The EU's next five years will define its next fifty

The 2024–2029 Commission faces digital regulation, industrial policy, defence, enlargement, and AI governance simultaneously. This is not a legislative cycle — it is a generational inflection point.

By VastBlue Editorial · 2026-03-26 · 14 min read

Series: The Chessboard · Episode 9

The inheritance

When Ursula von der Leyen secured her second term as President of the European Commission in July 2024, the margin was narrower than her allies had hoped — 401 votes in a European Parliament that had shifted rightward, with gains for nationalist and Eurosceptic parties across the continent. The arithmetic mattered less than what it signified. The political coalition that had underwritten European integration for seventy years — centre-right and centre-left parties alternating in broad agreement on the direction if not the speed of travel — was thinner than at any point since the Maastricht Treaty. Von der Leyen had a mandate, but it was a mandate delivered with conditions, qualifications, and an unmistakable message from European voters: the next five years had better produce results that the last five did not.

The institutional context makes the challenge clearer. The European Commission is not a government in any conventional sense. It proposes legislation but cannot pass it alone — that requires the European Parliament and the Council of the European Union, where twenty-seven national governments defend twenty-seven sets of national interests. It has a budget, but that budget is roughly one percent of EU GDP, compared to the twenty-four percent commanded by the US federal government. It can set strategic direction, but implementation depends on member states that vary enormously in capacity, political will, and domestic constraints. The Commission is, in structural terms, a strategic planning department with limited operational authority — powerful enough to shape the agenda, too weak to execute it unilaterally.

This structural reality has always constrained European ambition. What makes the 2024–2029 mandate different is that the challenges confronting Europe no longer permit the Union's traditional method of progress: incremental consensus-building over multiple institutional cycles, where every major initiative takes a decade from proposal to implementation and another decade from implementation to effect. The world is not waiting for Europe to build consensus. The United States is spending hundreds of billions to reshore strategic industries. China is exporting industrial overcapacity at prices that threaten European manufacturers. Russia's war in Ukraine has exposed defence vulnerabilities that cannot be resolved through communiqués. Artificial intelligence is transforming economic structures faster than any previous general-purpose technology. Each of these pressures demands a response measured in years, not decades. The Commission's five-year term — which is, in practice, closer to four years of productive legislative time — may be insufficient for any single one of them. It must address all of them simultaneously.

The digital regulation paradox

Europe's regulatory achievements of the previous Commission cycle were, by any objective measure, historically significant. The Digital Markets Act (DMA), the Digital Services Act (DSA), the AI Act, the Data Act, the Data Governance Act, and the European Chips Act collectively represent the most comprehensive attempt by any democratic jurisdiction to establish rules for the digital economy. The legislative architecture is impressive. The challenge now is making it work — and this is where the paradox emerges.

The DMA designated six companies as "gatekeepers" — Alphabet, Amazon, Apple, ByteDance, Meta, and Microsoft — and imposed a set of behavioural obligations designed to ensure contestability in digital markets. The obligations are specific and far-reaching: Apple must allow alternative app stores on iOS. Google must offer users a choice screen for default search engines. Amazon must not use third-party seller data to advantage its own products. Meta must allow interoperability for messaging services. Each obligation requires technical implementation, compliance monitoring, and enforcement capacity. The Commission's DG Competition, which leads enforcement, has substantial expertise but limited personnel — approximately 900 staff for a portfolio that now includes merger review, antitrust enforcement, state aid control, and the entirety of digital markets regulation.

The enforcement challenge is structural, not merely administrative. Each gatekeeper has deployed teams of hundreds of lawyers, engineers, and compliance specialists whose professional purpose is to interpret the DMA's obligations in the narrowest possible manner consistent with technical compliance. Apple's initial response to the alternative app store requirement — imposing a "Core Technology Fee" of €0.50 per app install beyond one million, effectively making alternative distribution uneconomic for most developers — demonstrated the pattern. The company complied with the letter of the regulation while arguably negating its intent. The Commission opened a non-compliance investigation in June 2024. The investigation is ongoing. The clock is ticking on whether the DMA produces genuine market change or becomes a monument to well-drafted rules that sophisticated actors circumvent.

