US climate change goals threatened by surge in AI-generated technological energy

Power-hungry artificial intelligence is consuming increasing amounts of energy from the creaking U.S. grid and threatening national efforts to combat climate change, according to the latest expert forecasts.

Unprecedented energy demand, driven in part by the expansion of data centers for artificial intelligence, combined with slower-than-expected growth in renewable energy sources and longer operating schedules for polluting coal-fired power plants, has prompted analysts to reshape their models to reduce greenhouse gas emissions.

The topic dominated discussions at New York Climate Week, held on the sidelines of the UN General Assembly last week, with a greater focus on technology companies than on historically polluting fossil fuel companies.

BloombergNEF’s latest report this week warned of slower U.S. progress in decarbonization, predicting that emissions would be reduced by just 34 percent by 2030 compared to 2005 levels.

The latest assessment pushes the U.S. trajectory even further away from its national goal of reducing emissions by 50-52 percent by 2030 compared to 2005 levels and achieving net-zero emissions by 2050, as pledged in the agreement Parisian.

“It’s not good in the long run,” said Tara Narayanan, chief energy analyst at BloombergNEF, calling the increase in demand for AI power a “big disruption” to supply.

“It’s very much like that moment when you’re deep in a movie and three different story lines emerge. It is unclear whether the problem will be resolved,” Narayanan said.

The lack of grid infrastructure is a major obstacle to progress in the green energy transition not only in the US but around the world.

China will commit an unprecedented $800 billion to spend the next six years trying to overcome pressures on its energy system as it rapidly shifts from coal-fired power to renewables.

In the US, energy demand has remained virtually constant for two decades. Currently, forecasters such as consulting group ICF expect it to grow 9% by 2028 and almost 20% by 2033, citing data center expansion, manufacturing relocation and electrification.

The Electric Power Research Institute predicted this year that data centers could double their share of U.S. electricity consumption by the end of the decade.

However, U.S. Energy Secretary Jennifer Granholm said she believes the country will still be able to meet its net zero goals and cope with the explosion in energy demand thanks to nearly $370 billion in environmental subsidies introduced in the Inflation Reduction Act by Joe Biden’s administration.

“We have to be aggressive, but the momentum has built and is not slowing down,” Granholm told the Financial Times.

Renewable energy project developers say generating green energy to meet historic levels of demand is hampered by the fact that it can take up to half a decade to bring new supplies onto the grid due to delays in permitting and commissioning.

“The need of the hour is to balance this,” said Sandhya Ganapathy, CEO of EDP Renewables North America. “Unfortunately, we may not implement (renewable) projects at the required pace.”

The proliferation of AI-powered data centers has led to a race among Big Tech companies to find sources of low-carbon, 24/7 power.

Last week, Constellation Energy and Microsoft signed a 20-year agreement to reopen the Three Mile Island nuclear power plant in Pennsylvania, the site of the nation’s worst nuclear accident.

Expectations of higher electricity demand have also caused U.S. operators to delay retiring coal-fired power plants. S&P Global Commodity Insights revised its expectations for coal-fired power plant closures by the end of the decade by 40 percent, despite a rise in renewable energy.

“As things currently stand, it is very difficult to imagine that the U.S. power system will be emission-free by 2035.” – said Akshat Kasliwal, energy expert at PA Consulting. “We are further away from that goal compared to where we were, say, three or four years ago.”

Pedro Pizarro, chief executive of utility company Edison International, said the surge in demand means gas-fired power plants will also have to stay in the energy mix longer to ensure reliable supply. The gas consists mainly of the highly heating molecule methane, which retains more heat than carbon dioxide, but has a shorter lifespan in the atmosphere.

“We are not a gas company. . . We don’t have a dog that’s hunting, trying to stop the gas,” Pizarro said. “We as the industry need to make sure we have a system that is reliable and resilient, given the greater number of extreme weather events, and at an affordable price.”

However, there is no shortage of renewable energy capacity in the US. Lawrence Berkeley National Laboratory estimates that nearly 1.5 terawatts of generating capacity awaits connection to the grid, enough for more than twice the size of the national power system.

But projects built last year had a five-year wait before they could connect to the grid, and a shortage of transmission lines makes it difficult to transport green energy from distant generation sites to centers of demand.

Research firm Rhodium Group found that if demand for data centers nearly triples by 2035 and developers struggle to install new wind and solar farms, emissions from the energy sector could be more than 56 percent higher than forecast in moderate emission forecasts.

But the high forecasts could also become much less optimistic as data centers become more efficient, technology group executives argue, and the wider use of artificial intelligence reduces energy use by streamlining daily operations.

“Although training models consumes energy, the models created will do the work much more energy efficiently,” Jensen Huang, chief executive of Nvidia, the fastest-growing AI chipmaker, told the Biptial Policy Center on Friday. “Energy efficiency and the productivity gains we will get from (AI). . . It will be unbelievable.”