Artificial intelligence (AI) is rapidly reshaping the economic and social landscape, confronting Europe with critical strategic choices. In a world where growth, sovereignty and the preservation of fundamental values interact in increasingly complex ways, the approach Europe adopts towards AI has become an overriding strategic imperative.
Two of the most influential voices in the European public debate, Mario Draghi and Christine Lagarde, approach the same critical dilemma from different perspectives, yet converge on a common conclusion: inaction is not an option.
I. Europe’s Difficult Choices: Mario Draghi’s Analysis
In his paper, “Europe’s Difficult Choices on AI”, Mario Draghi highlights Europe’s fundamental trilemma:
- The need for growth, in order to continue financing its social model and addressing emerging strategic priorities, including defence and the energy transition.
- The pursuit of sovereignty, enabling Europe to shape its economy in accordance with its own values and priorities.
- The objective of sustainable growth, without the adverse consequences of widening inequalities and environmental degradation.
The Productivity Gap
Europe has experienced a marked slowdown in the post-pandemic period, further widening the productivity gap with the United States. The adoption of AI has emerged as the most promising avenue for bridging this gap. According to European Central Bank (ECB) estimates, rapid AI adoption could add 0.3 to 0.4 percentage points annually to total factor productivity growth.
Growth in Tension with Sovereignty
However, this prospect creates a profound tension with the pursuit of technological sovereignty. Europe controls only a small share of the AI value chain, while the technology is expected to become ubiquitous, permeating every sector of the economy, from healthcare and education to energy and defence.
Disconnecting from AI once the economy has become dependent on it would have consequences comparable to being excluded from the US financial system. This creates opportunities for coercive dynamics, whereby third countries could potentially use access to AI models as leverage in negotiations over tariffs or digital taxes.
Data as a Strategic Advantage
Draghi points out that Europe’s sphere of sovereignty in AI remains extremely limited. European research teams struggle to compete financially with their American and Chinese counterparts, while the production of cutting-edge semiconductors lags significantly behind.
Nevertheless, Europe retains a substantial advantage in one particular area: data. The continent possesses vast amounts of data from both the public sector, including medical records, and industry. The European data economy is projected to exceed €800 billion by 2030.
The Critical Shortfall in Computing Capacity
To capitalise on these assets, Europe needs to retain control over their storage and processing, which requires large-scale, AI-oriented data centres.
At present, Europe hosts less than 5% of global AI computing capacity, compared with 75% in the United States. This gap is expected to widen further. If it is not addressed, European businesses will face an uncomfortable dilemma: either refrain from adopting AI due to concerns about data security or adopt foreign solutions at prohibitive cost.
Environmental concerns associated with data centres, including energy consumption and higher electricity bills, can be mitigated through clean energy, the reuse of waste heat and the fair allocation of infrastructure costs. Nevertheless, construction delays are also attributable to supply-side constraints, such as lengthy permitting procedures.
Demand Fragmentation as the Main Obstacle
According to Draghi, the most significant obstacle is fragmented demand.
Data centres require substantial capital commitments, which investors are willing to undertake only when they can secure bankable offtake contracts—long-term agreements guaranteeing demand for their services.
In the United States, such guarantees are provided by major cloud service providers and leading AI laboratories. In Europe, by contrast, demand is dispersed across millions of small and medium-sized enterprises.
Draghi proposes the creation of large, aggregated buyers: major companies and public-sector entities that pool their commitments into long-term contracts, following the model of Mistral European Compute Units.
This could establish a virtuous cycle:
Accelerated AI adoption → Financing of European computing capacity → Sovereign infrastructure → Faster economic growth → The ability to develop autonomous European AI models.
II. A New Age of Capital: Christine Lagarde’s Perspective
In her speech, “A New Age of Capital: Growth, Sovereignty and AI” (September 2026), Christine Lagarde places the AI challenge within a broader historical context, referring to Germany’s Gründerzeit, or the nineteenth-century “founders’ period”.
