Old Europe, New Fork in the Road

Then one line, dropped almost in passing halfway through, changed how I read it: Europe's delay arrives exactly as the US and China extend their lead in the global AI race. The rest of the piece, Brussels bureaucracy, percentages of recommendations enacted, is the mechanism explaining why that gap keeps widening and nobody closes it.

Mario Draghi is the man who, in 2012 as president of the European Central Bank, promised to do "whatever it takes" to save the euro during the debt crisis, and who later led Italy as prime minister between 2021 and 2022. In 2024 the Von der Leyen Commission commissioned him to write a report on European competitiveness: 383 recommendations, the most authoritative diagnosis the Union has ever received of its own structural lag. Mario Monti followed a similar arc, former EU commissioner, then Italy's technocratic prime minister during the 2011-2013 debt crisis, author himself of a 2010 single-market report that, he now says, may meet the same fate as Draghi's. Martin Lidegaard is Denmark's minister for business and competitiveness, the voice of someone who actually sits at the tables where these reforms get decided, or don't.

Only 15.7 percent of Draghi's 383 recommendations were fully implemented two years after his warning that Europe faced "slow agony." The reforms still stuck aren't the technically hard ones. They're the ones that ask a member state to give up power: capital fragmented across twenty-seven national markets instead of pooled into one. Lidegaard: "there is relatively high agreement about where we need to go, but not on how to get there." Monti, more cynical: recommendations get "a round of applause, which lasts about one day," then every government goes back to unraveling them piece by piece.

Here's where the comparison stops being rhetorical. The five largest US hyperscalers, Amazon, Alphabet, Microsoft, Meta, Oracle, have planned combined 2026 AI infrastructure spending of roughly $660 to $690 billion, nearly double 2025's approximately $380 billion . China trails in absolute terms, around $125 billion in 2025, but individual companies are accelerating fast: Alibaba alone has committed roughly $53 billion over three years. Europe's AI Continent Action Plan puts €200 billion on the table, but €150 of that €200 is private capital, meaning it depends on exactly the financial market integration the Draghi report certifies as barely moving. It isn't that Europe lacks the will to spend on AI. It's that the pipe carrying the money where it needs to go doesn't exist yet, or not at the speed required.

This is the delicate moment in which old Europe could choose to be different, and Germany this week shows both sides of that same coin. On one hand, whoever already holds the capital is bringing it in anyway: the UAE has just committed €40 billion to Germany, €10 billion earmarked for Bavaria alone, aimed at three sectors, artificial intelligence, energy, defense. That's capital arriving from outside to fill a gap European capital can't fill from within, and it arrives with its own priorities already written, not the ones Europe would have chosen if it had been the one putting the money on the table. On the other hand, Germany alone shows that money isn't enough: its 2025-2029 AI program is worth roughly €5.5 billion in public funds, framed as €130 billion once private capital is counted, yet the country is still described by analysts as "well-funded on paper, still catching up in practice," with 30 to 69 % of pilot projects never scaling past testing and only 14 % of German decision-makers comfortable delegating critical processes to AI. The problem was never just finding the money. It's turning it into something that scales.

The same week, an election result in a German state put pressure on Berlin's governing coalition, which confirmed it would press ahead with reforms already underway. That's not a judgment on who won or what it means. It's a fact worth noting for what it says about the mechanism, not the politics. Every cycle of domestic instability is time and attention not spent on structural reform, and that's exactly the kind of distraction that, multiplied across twenty-seven member states, explains why a diagnosis everyone has agreed on for two years is still stuck at execution.

So how could old Europe be different. Not by copying American scale, that requires a capital market Europe doesn't have yet, and forcing it before building one would just repeat the pattern already seen elsewhere: capital concentrated first, questions about who controls it later. The real move available to Europe is in the order of operations, not the speed: use the time fragmentation is imposing on it anyway to build, alongside whatever capital integration eventually arrives, the verification and accountability infrastructure that nobody, not the US, not China, not the Gulf states currently financing Bavaria, has chosen to build first. It isn't the fastest road to catching up. It's the only one where catching up doesn't also mean arriving late to the problem that race has already created everywhere else.