The AI ​​revolution has a power problem

In the race for artificial intelligence dominance, America’s tech giants have the money and the brands, but their ambitions have hit a new roadblock: electricity.

“The biggest problem we’re facing right now is not over-consumption of computing, it’s power and the ability to do builds fast enough near power,” Microsoft CEO Satya Nadella admitted in a recent podcast with OpenAI head Sam Altman.

“If you can’t do that, you might actually have a bunch of chips in stock that I can’t connect,” Nadella added. Echoing the 1990s dotcom frenzy to build internet infrastructure, today’s tech giants are spending unprecedented sums to build the silicon backbone of the AI ​​revolution.

Google, Microsoft, AWS (Amazon) and Meta (Facebook) are drawing on their massive cash hoards to spend around $400 billion in 2025 and even more in 2026 — for now with the backing of enthusiastic investors.

All that cash has helped mitigate an initial phenomenon: acquiring the millions of chips needed to race for computing power, and the tech giants are ramping up their in-house processor production as they seek to catch up with world leader Nvidia.

These will go into the racks that fill the massive data centers — which also use huge amounts of water for cooling.

The construction of the massive information warehouses takes an average of two years in the United States. It takes five to 10 years for new high-voltage power lines to become operational. The “hyperscalers,” as the big tech companies in Silicon Valley are called, saw the energy wall coming. A year ago, Virginia’s major utility provider, Dominion Energy, already had a data center order book of 40 gigawatts — equivalent to the output of 40 nuclear reactors. The capacity it has to deploy in Virginia, the world’s largest cloud computing hub, has since grown to 47 gigawatts, the company recently announced.

Data centers in the United States, already blamed for swelling household electricity bills, could account for 7 to 12 percent of national consumption by 2030, up from 4 percent today, according to various studies. However, some experts say the predictions may be exaggerated.

“Both utilities and technology companies have an incentive to embrace the forecast of rapid growth in electricity use,” Jonathan Koomey, a renowned UC Berkeley expert, warned in September.

As with the Internet bubble of the late 1990s, “many data centers that are talked about and proposed and in some cases even announced will never be built.”

If projected growth materializes, it could create a shortfall of 45 gigawatts by 2028 — equivalent to the consumption of 33 million American households, according to Morgan Stanley.

Several US utilities have already delayed closing coal plants, despite the fact that coal is the most climate-polluting energy source. And natural gas, which powers 40 percent of the world’s data centers, according to the International Energy Agency, is seeing renewed favor because it can be used quickly.

In the US state of Georgia, where data centers are proliferating, a utility has sought authorization to install 10 gigawatts of gas-powered generators. Some providers, including Elon Musk’s startup xAI, have rushed to buy used turbines from abroad to quickly build capacity.

Even recycling aircraft turbines, an old niche solution, is gaining ground. “The real existential threat right now is not a degree of climate change. It’s the fact that we could lose the AI ​​arms race if we don’t have enough power,” Interior Secretary Doug Burgum argued in October.

Tech giants are quietly downgrading their climate commitments. Google, for example, promised net zero carbon emissions by 2030, but removed that commitment from its website in June. Instead, companies promote long-term projects.

Amazon is championing a nuclear revival through Small Modular Reactors (SMRs), a still-experimental technology that would be easier to build than conventional reactors. Google plans to restart a reactor in Iowa in 2029.