During that era, European savers financed the transformative technologies of the time, including American railways. The outcome? Following the crash of 1873, Europe experienced two decades of sluggish growth, while the United States underwent an explosive expansion.
Today, history threatens to repeat itself. European savers are once again channeling capital into transformative technologies, particularly AI, with approximately €440 billion flowing to American technology companies.
Lagarde poses a critical question: Will Europe benefit from this growth this time around?
Progress and Outstanding Challenges
Europe has made progress. In 2026, AI accounts for 10% of total investment, while more than 50% of workers in the euro area use AI in their work.
Nevertheless, the gap with the United States remains substantial, as American digital investment is growing at twice the European rate.
This is particularly significant for a continent facing demographic ageing, with its workforce expected to shrink by more than one million people annually over the next 25 years.
Increasing productivity through AI is therefore vital. ECB estimates indicate that AI could raise productivity by up to 4% over a decade, a transformation that would have profound implications for public finances.
Lagarde emphasises a critical reality: Europe will pay the price of AI development whether or not it participates in it.
The rapid expansion of AI investment, primarily in the United States, is driving massive bond issuance by American hyperscalers. This affects European bond yields and, indirectly but inevitably, increases Europe’s borrowing costs.
Why AI Adoption Alone Is Not Enough
Unlike the railways of the nineteenth century—which were, to a significant extent, a European invention—AI companies are being built elsewhere.
The United States has produced 59 notable AI models, while China has produced 35, compared with just one in France and one in the United Kingdom.
Lagarde identifies three reasons why AI is uniquely problematic when imported from abroad.
1. Data
AI adoption means that sensitive data are processed through systems owned by third parties and subject to foreign legislation. The model provider gains deep insights into entire industries and potentially acquires the ability to compete with them.
It is no coincidence that almost half of European companies that have refrained from adopting AI have cited precisely these concerns.
2. Access
AI will become so deeply embedded in the management of commodities, tax returns, transport operations and banking payments that any interruption of access or unilateral change to contractual terms could affect every sector simultaneously.
This gives providers unprecedented bargaining power.
3. The Frontiers of Innovation
In cutting-edge sectors such as pharmaceuticals, finance and defence, even a small advantage in AI capabilities can translate into substantial gains. This creates “winner-takes-all” dynamics and can permanently entrench technological leadership.
The “Uncomfortable Choice” and the Way Forward
Confronted with these distinctive challenges, Europe faces an “uncomfortable choice”: either delay adoption, sacrificing growth, or adopt AI rapidly, becoming highly dependent on external providers and risking the loss of its freedom to organise its economy in accordance with its own principles.
To reconcile growth with sovereignty, Lagarde proposes three strategic priorities.
1. Building European Computing Capacity
The existing gap in data-centre capacity is expected to increase sixfold over the next decade. The cost of closing this gap is estimated at €600 billion.
While the United States draws on a broad range of capital markets, Europe continues to rely primarily on bank lending.
2. Developing “Good Enough” European AI Models
Europe should develop AI models that are sufficiently capable for its needs and operate on European infrastructure, thereby substantially reducing the risk of being cut off from essential services.
France’s Mistral already provides an example of this approach.
3. Securing Access to the Frontiers of Innovation
Europe must position itself as an indispensable link in the global AI supply chain, in the same way that ASML occupies a critical position in semiconductor lithography.
Achieving this requires sustained, targeted innovation.
The Question of Capital
All these objectives require more capital and different forms of capital.
Beyond the €600 billion required for infrastructure, developing new models and securing critical positions in the value chain will require equity capital, rather than debt financing alone.
European savings amount to approximately €1.4 trillion annually. However, Europe lacks the mature, integrated capital markets needed to channel these savings efficiently into productive domestic investment.
Lagarde concludes with a compelling proposition: AI represents precisely the kind of “transformative undertaking”that Europe needs to propel its capital markets forward.
It is a challenge of such magnitude that only markets can finance it effectively, while simultaneously offering Europe an opportunity to build its own future using its own savings.
Leave a Reply