London: Europe is entering a new phase of its artificial intelligence expansion as companies building large data centres move away from traditional technology hubs and search for places where electricity, land and grid connections are easier to secure.

The change reflects a simple but growing reality. Modern AI systems need enormous computing power, and that computing power requires enormous amounts of electricity. For data centre developers, access to reliable power is now becoming more important than being close to major cities.

New AI focused data centres are increasingly being planned in less traditional locations, including rural and edge of city areas. The average distance of new hyperscale projects from major European cities has risen to about 175 kilometres, compared with around 46 kilometres for projects built in recent years. Greenfield developments are also becoming much more common, making up about 39 percent of the future pipeline compared with only 8 percent previously.

The change is particularly visible in places such as northern Sweden, rural Spain and other areas where land is more affordable and electricity can potentially be connected more quickly.

Europe's established data centre markets remain important. London, Frankfurt and Amsterdam continue to attract investment because they have strong digital networks, established businesses and good connections to customers. However, these markets are facing growing pressure from limited land and increasingly difficult grid connections.

The cost difference can also be substantial. Powered land in Amsterdam can cost several million euros for each megawatt of capacity, while comparable sites in parts of France can be available at a fraction of that price. For companies planning facilities that may eventually require hundreds of megawatts or even several gigawatts, the difference can have a major effect on investment decisions.

The growing demand for electricity is also changing the competition between European countries.

France is emerging as one of the strongest candidates to benefit from the new trend. Its large electricity system, nuclear generation and government support for technology investment are attracting major interest. SoftBank has announced plans for a large data centre investment in northern France, with three facilities expected to have combined capacity of about 3.1 gigawatts by 2031. The planned investment could eventually become even larger.

Other countries are also looking for ways to benefit from the AI infrastructure boom while protecting their electricity systems. Portugal, for example, has said it wants data centres that bring wider economic benefits rather than simply attracting power intensive projects that provide limited value to the country.

The pressure on Europe's electricity networks is expected to grow significantly in the coming years. Global data centre electricity consumption is projected to nearly double by 2030, reaching about 945 terawatt hours. Europe is expected to account for a significant part of that increase as artificial intelligence becomes more widely used.

This creates a difficult challenge for governments. They want to attract investment in AI because it can create jobs, improve digital infrastructure and strengthen Europe's position in the global technology race. At the same time, large data centres can place heavy demands on electricity networks and local infrastructure.

The issue is already visible in countries such as Ireland, where data centres have become a major part of national electricity demand. Britain is also facing a large queue of proposed data centre projects competing for connections to the electricity grid.

The problem is not simply about generating more electricity. New transmission lines, substations and other infrastructure can take years to build. In some European markets, grid connection times for large data centres are becoming long enough to influence where companies choose to build their facilities.

Water is becoming another concern. Large data centres require cooling systems to keep computer equipment operating safely. As the number and size of facilities increase, governments and communities are paying closer attention to their effect on local water supplies.

Environmental concerns are also growing. Although technology companies are investing heavily in renewable energy, the speed of AI expansion can make it difficult to provide clean electricity quickly enough. In some markets, growing demand has increased interest in gas power and other sources that can provide reliable electricity when renewable generation is unavailable.

The result is a major shift in the geography of Europe's digital economy. The most important locations for AI infrastructure may no longer be the biggest cities. Instead, regions with abundant electricity, available land, strong fibre networks and supportive governments could become the new centres of Europe's AI industry.

This does not mean established technology hubs will disappear. London, Frankfurt, Amsterdam and other major markets will continue to play important roles in cloud computing, finance, connectivity and business services. But the largest and most electricity intensive AI facilities are increasingly likely to be built farther away from crowded urban areas.

Europe's AI race is therefore becoming a race for physical infrastructure. Access to advanced computer chips remains important, but having enough electricity, land, cooling capacity and grid connections may determine how quickly those chips can actually be put to work.

The coming years will test whether Europe can expand its AI infrastructure while keeping electricity affordable, protecting the environment and ensuring that local communities benefit from the investment. For data centre developers, the most valuable location may now be not the one closest to Europe's largest city, but the one where reliable power can be obtained quickly and at a reasonable cost.