The Old Continent seeks to overtake rivals in the AI race

Europe was slow to realize that artificial intelligence is more than just a technology. It is a force multiplier.

AI determines business productivity, the capabilities of public institutions, the effectiveness of armed forces, the pace of scientific research, and control over information flows. It shapes industrial competitiveness and, increasingly, a state's very ability to exercise sovereignty. This is why Brussels is now seeking to transform artificial intelligence from a subject of regulation into a strategic infrastructure.

The problem is that AI, often portrayed as an immaterial phenomenon, is fundamentally physical in nature. It requires advanced semiconductors, massive computing facilities, high-capacity networks, cooling systems, water, a continuous supply of electricity, and vast amounts of capital. The infrastructure comes before the algorithm. And it is precisely in this infrastructure that Europe's vulnerabilities become apparent.

Power does not reside in the cloud

By 2026, the European Union has 19 AI Factories, built around Europe’s leading supercomputers, and is preparing the next generation of AI Gigafactories.

These facilities are expected to house more than 100,000 advanced processors each, forming the physical core of the continent's new strategy. Brussels is also seeking to mobilize around €20 billion through InvestAI to expand computing capacity across Europe. This signals a conceptual shift.

For years, the European Union approached the digital sphere primarily through law: competition, privacy, consumer protection, and platform regulation. This was the typical stance of a regulatory power, capable of setting standards but far less capable of controlling the infrastructure needed to produce the technology it sought to regulate. Today, the limits of that paradigm are becoming clear.

Computing capacity has become a strategic resource, comparable in some respects to energy infrastructure or telecommunications networks. Without reliable access to computing power, even advanced industrial systems risk becoming dependent on external providers. After all, the cloud is not a cloud. It is territory.

It consists of buildings, cables, servers, transformers, power grids, and processors. Every artificial intelligence model presupposes a specific geography.

Algorithms consume energy

Europe's second vulnerability is energy.

According to the International Energy Agency, electricity demand across the European Union is expected to grow by an average of around 2.3 percent a year through 2030. In the coming years, Europe's power system will have to accommodate the electrification of transport, industrial decarbonization, the expansion of heat pumps, and the growth of data centers.

Data centers alone could account for more than 3 percent of the EU's total electricity consumption by 2030. This makes the issue geopolitically significant, as artificial intelligence is introducing a new dimension to competition for energy resources.

Having access to electricity is not enough. What matters is an abundant, reliable, dispatchable, and competitively priced energy supply. This is where structural differences between European countries become apparent. In the second half of 2025, the average electricity price for non-household consumers in the European Union stood at around €18.37 per 100 kWh. Yet the variation across countries was substantial, ranging from just over €7 in Finland to more than €25 in Ireland.

The implications are clear: major digital investments will tend to concentrate in countries where energy is cheaper, power grids are more robust, and permitting processes are faster. The geography of AI therefore risks converging with the geography of energy.

Europe's Weakness Lies in the Supply Chain

Europe's dependence extends beyond electricity. The continent remains weak in the most critical segments of the advanced semiconductor supply chain, in hyperscale cloud computing, and in its ability to rapidly finance technology companies capable of competing on a global scale.

The problem is not a lack of expertise. Europe has universities, research centers, advanced industries, and highly skilled talent.

The problem is turning knowledge into scale.

As early as 2025, the European Investment Bank reported that 37 percent of European firms were using generative AI applications—a share comparable to, and in that particular survey even slightly higher than, that of US firms. But adoption does not mean control.

A company can adopt other people's technologies extremely quickly and still remain dependent on them. This happens when the highest value-added activities are concentrated in segments of the supply chain dominated by others: semiconductors, platforms, cloud computing, foundation models, and computing infrastructure.

This is Europe's paradox: strong on the demand side, weaker in the architecture that makes supply possible.

Capital as a Geopolitical Tool

The gap with the United States becomes most apparent when capital enters the equation.

Between 2020 and 2025, venture capital investment in artificial intelligence followed markedly different trajectories on the two sides of the Atlantic. In the United States, the availability of private capital enabled companies to scale rapidly, funding the enormous costs of developing models, building infrastructure, and hiring specialized talent.

Europe has savings, but struggles to channel them into strategic risk capital.

When a European startup grows and requires more than €25 million in funding, a significant share of that capital still comes from investors based outside the European Union. This is a crucial factor.

Capital is not neutral. It determines where companies are based, which markets they serve, which standards they adopt, and which industrial ecosystems they support. In a sector characterized by rising costs and powerful economies of scale, those with greater financial firepower inevitably tend to concentrate power.

European Workers Are Already Using AI

While institutions are trying to build the strategic architecture, the social transformation is already underway.

In 2026, more than half of workers in the euro area report using artificial intelligence tools in their professional activities. Adoption has accelerated rapidly in just two years, while companies are allocating a growing share of their investment to software, automation, and intelligent systems. This brings a second dimension of the issue into focus.

Artificial intelligence is reaching Europe through the workplace before it does through sovereignty.

It is entering offices, businesses, universities, professional practices, and decision-making processes. It is becoming part of everyday infrastructure even before the continent has decided how much control it intends to retain over it. This creates an asymmetrical relationship: the more widespread AI adoption becomes, the higher the cost of any eventual technological dependence.

An Ageing Power Seeks Productivity

For Europe, the stakes are even higher because artificial intelligence intersects with a structural challenge: demographics.

The continent is ageing. The working-age population is set to shrink. Welfare systems will come under increasing pressure. Productivity will therefore become essential to sustaining income levels, public spending, and competitiveness.

The European Central Bank estimates that the rapid adoption of artificial intelligence could raise Europe's productivity level by up to 4 percent over the course of a decade. For a continent that will have fewer workers and more retirees, this is no minor economic consideration. It is a strategic necessity.

AI can partially offset demographic decline. But only if Europe succeeds in embedding it in its productive economy without turning itself into a mere consumer market for technologies developed elsewhere.

Sovereignty Beneath the Surface

Europe's challenge, then, is not necessarily to develop the world's most powerful AI model. It is to control a sufficient share of the essential building blocks of its own autonomy.

Energy, chips, capital, computing capacity, cloud infrastructure, networks, expertise.

Technological sovereignty does not mean autarky. No major power today exercises complete control over all its own value chains. Rather, it means having the capacity to prevent external dependencies from becoming political vulnerabilities.

Europe now faces this challenge.

For years, it believed that digital power resided primarily in regulation. It is now discovering that rules matter only when they rest on a sufficiently robust material foundation.

Because behind every algorithm lies a machine. Behind every machine, a network. Behind every network, energy and territory. And it is there, far more than on the screen, that the question of who will be sovereign in the age of artificial intelligence will be decided.

This version is better suited to a geopolitics section because it places less emphasis on explaining the technology to a general audience and more on structural constraints, power relations, dependencies, supply chains, and sovereignty.

Mondo Internazionale APS - All Rights Reserved ® 2026