Christine Lagarde: A new age of capital: growth, sovereignty and AI
Speech by Christine Lagarde, President of the ECB, at “Hofburg im Dialog – Economy, Europe, Resilience” in Vienna, Austria
Vienna, 14 September 2026
Let me begin with a story that starts in this city.
In this country, the years before 1873 have a name: the “Gründerzeit”, or “founders’ era”.
The old city walls had come down and the Ringstrasse was rising where they had stood. Railways were pushing out across the empire. New banks and joint stock companies were being founded by the hundred.
But money did not stop at borders. On a scale never before seen, savers in Vienna, Berlin, Paris and London could lend abroad. Their money went into the transformative technologies of the day. French savers financed the Suez Canal. Larger sums still went from British and German savers into the American railways.[1]
The historian Eric Hobsbawm called these decades the age of capital.[2]
In 1873 that age came to an end. Eight days after the Emperor opened the World Exhibition in the Prater, the Vienna stock exchange collapsed, and four months later Wall Street followed.
The crash hit both continents. But what the money had built stayed in the United States, and so did the growth. By the end of the century, the US economy was the largest in the world, while Europe went through two decades of slow growth.
Today, European savers are again financing the transformative technology of the age: artificial intelligence. Again, most of the money is flowing to the United States. Euro area households hold about €440 billion in US technology firms.[3]
The question I want to ask today is whether, this time, Europe will get its share of the benefits. I believe it can. But only if we are clear about why we need this technology, why we cannot simply buy it, and why part of it must be built here.
Why Europe must adopt fast
Let me start with some good news.
Europe is often criticised when it comes to technology: good at running and regulating an advanced industrial society, but weaker at putting new digital tools to work. With AI, the picture is already different. Things are moving.
Firms are investing significantly. Euro area firms will devote around 10% of total investment to AI in 2026, and ECB staff estimate that AI-related borrowing accounted for about a quarter of the growth in credit to firms in the first quarter of this year.
Workers are adopting too. The share of euro area workers using AI on the job has doubled in two years and now exceeds 50%,[4] a level it took the internet about a decade to reach.[5]
Yet progress is not as fast as it could be, as the gap with the United States shows. Over the last two years, US digital investment has grown twice as fast as in the euro area. US workers spend two to three times as much of the working week using AI as workers in the largest euro area economies.[6]
The reason this matters comes down to growth.
An ageing society is straining our social model, and new strategic investment needs are adding to the bill: our analysis finds that existing budgets and EU instruments leave more than €100 billion a year of the public share of those investment needs uncovered.[7] With a workforce that will shrink by more than a million people a year over the next 25 years, there is no way to pay for all this without faster productivity growth.[8]
AI is therefore arriving at the right moment. After years of stalling productivity, it is the best chance we have to make the arithmetic work.
ECB estimates suggest that, if adopted quickly, AI could lift the level of productivity by up to 4% over a decade, which would be transformative for public finances.[9] To put that in context, the Draghi report estimated that a 2% rise in productivity over a similar horizon would cover up to one-third of the fiscal cost of Europe’s strategic investments.[10]
But the case for speed may be even stronger than these estimates suggest, because Europe will pay for this boom whether or not it shares in the growth.
AI investment is now growing so fast, above all in the United States, that the tech companies’ own cash flows no longer cover it. The big US hyperscalers issued more than USD 100 billion in bonds last year, most of it long-dated,[11] and they now account for close to a tenth of new euro bond issuance by non-financial companies.[12]
Measured against the economy, the buildout is not unprecedented, at about half the size of the IT surge of the late 1990s.[13] But it is arriving at a time when government deficits and debt are high, and market participants already attribute part of the rise in US long-term real yields to AI-related borrowing.[14]
Since euro area long-term rates move with US yields, Europe will bear part of the price of this boom in its own borrowing costs. The question is whether it will also get the growth that goes with it.
Why importing is not enough
Adopting, for Europe, means importing.
Here, there is an important difference compared with the world of 1873. The railways were a European invention, and it was the United States that was catching up. The Gründerzeit earned its name because Europe turned the inventions of that age into companies.
Today, the companies are being built elsewhere. Last year the United States produced 59 notable AI models and China produced 35. France and the United Kingdom produced one each.[15] Along the rest of the AI value chain, the picture is the same: the United States hosts three-quarters of the world’s AI computing capacity; Europe hosts 5%.[16]
There is nothing inherently bad about importing rather than producing new technologies. Historically, the gains from great technologies have mostly gone to those who adopted them, and Europe benefits from US investment in AI.
But there are three reasons why AI is special.
The first concerns data.
Much of what a firm knows now lies in its data. That knowledge is what sets it apart from its competitors, and economists have started to value it as an asset in its own right.[17]
Adopting AI means running that data through a system that belongs to someone else and sits under someone else’s law. Cloud computing already did this, but a cloud only stores data, whereas a model reads it. That puts the supplier of a model in a position to learn what an industry knows and to compete with it.
European firms understand this, and without domestic alternatives many will hold back. Eurostat asked firms that had considered AI and decided against it why they had done so. Almost half cited data protection concerns.[18]
The second reason is access.
Past technologies, once adopted, stayed with those who used them. The railway tracks that savers financed in the 1870s could not be taken up again from London or Berlin.