The AI Act, which entered into force in August 2024 with a staggered implementation timeline extending to 2027, presents a different category of enforcement challenge. The DMA targets a small number of identified companies with specific obligations. The AI Act applies across the entire economy, classifying AI systems by risk level and imposing requirements that scale with the classification. High-risk AI systems — used in employment decisions, credit scoring, law enforcement, critical infrastructure — face requirements for conformity assessment, technical documentation, human oversight, transparency, and ongoing monitoring. The number of organisations that will need to comply is not six. It is potentially hundreds of thousands.

The EU AI Office, established within the Commission to oversee implementation, began operations in 2024 with an initial staff of approximately 140. By comparison, the United States Securities and Exchange Commission — which regulates a single sector of the economy — employs over 5,000 people. The mismatch between the scope of the AI Act's regulatory ambition and the resources allocated to its implementation is not a criticism of the Office's competence. It is a statement about institutional capacity. The AI Act requires not just legal enforcement but technical expertise — the ability to audit algorithms, evaluate training data, assess model architectures, and make determinations about risk levels that require deep familiarity with machine learning systems. Building this capacity takes time that the technology's development pace may not allow.

Europe has built the most comprehensive regulatory framework for digital markets and artificial intelligence that any democracy has attempted. The question is no longer whether the rules are right. It is whether the institutions charged with enforcing them have the capacity, resources, and speed to make them matter.

Editorial observation

The deeper paradox is competitive. Europe designed its digital regulation to protect European citizens and create a fair market that European companies could compete in. But the regulatory burden falls disproportionately on the companies that must comply — and the largest, best-resourced companies can absorb compliance costs that smaller European competitors cannot. A startup building an AI-powered medical diagnostic tool faces the same conformity assessment requirements as Google DeepMind. The regulatory framework that was designed to level the playing field may, in practice, raise the barriers to entry. The Commission is aware of this risk — the AI Act includes provisions for regulatory sandboxes and lighter requirements for SMEs — but the practical effectiveness of these mitigations will not be known until the Act is fully operational. By then, the market may have already consolidated around the companies that could afford to comply first.

Industrial policy without a federal budget

The Draghi Report, published in September 2024, delivered a diagnosis that was as uncomfortable as it was unimpeachable. Europe needed €750 to €800 billion per year in additional investment to close the competitiveness gap with the United States and China. European productivity growth had lagged America's for two decades. European energy costs were two to three times higher than American costs and four to five times higher than Chinese ones. European venture capital investment was a fraction of American levels. The report's conclusion was framed as a policy recommendation but read as an indictment: Europe was falling behind, and the pace of decline was accelerating.

The Draghi diagnosis was not new. The Letta Report on the single market, published months earlier, reached similar conclusions from a different angle — focusing on how the fragmentation of European markets in services, capital, and energy was undermining competitiveness. What was new was the political moment. With American industrial subsidies under the Inflation Reduction Act pulling investment westward, Chinese overcapacity pushing European manufacturers out of their own supply chains, and European defence spending needing to roughly double, the argument for bold action had never been stronger. The institutional capacity for bold action had never been more constrained.

The fundamental constraint is constitutional, not political. The European Union does not have a federal fiscal capacity. It cannot issue common debt on a routine basis — the NextGenerationEU recovery fund, which raised €750 billion through common borrowing, was justified as an exceptional response to the pandemic, and several member states (led by the so-called "frugal four" — the Netherlands, Austria, Denmark, and Sweden) insisted that it remain a one-time instrument. Without common borrowing, the EU budget remains capped at approximately one percent of GNI — a figure that has barely changed in thirty years, despite the Union's responsibilities having expanded dramatically.

