On 18 November, Germany hosted a Franco-German summit to discuss European digital sovereignty. Angela Garcia Calvo and Bob Hancké take a closer look at what a European digital policy would entail.
On 18 November 2025, the French and German governments, led by President Macron and Chancellor Friedrich Merz, held a Summit on European Digital Sovereignty in Berlin. The idea behind the summit was simple and uncontested: since Europe is very dependent on the US and China in most layers of the digital stack, from network equipment to online platforms and everything in between, its industry and economy are vulnerable to any type of shocks emanating from these two powers. Reducing that dependence is a good idea, but the question arises, to what extent is that also possible?
In a series of recent papers on industrial policy and technology, we have developed the contours of a framework to evaluate public policies that aim at revitalising Europe’s industries, including its innovation system. In those texts, we warned that the way may not exist, however strong the political will, and emphasised the need to think well beyond political will and pay attention to existing capabilities and the congruence of policies with institutions.
If existing firms that could occupy an emerging sector are weak, proto-competitive firms are hard to find and the aim of policies are orthogonal to the incentives and constraints that institutional frameworks offer, policies are not very likely to succeed. Demand matters in this process, because new firms need to scale up rapidly to generate the revenue they need to keep up the pace of innovation and stay at the cutting edge. The single market and the important role of public procurement could provide the opportunities needed to scale, though it may require some rethinking of the single market.
Given those considerations, (how) could Europe gain a degree of sovereignty in the “digital stack”? Answering that question requires a closer look at its different layers and the most salient activities in each layer: hard infrastructure, soft infrastructure and applications.
Hard digital infrastructure – too late, but with some opportunities
The digital economy is to a remarkable degree made up of bricks and mortar (and metal and silicon). Semiconductors, networks, high performance computing, quantum computing and increasingly energy are the necessary tangible components of infrastructure that makes the new digital world run. Can Europe find a viable place here, and how?
On mature semiconductor nodes (clusters of integrated circuits) Europe faces a dilemma. They are deemed strategically necessary. Maintaining a presence in that segment may make sense, especially to prevent supply bottlenecks like we saw during the recent lockdowns and to prevent strategic chokeholds à la Nexperia in September 2025. Yet, investment costs in this segment are very high and profit margins very low, especially in a high-wage continent. Investment, public or private, is therefore hard to justify in narrow economic terms and will almost certainly incur significant losses.
Investing in the most advanced logic nodes and going head-to-head with companies like Nvidia and TSMC is out of the question. That part of the industry never really found a foothold in Europe and Europe is in a very weak position, mainly because there are no designers or manufacturers of advanced chips.
However, the semiconductor industry is rapidly evolving and there is an emerging opportunity with the shift to chiplets (small, modular integrated circuits), provided Europe manages to build up an appropriate innovation ecosystem. This would entail leveraging resources from across Europe (including the UK), collaborations outside the EU (especially with the US and Taiwan) and sufficient demand from industry.
The auto industry, particularly the shift to electric, software-defined vehicles, is a promising entry point into chiplets – assuming the European car industry manages to make the industrial turn into electric vehicles a success. Other large-scale European sectors including aerospace, defence and the biotech industry could provide additional demand for chiplets. A broad industrial and security strategy, together with a second EU Chips Act – something that the industry has been asking for – would be steps in the right direction.
In contrast to semiconductor plants, Europe (still) has very strong network equipment manufacturers such as Nokia, Siemens and a few others. Moreover, several countries have made pledges to ban Chinese equipment. But even though Europe appears in a strong position here, infrastructure is typically owned by private operators, not governments, and they base investment decisions on hard incentives, not political will.
Quantum computing is a growth area in which the EU has some initial advantages because of its strong research base and the fact that this field is still in its very early stages. Developing this, however, would require investing considerably more than has been the case, probably at least partly financed through significant long-term public R&D funding.
Soft infrastructure – a few positives, but probably not enough
The second broad area is made up of the underlying software that runs the apps on our computers: cloud computing, browsers, artificial intelligence, search engines, operating systems, online platforms and the like. While there are a handful of bright spots, this does not appear to be a field where Europe can easily muscle its way in. It lacks both a critical mass of potentially competitive firms and the few that have emerged perhaps tell us more about the constraints that all others face (by overcoming them) than about the way forward.
