.” The plan is much less detailed than the U.S. one, but it stresses a commitment to “sustainable AI” and to responding to “energy and environmental issues.” It comes two weeks after China mandated new data centers in “national hub nodes” to use at least 80% green electricity.

This approach plays into China’s strength; the country has become a renewables stronghold. But it also plays a role in China’s plan to become a global leader of both AI and the energy transition.. The plan unveiled this weekend calls for establishing a global AI cooperation organization in collaboration with the United Nations, to promote international cooperation on tech development and regulation. 

Matt Pearl, director of the Center for Strategic and International Studies’ strategic technologies program, said China is actively looking to contrast with the strategic posture of the U.S.; it’s a strategy based on collaboration, versus one based on dominating via exporting U.S. technologies.

“The U.S. plan continues with an America-first, America-led approach, and I think the reason the People's Republic of China put out this document was to emphasize that they're taking a multilateral, global public goods approach to things, with UN collaboration,” he said, adding that it’s unclear how much of this is true, and how much of it is just messaging. 

But China’s embrace of clean energy for AI is certainly a part of their desire to contrast with the U.S., he added. “I think the hope is that they can take advantage of tensions between the U.S. and the European Union, for which green energy is a priority, for instance,” Pearl said. “They’ll see if they can use that as a policy affinity that they could take advantage of to bring them in closer cooperation and alignment.” 

The question from here is whether  China is actually willing to “walk the walk” in terms of establishing global, multilateral cooperation for AI. China’s “willingness to learn from the rest of the world, including especially the U.S.,” both in terms of technologies and policies, can be a competitive advantage, Pearl said, as it was when the country was industrializing. And if so, China could pull ahead in the technology race.

But a multilateral approach could benefit the U.S. as well, even if there are no indications that the current administration would be willing to entertain one. “Every country is doing things that are in its self-interest,” Pearl said. “But seeing whether multilateral institutions and processes provide opportunities, and being flexible about them, means that you have the ability to maximize U.S. influence, depending on how things develop.”

FEATURED STORY

Under the hood of CAISO’s AI outage management pilot

MAEVE ALLSUP | The California Independent Systems Operator processes hundreds of outages daily — a painstaking task only made more challenging by the sheer complexity and interconnectedness of California’s grid and a rise in extreme weather events and distributed assets.

Managing planned and unplanned interruptions to parts of the grid is a core responsibility of independent system operators. But it’s a continuous, highly manual task distributed among hundreds of employees every single day. Outage requests submitted by transmission owners, generation companies, and utilities looking to take infrastructure out of service — whether for maintenance, upgrades, or unexpected faults — must be assessed for their impact on system reliability, market operations, and customer service.

CAISO operators in departments around the organization (transmission, generation, reliability) read those reports, track down relevant historical data and operational information from multiple sources, input relevant metrics into CAISO’s grid software system, and run calculations to evaluate the impact.

But the organization is moving slowly away from doing everything by hand. For the last several months, CAISO has been piloting a generative AI platform that it hopes will streamline the process and help it meet the compounding demands of outage coordination.

MORE NEWS

Latitude Media | What will Google’s $3 billion deal mean for aging hydropower plants? — Google’s deal with Brookfield, the world’s largest clean energy deal for hydropower, could establish the first major lifeline to an industry facing many of the same problems that plagued nuclear power for much of the past decade.

The Washington Post | The AI explosion means millions are paying more for electricity — “In Columbus, for example, households on the standard plan of the local utility began paying about $20 more a month as of June — or $240 a year — because of the demand from data centers.” 

*The Wall Street Journal | Who Pays? AI Boom Sparks Fight Over Soaring Power Costs — “Utilities are seeking assurance that tech companies will pay for any potential surplus if the AI boom spurs the development of more power lines and plants than U.S. data centers ultimately need. The tech companies, meanwhile, say… some of those upgrades also benefit other customers.” 

The Seattle Times | Cheyenne to host massive AI data center using more electricity than all Wyoming homes combined — “The data center, a joint effort between regional energy infrastructure company Tallgrass and AI data center developer Crusoe, would begin at 1.8 gigawatts of electricity and be scalable to 10 gigawatts.” 

The Hill | Tech companies building massive AI data centers should pay to power them — In this opinion article, David Klaus, former deputy undersecretary of the U.S. Department of Energy during the Obama administration, and author Mark MacCarthy list the objectives that they think should be the core of public policy when it comes to data centers’ energy use. 

Facilities Dive | Nuclear specialist Oklo inks data center partnerships — “It plans to co-develop advanced power and thermal management solutions with Vertiv…and deploy Liberty Energy’s natural gas generators and electric load management platform to power Oklo’s future data center customers.” 

INL Media Relations | Idaho National Laboratory accelerates nuclear energy projects with Amazon Web Services cloud and AI technologies — “The laboratory is developing a suite of technologies that use AI to reduce the costs and timeframes of designing, licensing, building and operating nuclear facilities. Ultimately, the tools could be used for safe and reliable autonomous operation of nuclear reactors and accelerating deployment of new advanced reactors.” 

The Deep View | Your ChatGPT follow-ups are destroying the climate — “A ChatGPT query consumes about five times more electricity than a simple web search, while one query uses approximately as much electricity as could light one light bulb for about 20 minutes. A University of Cambridge report warns AI growth could raise tech sector energy use 25-fold by 2040.” 

*Subscription required

CHART OF THE WEEK

This chart by financial research firm Quartr offers a snapshot of the many companies providing power distribution, security, and cooling solutions for data centers.

RESOURCES + EVENTS

• On August 7, 2025, we’re hosting a webinar on how onsite generation is reshaping data center strategy alongside Bloom Energy, which will unveil key findings from its 2025 Data Center Power Report: Mid-Year Pulse. You can sign up here. 

• On August 21, 2025, Latitude Media’s Maeve Allsup will speak with Ken Rahn at FlexGen and Keaton Horner at Orennia about how utility-scale BESS can accelerate interconnection, improve resilience, and manage AI workload power volatility. You can sign up here.

• On September 10, 2025, Latitude Media’s Scott Clavenna will participate in a cross-sector panel discussing the energy impact of AI deployments. The panel is part of Yotta 2025, a digital infrastructure event that will take place in Las Vegas. 

• The Smart Electric Power Alliance and NC Clean Energy Technology Center released a Database of Emerging Large-Load Tariffs (DELTa), which includes data on utility tariffs targeting large loads and data centers and other large loads.

• At New York City’s upcoming Climate Week, Economist Impact is hosting an event on AI’s reshaping of the energy sector, and on its growing energy footprint. It will take place on September 24, and you can apply to attend here. 

• This Friday, the Center on Global Energy Policy at Columbia University SIPA is hosting a a “panel discussion on the Trump administration’s recent executive actions on artificial intelligence and their implications for energy use in the United States and around the world.” You can register here.