The Numbers Don't Lie

Europe's energy-intensive industries paid roughly double what their U.S. counterparts paid for electricity last year. They paid 50% more than companies in China and India. This cost disadvantage reflects a structural problem that makes building AI infrastructure in Europe economically difficult.

Those figures come directly from the International Energy Agency.

Michael Brown, global investment strategist at Franklin Templeton, put it plainly to CNBC: "If I were making the next $7 billion data center, it would be in the U.S. or China." Full stop.

What the EU Is Actually Planning

The EU has an "AI Continent Action Plan" that aims to double data center capacity by 2030. The problem is the electricity math doesn't work.

A new policy brief by Matilde Ciani, researcher at the Kiel Institute, projects that data centers associated with that expansion could consume between 98.5 and 168 terawatt-hours of electricity by 2030. For scale — that's roughly equivalent to Poland's entire electricity demand in 2024, and up to 5% of total EU consumption.

"AI policy cannot be separated from energy policy," Ciani wrote. "Europe is planning ambitious digital infrastructure without ensuring that the electricity system can support it."

The Kiel brief warns of a three-way trap: Europe either burns more fossil fuels to meet demand, throttles economic growth to keep consumption flat, or falls behind in AI.

The Iran Crisis Made Everything Worse

The recent U.S.-Iran conflict sent energy prices surging again across Europe. Olivier Darmouni, associate professor at Columbia Business School specializing in energy transition, noted during a Tuesday briefing that the crisis triggered "renewed interest in electrifying the economy" — but the grid isn't ready for it.

Darmouni also flagged that rapid data center growth could inflate regional electricity costs by 20-40% in high-demand zones like Texas and Virginia in the U.S., or Slough and Paris in Europe. He called AI a "wake-up call" that forces treatment of the energy system as a matter of economic sovereignty.

"Affordability and inflation, competitiveness to the European companies and technological leadership in the form of AI — we can't get any of that if we don't fix the energy system," he said.

Winners and Losers Inside Europe

Not every European country is equally exposed. Data center investment will migrate toward wherever electricity is cheapest — meaning the continent will develop internal fractures between energy-rich and energy-poor regions.

Countries with cheaper hydroelectric power, like the Nordic nations, stand to capture investment. Countries still struggling with high gas dependence will watch projects walk out the door. It's already happening.

The European Council on Foreign Relations flagged the same issue, warning that slow decision-making across Europe undermines the continent's security and prosperity — from defense to energy to tech. Europe faces both a bureaucratic speed problem and an energy price problem.

What Mainstream Coverage Is Missing

Most coverage of this story frames it as a clean energy vs. AI dilemma — the left worries about carbon, the tech press worries about compute capacity, and nobody connects the dots.

This is a geopolitical crisis, not just an economic inconvenience.

If AI infrastructure concentrates in the U.S. and China because those countries have cheap, abundant power, Europe doesn't just lose market share. It loses strategic leverage. It becomes dependent on AI infrastructure it doesn't control, built by companies that answer to Washington or Beijing — NOT Brussels.

Data centers now consume 2% of the world's electricity, up from 1.7% in 2024, according to the International Energy Agency. That number is accelerating fast. Every percentage point of growth is a moment where geography and energy policy determine who wins.

The Fix Exists — Europe Just Has to Want It

Ciani at the Kiel Institute doesn't just diagnose the problem. She prescribes the solution: link data center expansion directly to new low-carbon electricity supply. Build the power generation alongside the data centers. Don't assume the grid will absorb the load — because it won't.

Public-private partnerships, stronger coordination between energy and digital planning, renewable generation tied contractually to new AI infrastructure. These aren't radical ideas. They're just not happening at the speed required.

The End Game

Europe is attempting to run an AI race with a flat tire. Double the energy costs, an aging grid, and bureaucratic inertia that moves at the speed of committee meetings — while the U.S. and China sprint ahead.

This isn't about green energy versus fossil fuels. It isn't about which political party runs Brussels. It's about whether Europe can make decisions fast enough to matter. So far, the answer is no. And the clock is running.