Where Europe has remained reliant on others for inputs or services, the damage a cut-off could cause has been painful but bounded. Halving imports of critical inputs from China would cost us 2-3% of manufacturing value added, for example.[19]
With AI, the loss would be of a different kind. Within a few years it will be screening goods at the border, deciding which tax returns are audited, dispatching trains, watching patients on wards and clearing payments at banks. A withdrawal of access, or a change in its terms, would then reach every sector at once.
That is leverage of a kind no trade partner has ever held over Europe, and it could be used in any negotiation, on tariffs or on digital taxes, for example.
The third reason concerns the frontier.
If we move to a world of many models, including open ones, the threat of a cut-off will diminish. For most of the economy, a model a year behind the best will do the job. But even then the frontier will matter in the most innovative sectors.
Where competition is cut-throat, a small edge in capability can produce outsized gains. Pharma, finance and defence, for instance, all work this way. And this can produce winner-takes-all dynamics: the firm with the best model earns the most from it and can spend the most on the next one, which can entrench its lead for good.
Faced with these properties of AI, Europe has an awkward choice.
Either it holds back on adopting, because it cannot protect its data, and forgoes the growth. Or it adopts AI quickly, becomes highly dependent, and risks losing the freedom to organise its economy according to its own values.
Financing sovereign AI
So what can we do to unite growth with sovereignty?
The foundation is building more European computing capacity. Companies should be deciding whether to adopt AI on its merits, without worrying about where their data ends up.
There is much debate about whether the AI buildout in the United States is a bubble. ECB staff have looked at past technological revolutions, and corrections have been part of the pattern. No one knows whether this one will follow a similar route, but if it does, we will feel the effects here in Europe.[20]
Most European savers’ exposure to US tech giants runs through investment funds, which would have to sell into a falling market to meet redemptions. Making sure our financial system can absorb that is a pressing task for supervisors and regulators. Banks, which the ECB is responsible for supervising, are well capitalised.
None of this is a reason to delay building on this side of the Atlantic. Europe already has too little data centre capacity to meet its own demand, and on current trends, that gap is projected to grow more than sixfold within a decade.[21]
The EU’s new gigafactories are a start, but they will fill only a fraction of it. Europe needs to build at a different pace altogether.
Next, Europe needs models that are “good enough” for most tasks and that run on European infrastructure, so that the threat of being cut off loses its force.
Today such models are mostly released with open weights, and the best of them are currently Chinese, such as DeepSeek and Kimi. But they offer no guarantee: a lab can stop releasing new versions openly whenever it chooses, and the licence terms can change with each one.
Europe has something to build on: Mistral in France has released its latest generation of models with open weights, and a firm in San Sebastián has compressed a Chinese open model into a version that now scores higher on independent benchmarks than any other European model. We need to grow that potential to give Europe a range of open models of its own.
The final step is securing access to the frontier, so that Europe stays competitive in the innovative sectors that will drive the fastest productivity growth.[22]
No country controls the whole AI supply chain today. It runs through China for rare earths, Taiwan for chips, Europe for lithography and the United States for models. What countries can do is protect their access by being indispensable to the frontier.
Some call this holding a chokepoint. I prefer to see it as being a link in the chain that no one can afford to remove.
Europe holds one of the most important links of all in ASML. But indispensability does not last on its own; others will try to build ways around it. When China restricted rare earth exports in 2010, Japan cut its dependence sharply within a few years. China is now spending heavily to replicate what ASML does.
So Europe must keep innovating, to find the links in the chain where it can become indispensable next.
Each step calls for many policies, and I will not list them all. But all of them come back to capital. Europe needs more of it, and it needs a different kind.
More of it, because data centres are expensive. The cost of closing Europe’s data centre gap over the next decade could be as high as €600 billion, chips included.[23]
In the United States the whole of the capital market has been drawn in to make the buildout possible: bonds, private credit and securitisation alongside traditional lending. Europe is still relying largely on its banks.
A different kind of capital, because new models and the links in the chain that keep us indispensable cannot be financed with debt alone. They take equity, and equity willing to carry a firm through years of losses before it earns anything.
The news is better than it was. Mistral’s funding round last week was the largest equity raise ever by a European technology company. But that is one round, and Europe needs many more.
In each case the money is there: European households save around €1.4 trillion a year.[24] What is missing are the markets that channel it to our own uses.
We hear a great deal about the savings and investments union. This, fundamentally, is what it is for. Savings will always go abroad in search of returns, as they have done since the age of capital. But Europe must first be able to deploy its savings at home.
Conclusion
Let me conclude by stressing the opportunity in front of us.
Three years ago I argued that capital markets do not spring up because policymakers call for them. They emerge to finance transformative projects that are beyond the reach of banks.[25]
That is what happened in the 1860s and 1870s, when US railways needed financing on a scale no bank could provide, and capital markets sprung up to channel savings, many of them European, to where they were needed.
I said then that Europe needed a project of its own to drive its capital markets forward. I believe AI is that project. It needs capital on a scale that only markets can provide, and the stakes for Europe could hardly be higher.
Governments across Europe have understood this. The Commission has put its proposals on the table. What remains is to act on them, so that this time Europe’s savings can build Europe’s future.
Thank you.
Legal Disclaimer:
EIN Presswire provides this news content "as is" without warranty of any kind. We do not accept any responsibility or liability for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this article. If you have any complaints or copyright issues related to this article, kindly contact the author above.