Consider first cloud computing, an area where the large US tech firms are spending serious money, but where European players are very small and losing market share (or maintaining low market shares while the market grows). Moreover, incumbent firms elsewhere on the globe have serious competitive advantages that would allow them to see off a challenge from European companies, unless European policymakers are willing to regulate and subsidise Europeans into existence.
It is not very likely that a US administration of any colour would look favourably upon such blatantly protectionist EU efforts. The idea of building a European cloud is therefore not very realistic. If incumbency is a problem in cloud computing, it is even more of an obstacle in operating systems, given the near monopoly of iOS and Android.
Online platforms are another area where European markets are fully dominated by the US, and, increasingly, Chinese contenders. Opportunities may exist in this space, potentially through a focus on open source alternatives to established contenders, the approach India has taken.
There are, perhaps, also other niches in business software systems on the back of existing networking capabilities that can be explored. In sum, while there are some small sub-areas where European technology forms could excel, the field as a whole appears pretty well stitched up by American and Asian competitors.
Artificial intelligence
A lot has been said about an emergent European artificial intelligence ecosystem. There are a few very promising AI startups, especially in France and Germany. But taken together, China and the US are the undoubted leaders in AI-related activities, with Europe a distant third. The problem in Europe is not the initial ideas but scaling up projects – growing fast enough to occupy a corner in the market from where to grow more and faster.
Succeeding there requires access to large databases and infrastructure to crunch the data. As even a cursory glance at the financial pages makes clear, Europe has none of those large firms that could easily grow. Some of the successful European firms will therefore be bought by US (and possibly Chinese) companies – as has already happened.
This could leave Europe in the frustrating position of becoming an incubator for AI leaders elsewhere. In such a scenario, these companies not only benefit from Europe’s talent and investment in startups but also come back to dominate EU markets, deepening the region’s dependencies.
Industrial policy is often touted as a solution to the problem of AI scalability, but it is not immediately clear how that would work. One viable road could lie in specialised publicly funded databases – in healthcare for example – to which European AI startups would have near-exclusive access.
The developing segment of autonomous vehicles (AVs) could also be a significant source of data, but given the slow rate of progress on AVs and the dismal state of the European automotive industry today, reasonable doubt exists about whether Europe can develop the necessary speed for this to succeed.
In addition, European car firms looking for AVs seem more willing to partner with established US tech companies than European ones. Perhaps some targeted policies to shift incentives for European automakers to invest in European software companies could address that, but they are currently not on the table.
In sum, Europe seems, with very few exceptions, to have missed the opportunity to be an important player in both the hard and the soft infrastructure of the digital economy – at least in its current phase. Something similar can be said about the applications that form the key interface between the digital economy and consumers.
Is this an area for government policy?
Most of us know the digital world through the many apps (or online platforms built on other platforms) that we use on our devices. In a way, this is the easiest area for growth, because the market is open: Europe has many smart ideas and firms face few significant barriers to entry.
But despite these favourable preconditions, there is a dearth of large European apps. There are a few like Spotify and Booking.com, but the absolute number and density of large apps per capita remain very low.
The usual answer to this conundrum is that the EU service markets are very fragmented along regulatory, linguistic and cultural lines, which make scaling up (again) very difficult. While certainly plausible at first sight, this does not explain the success of US and even Chinese apps across Europe.
There is probably some low-hanging fruit here but are government policies the best instruments to exploit these opportunities? Is it really convincing to look at industrial policy, even formulated by a broad committee of industry experts, to identify and develop successful apps when private investors and entrepreneurs have been unable to do so?
Quo vadis, Europa?
In short, there are some areas in the digital stack that show promise: chiplets, possibly quantum computing and some specialised AI applications. In those areas, the market is still open, European tech has competitive strengths and they speak to the broad institutional underpinnings of the European industrial innovation system.
But success will probably happen in collaboration with tech firms elsewhere, leverage existing strong European industries like biotech and the automobile industry, and perhaps rely on access to public databases (which will require some shifts in sensitive privacy regulations).
Despite these few bright lights, though, Europe is likely to remain a small niche player, dependent on others in most areas of the stack, at least in this stage of the digital transition. What may be needed, perhaps, is not a summit to catch up with the rest of the world, but one that identifies long-term future digital growth areas where Europe currently has embryonic strengths.
Note: This article gives the views of the authors, not the position of LSE European Politics or the London School of Economics.
Image credit: EUS-Nachrichten provided by Shutterstock.
Discussion about this post