There’s something almost surreal about how big things are right now. In rural Virginia, on the edges of Shanghai, and in the flatlands of Frankfurt, huge buildings are being built that are climate-controlled, have no windows, and are buzzing with electricity, but no one seems to notice. The race to build these data centers has become one of the most important geopolitical stories of the decade.
It’s hard to fully understand the numbers. By the beginning of 2025, 122.2 gigawatts of IT power had been installed in data centers around the world. About 44% of that comes from just the United States. Around 17% is held by Europe. China has one of the biggest economies in the world, but they only claim about 16% of the market. This number probably doesn’t reflect the truth, though, because Chinese operators tend to keep facility locations secret. Just the fact that there is a gap in transparency shows how differently these three powers see the same race.
Obviously, artificial intelligence is what’s making all of this happen. AI doesn’t set goals for itself. It needs servers, cooling systems, and a lot of electricity to work. The International Energy Agency says that by 2030, the amount of electricity used by data centers will have more than doubled, making up almost 3% of all the electricity used in the world. It’s not a rounding mistake. That’s a big change in how people use energy that’s happening faster than most power grids were made to handle.
In the US, the model is mostly run by the private sector with help from the government. The big hyperscalers, or cloud giants whose names are on every earnings call, are doing most of the work. They are signing long-term contracts for clean energy, co-investing in renewable projects, and moving into states with cheap land and flexible zoning laws. In Virginia, Georgia, and Ohio, these clusters are growing so quickly that they are already putting stress on the local grids. Sometimes the federal government steps in, but most of the time it’s just to remove problems and not make decisions about strategy. It’s messy and doesn’t always work well, but it moves.
China’s way of doing things is very different. There, the market and the government are working together, not against each other. Beijing’s “Eastern Data, Western Compute” program is a good example. It moves non-urgent computing tasks to western provinces with lots of renewable energy, while work that needs to be done quickly stays near the coasts. To make things even more efficient, the country wants a Power Usage Effectiveness standard of 1.25 by 2025. This is a lot stricter than what Germany will require starting in 2027. It’s not clear if China hits all of its targets, but the country is clearly moving in the right direction.
The situation in Europe, on the other hand, is not good. It has some of the strictest rules for data centers in the world. These include detailed sustainability metrics, mandatory reporting frameworks, and a real dedication to building low-carbon infrastructure. Those things are big. But rules without ready infrastructure are like having great traffic laws for roads that aren’t even built yet. Delays in getting permits are real. The amount of data that can be sent is limited. A number of EU countries have already put limits on new connections between data centers and the grid. In Ireland, the Netherlands, Belgium, and Germany, the tension between digital demand and energy supply has been felt in real, everyday ways.
It’s still not clear if Europe can change its course quickly enough. The global data center market was worth more than $527 billion in 2025 and is expected to keep growing. However, most of that money goes to places where energy is cheap, permits are easy to get, and incentives are strong. Europe’s identity and, some might say, its strength is its strict regulations. But more and more people think that strictness won’t be enough to win this race.
It’s becoming clear that data centers are more than just places to store data. At the same time, they are political assets, questions about energy policy, and economic bets. Any country that can power them quickly, cleanly, and on a large scale will be the leader in more than just AI. They will decide who can be a part of the next phase of the digital economy and how it looks.