This means that when the Commission announces an "industrial strategy" or a "sovereignty fund" or an "investment plan," the money must come from one of three sources: reallocation within the existing EU budget (robbing one priority to fund another), relaxation of state aid rules to allow member states to spend their own money (which advantages rich countries over poor ones), or creative financial engineering through the European Investment Bank and other institutions that can leverage public capital to attract private investment. Each approach has been deployed. None approaches the scale that the Draghi Report identified as necessary.

The Clean Industrial Deal, proposed by the Commission in early 2025 as the successor to the Green Deal Industrial Plan, illustrates both the ambition and the constraints. It proposes to simplify permitting for clean-energy projects, create lead markets for European clean technology, expand carbon border adjustments, and mobilise investment for industrial decarbonisation. The proposals are directionally sound. The funding mechanisms rely heavily on private capital mobilisation and EIB lending rather than direct public expenditure. For a European battery manufacturer competing against a Chinese rival backed by state-directed credit at below-market rates, the distinction between a direct subsidy and a leveraged loan guarantee is not academic — it is the difference between competitive viability and market exit.

The state aid dimension adds a layer of internal tension that the Commission must navigate with care. When France offers €1.3 billion to attract battery manufacturing and Germany pledges €3.5 billion for semiconductor fabrication, smaller member states — Portugal, Greece, the Baltic states, the Balkans — watch their competitive position erode not because of external competition but because of the fiscal asymmetries within the Union itself. The single market, Europe's most transformative economic achievement, is being strained by the same forces that are supposed to protect it. A credible European industrial policy must find a way to channel investment toward competitiveness without fragmenting the internal market that makes European-scale competitiveness possible. No one has yet designed an instrument that does both. The 2024–2029 Commission must try.

Defence: the end of the peace dividend

For thirty years after the Cold War, Europe operated on an implicit strategic assumption: that large-scale conventional conflict on the European continent was a historical phenomenon, not a current risk, and that defence spending could therefore be treated as a residual budget category — funded after healthcare, pensions, education, and infrastructure had taken their shares. NATO's two percent of GDP spending target, agreed at the Wales Summit in 2014, was treated by most European member states as an aspiration rather than an obligation. In 2021, the year before Russia's full-scale invasion of Ukraine, only eight of NATO's thirty-one members met the target. The European average was approximately 1.5 percent.

February 24, 2022, ended the strategic holiday. Russia's invasion of Ukraine — the largest conventional military operation in Europe since 1945 — demonstrated that territorial aggression by a nuclear-armed power was not a theoretical scenario but an operational reality. The European response was swift in political terms: unprecedented sanctions against Russia, military aid to Ukraine, and a rhetorical commitment to a "Zeitenwende" — the term used by German Chancellor Olaf Scholz to describe the tectonic shift in European security. Germany announced a €100 billion special fund for its military. Poland committed to spending four percent of GDP on defence. Finland and Sweden abandoned decades of military non-alignment to join NATO. The political will, dormant for a generation, appeared to materialise overnight.

Three years later, the gap between rhetoric and reality is becoming visible. Germany's Sondervermögen — the special fund — has been allocated, but the procurement process has been hampered by a defence-industrial base that atrophied during the peace dividend decades. Rheinmetall, Europe's largest ammunition manufacturer, has expanded capacity but cannot yet meet the volumes that a sustained conflict or credible deterrence posture requires. European artillery ammunition production in 2024 was estimated at approximately 1.3 million rounds per year — an impressive increase from the 300,000 rounds of pre-2022 capacity, but still below the rate at which Ukraine was consuming ammunition in active combat. Building defence-industrial capacity is not like building software — it requires specialised machine tools, skilled workers with security clearances, hardened supply chains for energetic materials, and facilities that take years to construct and certify.

The Commission's response has been to treat defence as an industrial policy challenge as well as a security one. The European Defence Industrial Strategy (EDIS), proposed in March 2024, set targets for intra-EU defence procurement: 50 percent of defence equipment budgets should be spent within the EU by 2030, and 60 percent by 2035. The Defence Industrial Readiness Programme proposed incentives for joint procurement, common standards, and cross-border defence supply chains. A proposed €1.5 billion European Defence Fund aimed to co-finance research and development. These are meaningful steps. They are also modest relative to the scale of the challenge. The United States spends more on defence in a single month than the European Defence Fund will spend in its entire programming period.

