AI demand and infrastructure requirements surged in 2025. At Davos, the narrative for AI in 2026 continues unchanged.

AI Expansion Drives Unprecedented Energy and Infrastructure Growth

The rapid advancement of artificial intelligence has led to a significant increase in energy consumption and infrastructure projects throughout 2025. At the World Economic Forum in Davos, global leaders anticipate that these trends will remain central to the AI narrative as 2026 approaches.

During his address in Davos, President Trump highlighted the immense energy requirements of AI, stating, "We needed more than twice the current national energy supply just to support AI facilities, and that's simply not feasible." He reiterated the United States' dedication to strengthening its domestic power infrastructure.

Trump emphasized that his administration is focused on turning infrastructure development into a thriving sector to fuel America's AI ambitions. Rather than shutting down energy plants, he declared, "we're opening them up."

According to Goldman Sachs, global data center electricity usage is projected to climb from approximately 55 gigawatts to 84 gigawatts within the next two years.

However, companies constructing the data centers that power AI are facing significant delays, with some components like natural gas turbines backlogged for years. Connecting new-generation equipment to the US grid can take over a decade.

Olivier Blum, CEO of Schneider Electric (SU.PA), described this as the core dilemma of today's economy: "AI is a major driver of digital growth, but it also consumes vast amounts of one of the world's most critical resources—energy."

Against this backdrop, the Industrials sector (XLI) has outperformed, gaining 17.5% over the past year compared to the S&P 500's (^GSPC) 13.5% return. Caterpillar (CAT), the leading US heavy machinery company, saw a remarkable 58% return in the same period.

The Most Ambitious Infrastructure Expansion Ever

President Trump asserted that both infrastructure and energy output are on the rise in the United States. He noted that steel production is "doubling and tripling," with new steel plants emerging nationwide. Factory construction has surged by 41%, and US natural gas output has reached record highs.

These domestic investments reflect the relentless need for new infrastructure to support AI's rapid expansion, as echoed by many at the Davos summit.

In a discussion with BlackRock (BLK) CEO Larry Fink, Nvidia (NVDA) CEO Jensen Huang remarked that, despite the hundreds of billions already invested by major tech firms, AI's growth will demand "trillions of dollars" for what he described as the “largest infrastructure build-out in history.”

Image: Nvidia and BlackRock Leaders at Davos

Strategists at BlackRock echoed this sentiment in a recent report, emphasizing the urgent need for expanded energy and infrastructure. They wrote, "Infrastructure bridges the gap between economic goals and practical capabilities. For AI, reliable and abundant power remains a critical challenge."

Yet, expanding the nation's power generation and infrastructure is a lengthy process, and US investment in the power grid has lagged. According to Bank of America, 31% of transmission equipment and 46% of distribution equipment in the US are within five years of retirement or already obsolete. The US averaged 1,700 miles of new transmission lines annually in the early 2010s, but that figure dropped to just 645 miles per year in the latter half of the decade.

With grid connections taking years, both corporations and government officials are exploring alternatives. Major tech firms are investing in on-site "behind-the-meter" power solutions, while leading oil and gas companies are repurposing natural gas infrastructure to supply electricity to Texas data centers. Apple (AAPL) is considering battery and solar projects.

A key topic at Davos, and for President Trump, has been what some call the "nuclear renaissance."

Trump stated, "We're making a major push into nuclear. I was once skeptical due to the risks, but the advancements in nuclear technology are astonishing."

He added, "We're fully embracing nuclear energy—it's now affordable, extremely safe, and we're leading the world in AI, far ahead of China."

Trump highlighted executive orders signed in May 2025 aimed at adding 300 gigawatts of new nuclear capacity by 2030, boosting domestic nuclear fuel production, and streamlining reactor approval processes.

Bank of America analysts noted that nuclear energy has seen renewed interest, driven by AI, electrification, industrial growth, and electric vehicles. The bank forecasts 18 gigawatts of new nuclear capacity annually between 2025 and 2040, with nuclear poised to become a $10 trillion industry.

Image: Cooling Towers at Vogtle Nuclear Plant

Infrastructure Expansion Faces Financial and Capacity Challenges

Such a massive infrastructure effort comes with a hefty price tag. As Jensen Huang pointed out, hyperscale companies have already invested hundreds of billions in the early stages of the AI surge. However, this spending coincides with rising utility bills for US consumers as data centers consume more electricity.

Residential electricity costs have increased by about 13% since January 2025, according to the Energy Information Agency. The average utility bill for electricity and gas rose 3.6% year-over-year in the third quarter.

On Friday, the Department of the Interior announced a plan to fund new power generation projects and require tech companies to contribute more to these costs, as President Trump has advocated.

A major initiative involves President Trump and several governors urging PJM, the utility serving over 67 million people in the mid-Atlantic, to hold an emergency auction where large data center operators can secure 15-year contracts for energy, encouraging investment in new power generation.

At Davos, Microsoft (MSFT) CEO Satya Nadella remarked that energy costs and infrastructure will determine the leaders in the AI sector. "Are you able to produce energy cheaply? Can you build data centers? What is the cost of silicon in your systems?" he asked.

Nadella added, "Tokens are the new commodity. Every economy and business must convert these tokens into economic growth."

When Amazon (AMZN), Google (GOOG), Meta (META), and Microsoft release their fourth-quarter results later this month, they are expected to significantly raise their AI capital expenditure forecasts after previous increases last quarter.

Nonetheless, HSBC strategists warn that these companies will face delays due to limited cloud capacity, which will take time to resolve.

Nicolas Cote-Colisson, global head of tech platforms at HSBC, wrote, "The three largest Western cloud providers (Amazon, Microsoft, and Alphabet) are all experiencing capacity limitations amid strong demand and order backlogs."

"Given the lengthy timelines for infrastructure projects, we do not anticipate this situation changing in 2026."