AI power and infrastructure needs boomed in 2025. At Davos, the AI story for 2026 remains the same.

The AI boom triggered a surge in energy demand and infrastructure development throughout 2025. Global leaders gathered at this week’s World Economic Forum in Davos expect these same themes to be a major part of the AI story in 2026.

“I said, you can’t create this much energy. We needed more than double the energy currently in the country just to take care of the AI plants, and I said we can’t do that,” President Trump said in comments during his speech in Davos Wednesday, during which he repeatedly emphasized the US’s commitment to powerful domestic electricity infrastructure.

The administration, Trump said, is intent on making infrastructure development in the US a booming industry to power the country’s AI ambitions.

“Instead of closing down energy plants, Trump said, “we’re opening them up.”

Global power usage by data centers is expected to grow from a current level of around 55 gigawatts to 84 gigawatts in only the next two years, according to research from Goldman Sachs.

Yet, the heavy machinery companies building the data centers that underpin AI’s computational power are facing years-long backlogs for parts such as natural gas turbines, and the process for getting new-generation equipment connected to the US power grid can take more than a decade.

“This is the fundamental paradox of the modern economy,” said Schneider Electric (SU.PA) CEO Olivier Blum in comments for the Davos summit. “AI is the digital engine of growth, but it is also a massive consumer of one of the world’s most in-demand resources — energy.”

Against this backdrop, the Industrials sector (XLI) has soared, picking up 17.5% over the past year against the S&P 500’s (^GSPC) return of 13.5%. Caterpillar (CAT), the largest heavy machinery operator in the US, has returned a staggering 58% over the same timeframe.

During his comments on Wednesday, President Trump insisted that infrastructure and energy production are on the rise in the US.

Steel production is “doubling and tripling, and we have steel plants being built all over the country,” the president said. Factory construction is “up by 41%, and that number is really going to skyrocket right now.” US natural gas production is “at an all-time high by far.”

The domestic investments are indicative of the insatiable demand for new infrastructure to feed AI’s growth, according to a chorus of voices at the Davos summit.

During a conversation with BlackRock (BLK) CEO Larry Fink during the summit, Nvidia (NVDA) CEO Jensen Huang said that — despite the hundreds of billions already spent by Big Tech companies — AI’s development will require “trillions of dollars” of spending on what the chipmaking leader called the “largest infrastructure build-out in history.”

In a report published Wednesday, strategists at Fink’s firm echoed the same point, noting the need for “pipes and power.”

“Infrastructure is linking economic ambition with real-world capacity,” BlackRock’s strategists wrote, “For AI infrastructure, power availability & reliability remains a key constraint.”

But building out a massive new wave of electricity generation and infrastructure capacity is not a quick task, and in the US, power grid funding and development have slumped.

Thirty-one percent of transmission equipment and 46% of distribution equipment in the US are within five years of the end of their useful life or have already passed that point, according to Bank of America. The US built an average of 1,700 miles of transmission infrastructure per year through the first half of the 2010s, but only 645 miles per year on average per year in the second half, according to Grid Strategies.

Given the years-long timelines for grid connection, corporate and government leaders have sought out alternatives. Big Tech companies have begun developing on-site “behind-the-meter” power, and the US’s largest oil and gas companies are signing deals that will see them repurpose natural gas infrastructure to provide electricity to data centers in Texas. Apple (AAPL) is looking at battery and solar farms.

Top of mind at Davos — and for President Trump — has been, to pull from the title of a panel during the conference, the “nuclear renaissance.”

“We’re going heavy into nuclear,” Trump said Wednesday. “I was not a big fan because I didn’t like the risk, the danger, but … the progress they’ve made with nuclear is unbelievable.”

“We’re very much into the world of nuclear energy, and we can have it now at good prices and very, very safe, and we’re leading the world in AI by a lot. We’re leading China by a lot.”

During his comments, Trump touted a suite of executive orders signed in May 2025 that aim to, among other goals, add 300 gigawatts of new nuclear capacity by 2030, increase domestic production of nuclear fuel, and speed up the regulatory process for reactor construction approvals.

“In many ways nuclear energy has been recently ‘rediscovered,’ ignited by the rise of AI/data centers, building electrification, industrial growth and electric vehicle adoption,” Bank of America analysts wrote in a note.

The bank is predicting that 18 gigawatts of new nuclear capacity will be built annually between 2025 and 2040, with nuclear energy primed to become a $10 trillion industry.

An infrastructure build-out of this scale, of course, doesn’t come cheap.

As Huang noted, hyperscalers have poured hundreds of billions of dollars into the early innings of the AI boom. But that spending comes at a moment when US rate-paying consumers face utility bills that are steadily ticking up as data centers eat up power supply.

Monthly costs for residential electric power have risen roughly 13% since January 2025, according to the Energy Information Agency, while the average utility payment for electricity and gas rose by 3.6% year over year in the third quarter.

On Friday, the Department of the Interior rolled out a plan to both fund power generation build-outs and put the onus on tech companies to “pay their own way,” as President Trump has put it.

Key to the efforts is a push from President Trump and a coalition of governors for PJM, the utility serving a mid-Atlantic region of more than 67 million people, to hold an emergency power action where large-scale data center operators will bid for 15-year contracts to lock in energy prices and thereby encourage those developers to build new power generation equipment.

In comments at Davos, Microsoft (MSFT) CEO Satya Nadella said that energy and energy infrastructure costs will be the key driver of who wins the AI race.

“Are you a cheap producer of energy? Can you build data centers?” Nadella said. “What’s the cost of the silicon in the systems?”

“We have got a new commodity, it is tokens. The job of every economy, and every firm in the economy, is to translate these tokens into economic growth.”

When Amazon (AMZN), Google (GOOG), Meta (META), and Microsoft — some of the largest “hyperscaler” spenders on AI development — report fourth quarter earnings later in January, all four are expected to substantially increase their AI capex estimates after previous increases disclosed in the third quarter.

But they will be slowed down by cloud capacity constraints that will take time to overcome, according to HSBC strategists.

“The three largest Western cloud players (Amazon, Microsoft, and Alphabet) all note capacity constraints, against a backdrop of strong demand and order backlogs,” Nicolas Cote-Colisson, the global head of tech platforms at HSBC Bank, wrote in a recent note to clients.

“Given long infrastructure project build times, we do not expect a change to this narrative in 2026.”

Jake Conley is a breaking news reporter covering US equities for Yahoo Finance. Follow him on X at @byjakeconley or email him at jake.conley@yahooinc.com.

Click here for in-depth analysis of the latest stock market news and events moving stock prices

Read the latest financial and business news from Yahoo Finance