Why Europe could be the real winner of the AI boom

Softbank

01/10/2026 16:39

The author is a Reuters Breakingviews columnist.  The opinions expressed are his own.
By Felix Martin
LONDON, Oct 2 (Reuters Breakingviews) - “America innovates, China replicates, Europe regulates.” So goes the well-known critique which mocks the continent’s failure to build a competitive tech sector. The result is a divergence in growth which means the euro zone’s gross domestic product last year was less than 60% of the US level, down from near parity in 2008. Now American dominance of the artificial intelligence revolution is driving euro-pessimism to new depths. Yet a closer look at the emerging economics of AI reveals a different picture where Europe is well placed – regardless of whether the current boom is sustainable, or not.
Start with the bubble scenario. Gigantic investments in AI infrastructure have sucked up the free cash flows of tech giants like Meta Platforms META.O and Alphabet GOOGL.O. Spending by these so-called hyperscalers is now increasingly financed by a proliferating range of debt supplied by institutional investors. A frenzy of circular financing between chipmakers, cloud computing providers, and AI frontier labs has added a further estimated gross $3.6 trillion of trade credit. Most recently, investment bankers have arrived promoting new derivative instruments to further financialise the underlying investments and redistribute the associated risks.
Parallels with the last great speculative bubble in US real estate before 2008 are hard to miss. Investors recall how its bursting not only pole-axed Europe’s banks but contributed to the continent’s sovereign debt crisis. Money managers fret that the old continent is going to be the scapegoat for American excesses once again.
Yet there is a critical difference between the property and AI booms. In 2007 European institutions were intimately entangled in the asset classes at the heart of the US frenzy. A 2018 postmortem by the Bank for International Settlements found that six European lenders had a 35%-40% market share in underwriting mortgage-backed securities backed by US subprime housing loans. Royal Bank of Scotland was the biggest underwriter of all. German state banks, meanwhile, were notorious as the suckers which ended up holding much of the worthless paper from across the Atlantic.
By contrast, the financial engine powering the AI infrastructure boom is so far an overwhelmingly American affair. Debt issued by hyperscalers has focused almost exclusively on the US market. European exposure is currently negligible. SoftBank’s 9984.T execution last week of the largest high-yield corporate bond issue on record to fund its investment in OpenAI is a case in point. Only approximately a tenth of the $11.1 billion raised by the Japanese tech investor was denominated in euros. Meanwhile, the elaborate vendor-financing superstructure underpinning much AI activity consists almost entirely of contracts between US companies.
Of course, the sheer scale of the AI frenzy means the economic pain of any implosion will be felt around the world. But Europe is almost certain to suffer less, precisely because it has been largely absent from developing and financing AI. Investors seeking protection from the fallout should update their watchwords. Perhaps “America innovates, China replicates, Europe insulates” is more appropriate.
What if the doomsters are mistaken, and the AI boom keeps going? The consensus in that case is that America’s dominance of frontier models and building data centres condemns Europe to economic vassalage. Yet there are reasons why the continent may also outperform in this scenario.
First, it is still far from clear where the real value in the AI technology stack will sit. Spectacular valuations of US frontier labs like OpenAI and Anthropic, which is preparing a stock market listing with a possible $2 trillion price tag, assume that the financial advantage lies in selling metered access at premium prices to state-of-the-art proprietary AI models. Yet cheaper open-source competitors increasingly pose an existential challenge to this notion. Telecom giant AT&T T.N has already switched 40% of its AI use to open-source models, citing 80% cost savings, according to the New York Times. Eight of the top ten AI models by usage in September were open-source, according to Openrouter.
This suggests an alternative future where companies offering specific services running on cheap, commoditised models capture most of the profit from AI. Since this so-called “application layer” does not yet exist in many industries, Europe is not in fact behind.
Indeed, the dominance of US tech giants over the existing online economy may prove a handicap. As Thomas Reiner of Altimeter Capital argues in a recent note, the steep drops in the share prices of online travel agents such as Booking BKNG.O, Expedia EXPE.O, and Airbnb ABNB.O following the launch of Meta’s Muse AI agent last month may augur a more general trend. Agentic AI will disrupt online aggregators across the board, allowing their long-suffering business customers to keep a larger share of the value of not just flights, hotel bookings, and car rentals but potentially all other goods and services sold on online marketplaces. In this scenario, Europe’s failure to build the last generation of tech champions may give it a head start in what comes next.
The final reason to bet on Europe is the simplest but most important of all. If there is one lesson of economic history on which scholars universally agree it is that the real value from technological revolutions does not accrue to the innovating company or country. Instead, it is the users of new technologies, wherever they are, who harvest the overwhelming bulk of the economic rewards. The canonical study by the economist William Nordhaus put the proportion at nearly 98%. It is AI adoption, more than innovation, that will boost prosperity and growth.
It is easy to assume that because the AI industry is dominated by American firms, the US leads in AI usage too. Yet Microsoft’s June 2026 Global AI Diffusion report found it ranks only 21st globally. European countries, meanwhile, account for six of the top ten worldwide. Europe has indeed lagged in inventing AI. But it currently has the edge in absorbing and deploying it. In other words: “America innovates, China replicates, Europe integrates”.
Follow @felixmwmartin on X

(Editing by Peter Thal Larsen; Production by Pranav Kiran)
((For previous columns by the author, Reuters customers can click on MARTIN/))