Europe is trying to become a geopolitical power without yet deciding who is going to pay for it.
The old EU political economy was built around the single market, relatively cheap (including Russian) energy, US security, Chinese demand and supply chains, German export surpluses, fiscal restraint and the assumption that geopolitical risk could largely be externalised.
Almost every element of that settlement has weakened. What is replacing it is a much more state-directed, security-driven European economy: defence procurement, energy security, AI infrastructure, strategic industries, critical minerals, subsidies and protection of European production. But the institutional architecture remains half-built.
The numbers show the problem. The Commission expects the EU deficit to rise from 3.1% of GDP in 2025 to 3.6% in 2027, while debt rises from 82.8% to 85.3%. Higher interest costs, defence spending and energy measures are part of the reason. So this is not a return to post-2008 austerity, but neither is there unlimited fiscal room. Europe is entering a period in which governments need considerably more capital at exactly the moment capital has become more expensive.
That gives you the central contradiction. Defence and AI may actually become important engines of European investment and growth, but they also worsen the fiscal struggle over resources. The ECB itself now identifies defence, infrastructure and AI as significant supports for investment; it estimates investment could account for almost 40% of euro-area growth between 2026 and 2028. But Europe still faces the Draghi-scale investment requirement of roughly €750–800 billion extra annually by 2030, before fully incorporating the newer defence burden.
Defence is therefore becoming a de facto industrial policy, not merely military expenditure. SAFE allows up to €150 billion of EU borrowing for member-state defence loans, while Readiness 2030 is intended to mobilise more than €800 billion overall. Procurement rules deliberately favour European, Ukrainian and eligible partner-country production. That means defence expenditure is being used to manufacture a European industrial base, create demand, consolidate suppliers and reduce dependence on the US and others.
But this is where the politics becomes interesting. Six large net contributors – Germany, the Netherlands, Sweden, Denmark, Austria and Finland – have just demanded hundreds of billions of euros of cuts to the Commission’s proposed nearly €2 trillion 2028–34 budget, while simultaneously saying the EU must spend more on defence, competitiveness, migration and sovereignty. They also oppose more common borrowing. That is the European problem in one paragraph: do more collectively, spend less collectively, borrow less collectively.
So Europe is moving towards selective mutualisation: when the threat is large enough, common borrowing suddenly becomes permissible – COVID recovery first, now SAFE – but governments remain unwilling to concede a permanent European treasury. The likely result is repeated emergency instruments.
China pushes the EU in the same direction. China supplied 21.9% of all extra-EU goods imports in Q2, worth €153.6 billion, and China still supplies nearly half of EU rare-earth imports. Europe therefore cannot simply “decouple.” Instead it is moving towards managed interdependence: tariffs, local-content requirements, subsidies, critical-mineral diversification and “Made in Europe” policies. Those policies are already creating friction.
AI shows the same transformation. Europe spent the first phase of the AI era worrying primarily about regulation, but is now increasingly worrying about capacity. The AI Act is being enforced, but simultaneously simplified, while the Commission has launched AI gigafactory plans intended to unlock more than €30 billion in public and private investment. The political-economic question is no longer simply how should AI be regulated? It is who owns the computing capacity, energy, chips, data centres and firms on which European AI will run?
And then there is Ukraine. Europe has already provided about €220 billion in overall assistance, including €88.7 billion for Ukraine’s armed forces, and the EU has agreed a €90 billion loan for 2026–27. Kyiv is already asking for some 2027 money to be brought forward because of a large military funding gap. Meanwhile Poland, Spain, Sweden and the Netherlands are reopening the politically explosive question of using roughly €210 billion of frozen Russian central-bank assets.
So the direction of travel is reasonably clear: The EU is evolving into a security-and-investment space. But it is doing so without having solved the fundamental distributional question: who pays – the taxpayer, national governments, EU borrowers, consumers through higher prices, traditional cohesion/agricultural recipients through budget cuts, or holders of private European savings?
We might not have to wait long to find out.
Leave a Reply