The deeper strategic question is whether Europe can build credible defence capacity while maintaining its alliance with the United States — and what happens if that alliance weakens. The 2024 US presidential election and its aftermath raised questions about American commitment to European security that would have been unthinkable a decade ago. The idea that Europe might need to provide for its own conventional defence — deterrence against Russia without guaranteed American reinforcement — implies a level of military spending, industrial mobilisation, and strategic coordination that the EU's current institutional architecture was not designed to deliver. Defence remains a national competence. The EU has no army, no defence minister with operational authority, and no command-and-control infrastructure for collective defence. NATO provides the alliance framework, but NATO without robust American participation is a fundamentally different organisation than NATO with it.

Europe spent thirty years treating defence as a line item that could be cut without consequences. The consequences have arrived. Rebuilding credible military capacity will take a decade, cost hundreds of billions, and require institutional innovations that the EU has not yet imagined.

Editorial observation

The ReArm Europe plan — later rebranded "Readiness 2030" — proposed in early 2025, attempted to bridge the gap between ambition and fiscal reality. It proposed activating the national escape clause of the Stability and Growth Pact to allow member states to increase defence spending without triggering excessive deficit procedures, and proposed up to €150 billion in common EU borrowing specifically earmarked for defence. The proposal was politically significant — it acknowledged that defence spending could not be funded from existing budgets — but the €150 billion figure, while substantial, was a fraction of what sustained defence modernisation would require. Multiple defence analysts estimated that bringing European NATO members to a credible deterrence posture against Russia would require cumulative additional spending of €500 billion to €1 trillion over the next decade. The Commission's proposal was a down payment, not a solution.

Enlargement: the promise and the architecture

The EU's enlargement agenda has been revived with a speed that would have seemed impossible before 2022. Ukraine and Moldova received candidate status in June 2022, a decision driven by the geopolitical imperative of anchoring both countries to the European project in the face of Russian aggression. Georgia's candidate status followed in December 2023, though democratic backsliding has since complicated the relationship. The Western Balkans — Serbia, Montenegro, Albania, North Macedonia, Bosnia and Herzegovina, and Kosovo — have been in various stages of the accession process for over a decade, with Montenegro and Albania furthest advanced. The current enlargement pipeline, if fully realised, would expand the EU from twenty-seven to approximately thirty-five members.

The strategic case for enlargement is strong. A European Union that includes Ukraine, Moldova, and the Western Balkans would extend the zone of democratic governance, market economy, and rule of law across the European continent — denying Russia a sphere of influence and creating a single market of over 500 million people with a combined GDP rivalling the United States. The normative argument aligns with the strategic one: these are European countries whose citizens overwhelmingly aspire to EU membership, and the credibility of the European project depends on keeping the door open to those who meet the criteria.

The institutional case is far more problematic. The EU's decision-making architecture was designed in the 1950s for six founding members, reformed for fifteen in the 1990s, and stretched to accommodate twenty-seven after the 2004 and 2007 enlargements. Each enlargement strained the system. The unanimity requirement in foreign policy, taxation, and certain other domains gives every member state an effective veto — a principle that was manageable with six members and has become a structural impediment with twenty-seven. Adding eight to ten more members without reforming the decision-making process would make the EU effectively ungovernable on any issue where national interests diverge significantly.

The Franco-German expert group report of September 2023, commissioned by the foreign ministers of France and Germany, proposed a set of institutional reforms designed to make enlargement viable: extending qualified majority voting to foreign policy and taxation, creating a multi-speed or "concentric circles" model where core members could integrate more deeply while newer members caught up, reforming the Commission's composition to avoid a body with thirty-five or more commissioners, and revising the EU budget to accommodate the fiscal implications of admitting countries whose GDP per capita is, in some cases, less than a third of the EU average. The proposals were intellectually rigorous. They required treaty change — a process that requires unanimity among existing member states and ratification by national parliaments or referenda. The last attempt at significant treaty reform, the Constitutional Treaty of 2005, was rejected by French and Dutch voters. The appetite for another such exercise is, charitably, limited.

