The surge in gas plants driven by AI and the efforts to combat it.
The AI expansion has achieved something the fossil fuel sector failed to accomplish. It has initiated the largest construction boom ever seen in natural gas-powered plants, according to a report by the Associated Press.
Outdated coal power plants are also being kept operational beyond their intended retirement dates. Utilities, plant operators, and the federal government have all lobbied to delay these shutdowns.
The underlying issue is straightforward math. Some data centers use more electricity than a medium-sized city, and the development of wind and solar energy cannot keep pace with demand.
Several states are taking legislative action to address this. A bill currently with New York Governor Kathy Hochul would require large data centers to achieve renewable energy targets starting in 2030, reaching at least 90% renewable energy by 2040.
State Senator Kristen Gonzalez, the bill’s sponsor, believes the goals are attainable. She argued that the wealthiest companies in the world, which have the capacity to invest billions in data centers, can also invest in the energy infrastructure needed to power them.
Michigan, Oregon, and Minnesota were the first to act. In the past 18 months, all three states enacted laws to uphold their commitments to achieve emissions-free electricity by 2040. Michigan’s law links clean energy goals to financial incentives, requiring hyperscale data centers to attain 90% clean energy within six years in order to maintain a valuable sales tax exemption. Similar legislation has emerged in California, Illinois, New Jersey, Pennsylvania, and Virginia.
An honest acknowledgment from the opposition comes from Bob Jenks of the Oregon Citizens’ Utility Board, who admitted that meeting the 2040 target would be difficult with the presence of data centers and nearly impossible without them.
This outlines the challenge at hand. The clean energy destination was already ambitious, and the advent of AI has made it even more elusive.
Households are the first to feel the impact. Electricity rates are rising in many regions, with AI data centers contributing to increased energy costs for factories in the Rust Belt.
In response to the construction boom, advocates are targeting regulatory changes instead. The strategy is to persuade regulators to allow large power consumers to develop their own clean energy generation and connect it to the power grid.
Colorado has instructed Xcel Energy to implement such a program. In an April filing, Xcel acknowledged the potential benefits for customers, referencing Google’s initiatives that involve connecting 115 megawatts of geothermal energy in Nevada and 1,900 megawatts of wind, solar, and storage in Minnesota.
Google’s agreement with NV Energy is considered a pioneering model, and the company asserts that similar agreements are already approved or in discussion across eight additional states. The Corporate Energy Buyers Association has also struck a comparable deal with Georgia Power and is currently pursuing opportunities in North Carolina.
The appeal to utilities emphasizes commercial interests rather than moral obligations. They would secure a lucrative long-term customer that finances grid expansion, rather than allowing that customer to develop independent energy sources and withdraw from the grid.
The real crux of the matter lies in grid access, where crucial outcomes are established, not in legislative halls. Regulators have been expediting grid connections for data centers, and the entity that manages this queue dictates what gets constructed.
Financial investment is gravitating toward the same constraint, with startups backed by Nvidia raising funds to address power challenges for data centers. Energy, rather than silicon, is now the primary limitation affecting AI development.
Communities are independently opposing these projects, having successfully halted 75 data center initiatives valued at $130 billion in just one quarter. Congress is also weighing in, with the House voting on legislation aimed at shifting energy costs associated with data centers back to the companies responsible for them.
CEBA’s policy director believes that the decisions being made now will influence energy policy for the next two or three decades. This assessment is likely accurate, underscoring the need to pay more attention to the technical discussions surrounding grid interconnection.
As gas plants are being constructed, the regulatory framework is still being shaped, and typically, concrete projects prevail in these contests.
Other articles
The surge in gas plants driven by AI and the efforts to combat it.
AI data centers have triggered the most significant surge in gas plant construction to date. Advocates for clean energy are unable to match this pace, so they are targeting regulators instead.
