Europe is confronting a complex challenge as it seeks to balance economic growth, sovereignty, and social values in the era of artificial intelligence (AI). The continent aims to boost productivity and innovation to support its social model and meet emerging demands in areas such as defence and energy, while avoiding the pitfalls of wealth inequality and environmental harm seen elsewhere.

Since the pandemic, Europe’s growth trajectory has slowed relative to its own past performance and compared to the United States. Estimates indicate that the productivity gap between the euro area and the US widened significantly, from around $9 per hour in 2018 to approximately $21 in 2025. Among various strategies to stimulate growth, widespread AI adoption is seen as a promising avenue. According to scenarios from the European Central Bank, rapid AI integration could contribute an additional 0.3 to 0.4 percentage points annually to total factor productivity—a measure of efficiency gains not reliant on increased labor or capital inputs—which has stagnated since 2022.

However, this growth potential introduces tensions with Europe’s pursuit of technological sovereignty. Currently, the continent controls only a small portion of the AI value chain, making it vulnerable to potential restrictions on access from major players, namely the United States and China. The pervasive role AI is set to play across sectors—including health, education, energy, and defence—means that dependency could have serious strategic and economic consequences if access were to be limited by foreign governments.

Europe’s ability to assert sovereignty in AI is constrained by its underdeveloped frontier labs and chip manufacturing compared to American and Chinese capabilities. Nonetheless, data storage and processing remain areas where Europe can maintain greater control. The continent possesses extensive public-sector data, including decades of health records and industrial information, which the European Commission estimates could underpin a data economy exceeding €800 billion by 2030, representing more than 5 percent of GDP.

To capitalize on this advantage, Europe requires the expansion of large-scale AI data centers under its control. Yet constructing such facilities faces challenges. Local communities express environmental and energy cost concerns, and much of the current European data center capacity is operated by American technology firms. Unlike the debate in the US about overbuilding data centers, Europe actually suffers from a significant deficit. The EU accounts for less than 5 percent of the world’s AI compute capacity, compared to 75 percent for the US, and there is a widening gap between demand and supply for data center infrastructure projected to reach 14 gigawatts by 2030.

This shortage threatens to restrict companies from keeping sensitive data and AI workloads within Europe. While American firms offer so-called sovereign cloud services hosted on European soil, ultimate control remains foreign, prompting the European Commission's proposal of a tiered regulatory framework in the Cloud and AI Development Act to ensure critical infrastructure is governed locally.

Obstacles to scaling up data center capacity include lengthy permitting processes—42 months in Germany compared to 24 in the US—and higher electricity costs in parts of Europe. Nonetheless, there are regions like Sweden where costs are more competitive, roughly 10 percent higher than in China.

A central issue is Europe’s fragmented demand. Unlike in the US, where large technology companies provide stable, sizable contracts to support investment in data centers, European demand is dispersed across millions of firms. Companies typically purchase compute resources on short-term, uncertain horizons, limiting investor confidence. This dynamic has resulted in European AI cloud providers like Nscale selling much of their capacity to US customers.

To overcome this, Europe needs to create concentrated demand through consortia of large companies pooling commitments into multiyear contracts that underpin investment. An existing example involves firms such as ASML, Capgemini, and Amadeus investing in Mistral’s European Compute Units, aiming to support 1 gigawatt of capacity by 2030. Expanding such arrangements could accelerate AI adoption, establish sovereign compute infrastructure, and thereby enable Europe to develop its own AI capabilities more robustly.

A further concern is the potential concentration of wealth and power among a few technology giants as AI grows in influence. The proposed model of pooled demand and diversified providers intends to promote competition and prevent oligopoly formation. Additionally, measures like progressive taxation could help distribute AI-generated gains more broadly across society.

By pursuing this strategy, Europe aspires to reconcile the sometimes conflicting goals of growth, sovereignty, and social fairness, positioning itself to benefit from AI without compromising its values or becoming overly dependent on external actors.