The Commission's approach has been to pursue what might be called "enlargement without treaty change" — a strategy of preparing candidate countries for accession while hoping that the institutional reforms necessary to absorb them will materialise through political will rather than formal constitutional redesign. The 2024 enlargement package proposed a "gradual integration" model, where candidate countries would gain access to specific EU policies and funding instruments before full membership. Ukraine could participate in the single market for certain sectors. Western Balkan countries could access cohesion funds under pre-accession arrangements. The approach is pragmatic. It is also a workaround for a fundamental problem that no workaround can permanently solve: the EU needs institutional reform to function with thirty-five members, and the process for achieving institutional reform is itself one of the things that needs reforming.

AI governance: regulating the unfinished

The EU AI Act was negotiated and largely finalised before the release of GPT-4, before the explosion of generative AI into mainstream use, and before the AI industry's trajectory made clear that the technology's capabilities — and risks — were evolving faster than any regulatory framework could anticipate. This is not a failure of foresight by the Act's drafters. It is a structural feature of attempting to regulate a general-purpose technology during its most rapid phase of development. The AI Act is, in a meaningful sense, a regulation designed for a technology that did not yet exist in its current form when the drafting began.

The Act's risk-based classification system — minimal risk, limited risk, high risk, and unacceptable risk — provides a conceptual framework that is elegant in principle. AI systems used for social scoring are banned outright. Systems used in critical domains like healthcare, employment, and law enforcement face extensive requirements. General-purpose AI models face transparency obligations that scale with their capability. The framework makes intellectual sense. The difficulty lies in applying static risk categories to a technology whose capabilities shift discontinuously. An AI model that qualifies as "limited risk" today may, through fine-tuning, additional training, or novel prompting techniques, cross into "high risk" territory tomorrow — without any change to its underlying architecture that regulators can observe in advance.

The general-purpose AI provisions — Article 52 and the subsequent delegated acts — represent the Act's most forward-looking and most uncertain elements. Providers of general-purpose AI models face transparency requirements: they must disclose technical documentation, comply with EU copyright law, and publish training data summaries. Providers of models deemed to pose "systemic risk" — defined by computational thresholds that currently map to the largest frontier models from OpenAI, Anthropic, Google DeepMind, and Meta — face additional obligations including adversarial testing, incident reporting, and cybersecurity protections. The systemic risk provisions are designed to be technology-neutral and scalable. Whether they are enforceable against companies headquartered in San Francisco, with training data assembled from global internet scrapes, operating compute infrastructure distributed across multiple non-EU jurisdictions, is an open question that the AI Office will need to answer in practice rather than theory.

The competitive dimension of AI governance is the aspect that receives the most political attention and the least honest analysis. European policymakers frequently frame the AI Act as a competitive advantage — a "trustworthy AI" framework that will attract investment from companies seeking regulatory certainty. The argument has some merit: regulatory clarity can reduce uncertainty costs for companies that plan to operate in the EU market long-term. But it coexists with a less comfortable truth. Europe has no frontier AI lab of its own. Mistral AI, the most prominent European large language model developer, was founded in 2023 and has raised significant capital but remains orders of magnitude smaller than OpenAI, Anthropic, or Google DeepMind in terms of compute, training data, and model capabilities. Aleph Alpha, the German AI company, pivoted away from foundation model development in 2024, citing the capital requirements as prohibitive. The European AI ecosystem is populated by application-layer companies — firms that build on top of American-developed foundation models — rather than infrastructure-layer companies that build the models themselves.

This creates a dependency that the AI Act cannot resolve. If European businesses, governments, and citizens rely on AI infrastructure built by American companies, subject to American corporate governance, trained on data curated according to American values, and operated from data centres located primarily in the United States, then European AI regulation governs the terms of access to that infrastructure but not the infrastructure itself. The AI Act ensures that AI systems deployed in Europe meet European standards. It does not ensure that Europe has the capacity to build AI systems that meet any standard at all. The 2024–2029 Commission faces the uncomfortable task of enforcing a world-leading regulatory framework while simultaneously trying to build the industrial capacity that would make European AI sovereignty more than a rhetorical aspiration.

Europe wrote the rules for artificial intelligence before it built the industry. The risk is that it has regulated a technology it does not control, setting standards for products it does not make, in a market where the infrastructure belongs to someone else.

Editorial observation

The convergence

The five agendas outlined above — digital regulation enforcement, industrial policy, defence, enlargement, and AI governance — are typically discussed in separate policy silos, by separate directorates-general, in separate Council formations, and by separate parliamentary committees. This institutional separation reflects administrative convenience, not strategic reality. The agendas are deeply interconnected, and the Commission's success or failure will depend on whether it manages them as a system rather than a collection of independent priorities.

Defence-industrial capacity depends on industrial policy. You cannot build ammunition factories, drone production lines, or missile defence systems without the same permitting reform, supply chain resilience, and investment mobilisation that the Clean Industrial Deal proposes for clean energy. The skills required — advanced manufacturing, precision engineering, materials science — overlap significantly. A European industrial policy that treats defence and clean energy as separate domains will duplicate effort, compete for the same scarce technical workforce, and miss synergies that integrated planning could capture.

AI governance shapes industrial competitiveness. If the AI Act's compliance costs are too high for European AI startups but manageable for American Big Tech, the regulatory framework will consolidate the market in favour of exactly the companies it was designed to constrain. If the AI Office lacks the technical capacity to distinguish between genuine high-risk applications and innovative low-risk ones, it will either under-enforce (rendering the Act symbolic) or over-enforce (chilling European AI innovation). Getting AI governance right is not just a consumer protection question. It is an industrial policy question.

Enlargement determines the future shape of the single market. A European Union of thirty-five members — with Ukraine's agricultural capacity, the Western Balkans' manufacturing potential, and Moldova's strategic geography — is a fundamentally different economic entity than the current EU of twenty-seven. The single market that European industrial policy is designed to leverage would be larger, more diverse, and potentially more dynamic. But only if the institutional architecture can accommodate it. A dysfunctional EU of thirty-five members would be worse for European competitiveness than a functional EU of twenty-seven.

Digital regulation enforcement determines whether Europe's single market is a genuine single digital market or twenty-seven fragmented national markets connected by regulatory text that large platforms circumvent through creative compliance. The DMA's promise — that European app developers, online retailers, and digital service providers can compete on fair terms with global platforms — is an industrial policy intervention disguised as regulation. Its success or failure will be measured not in fines levied or investigations opened, but in whether new European digital businesses emerge and scale in the space that enforcement creates.

The convergence point is fiscal. Every one of these agendas requires money at a scale that the EU's current fiscal architecture cannot provide. Defence modernisation requires hundreds of billions. Industrial policy at IRA-competitive scale requires hundreds of billions more. Enlargement will strain the cohesion and agricultural budgets. AI infrastructure — the compute, data centres, and talent investment necessary for European AI sovereignty — requires tens of billions. The EU budget, at roughly €170 billion per year, cannot fund even one of these priorities at the necessary scale, let alone all five. Something must give: either the budget expands through new own resources and common borrowing, or the ambitions contract to match the fiscal reality, or the gap between what Europe says it will do and what it actually does continues to widen.

Five years, fifty years

The title of this episode is not hyperbole. The decisions that the 2024–2029 Commission makes — and, more importantly, the institutional capacities it builds or fails to build — will establish path dependencies that persist for decades. If Europe builds credible defence-industrial capacity in the next five years, it will have strategic options in 2040 that a disarmed Europe will not. If Europe develops genuine AI infrastructure, it will participate in the technological transformation of the 2030s as a producer, not merely a consumer and regulator. If Europe enlarges successfully, it will be a continental-scale economic and political entity capable of competing with the United States and China as an equal. If Europe reforms its fiscal architecture to enable common investment at scale, it will have the tools to respond to future crises — whatever form they take — with the speed and resources that the current architecture denies.

The inverse is equally true. If Europe fails to enforce its digital regulations effectively, the regulatory model it has spent a decade building will be exposed as an elaborate exercise in rule-writing that changes nothing. If it fails to fund industrial policy at competitive scale, the manufacturing base that employs millions of Europeans and anchors entire regional economies will continue to erode. If it fails to reform its institutions for enlargement, it will either admit new members into a dysfunctional system or break its promise to countries that reoriented their societies toward European membership. If it fails to build AI capacity, it will regulate a technology that shapes every aspect of economic and social life while depending entirely on non-European companies to provide it.

The historical parallel that EU officials invoke most often is the period between 1985 and 1992 — the Single European Act and the Maastricht Treaty, which transformed the European Economic Community into the European Union, created the single market, and laid the groundwork for the euro. That period succeeded because a combination of political leadership (Delors, Mitterrand, Kohl), external pressure (the end of the Cold War, the reunification of Germany), and institutional momentum produced a window of opportunity that was seized decisively. The current moment bears structural similarities: external pressure is intense, the case for deeper integration is compelling, and the consequences of inaction are severe.

But there are differences that counsel caution. The EU of the late 1980s had twelve members, a broadly shared Franco-German strategic vision, and a permissive geopolitical environment that made deeper integration feel like a natural evolution rather than a crisis response. The EU of 2025 has twenty-seven members with divergent interests, a Franco-German relationship that is more strained than at any point since reunification, and a geopolitical environment that is not permissive but threatening. The window for decisive action is open, but it is narrower, and the consensus needed to act through it is harder to assemble.

The Commission cannot control the outcome. It cannot force member states to spend on defence, agree to common borrowing, reform voting rules, or invest in AI infrastructure. What it can do — what it must do — is frame the agenda with sufficient clarity and urgency that the political cost of inaction exceeds the political cost of action. This is a Commission whose legacy will be determined not by the legislation it drafts but by the institutional capacities it builds. Regulations are written on paper. Capacities are built in institutions, industrial facilities, research laboratories, military logistics chains, and the trust between member states that makes common action possible. Paper is easy. Capacity is hard. Europe has excelled at the former for decades. The next five years will determine whether it can master the latter.

The 2024–2029 Commission will not be judged by the legislation it drafts. It will be judged by whether Europe in 2030 has the industrial capacity, defence capability, institutional architecture, and technological sovereignty that Europe in 2024 manifestly lacks. That is the only metric that matters.

Editorial observation

Sources

1. Draghi Report — The Future of European Competitiveness — https://commission.europa.eu/topics/strengthening-european-competitiveness/eu-competitiveness-looking-ahead_en
2. European Commission — AI Act Official Text and Implementation Timeline — https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai
3. European Commission — Digital Markets Act: Gatekeeper Designation and Compliance — https://digital-markets-act.ec.europa.eu/gatekeepers_en
4. European Commission — European Defence Industrial Strategy (EDIS) — https://defence-industry-space.ec.europa.eu/eu-defence-industry/european-defence-industrial-strategy_en
5. Letta Report — Much More Than a Market — https://www.consilium.europa.eu/media/ny3j24sm/much-more-than-a-market-report-by-enrico-letta.pdf
6. Franco-German Expert Group Report on EU Institutional Reform — https://www.auswaertiges-amt.de/en/aussenpolitik/europe/franco-german-group-report/2617206
7. European Defence Agency — Defence Data 2023-2024 — https://eda.europa.eu/publications-and-data/defence-data
8. European Commission — Clean Industrial Deal Communication — https://commission.europa.eu/topics/strengthening-european-competitiveness/clean-industrial-deal_en