After briefly commenting on the EuroStack, politics have continued to escalate, warranting another, slightly longer post on the broader European perspective. What is the policy context, what kind of sovereignty is implied in the current policy developments, and what is to make of this?
Europe’s ambiguity: what exactly is the problem with the US authoritarian turn?
European policy-makers have been shocked by how fast the second Trump administration pushed ahead with authoritarian policies. They respond with fear and indignation, in particular to the imperialist ambitions of the US, but also with an interest to follow suit and apply related policies in the EU, in particular regarding the role of the state, most prominently embodied by Elon Musk’s DOGE (Department of Government Efficiency). In this ambiguity, it becomes clear that European countries are becoming more united in the face of the US aggression; however, their main concern seems to be European economic interests rather than authoritarian views seeping from the US into European politics.
Zooming in on the military: US-American politicians like Trump and JD Vance have openly threatened to withdraw from NATO, the military defence alliance. In addition to this, Europeans were alarmed both by Trump’s ambitions to annex Greenland and by the level of resentment towards Europe that top officials expressed in their Signal chat on a military strike on the Houthi in Yemen.
It is not the first time that EU countries are becoming aware that the USA is becoming a less reliable ally. They already increased military spending by 30% over the past three years, exceeding €300bn/year. But what has happened in the past month is quite staggering: The German parliament agreed on a complicated reform of the debt brake to remove the limits from military spending, with €500bn a key figure that is being discussed for the medium term. The EU wants to spend an additional €800bn on defence (but are still negotiating on how to fund this). Is this the investment into public infrastructures that has been called for? Well, investment it is, and there is hope for dual-use spillovers – that military investments might also create something which might be useful during peaceful times. German universities are rethinking their separation between civilian and military research, and this is in sync with Google ending its ban on AI for military purposes in early 2025. But before getting into tech, let’s briefly touch on the role of the state.
Less regulation, less welfare, less migration are slogans that seem to resonate on both sides of the Atlantic.
DOGE and the libertarian state: Less regulation, less welfare, less migration are slogans that seem to resonate on both sides of the Atlantic. While EU countries have, so far, abstained from burning down their public infrastructure as aggressively as Elon Musk has been doing, last year’s Draghi report and recent developments such as the omnibus simplification package echo a deep suspicion of regulation. Instead, “growth and innovation” are the sole focus of politicians, with the UK leading the way. Instead of competition enforcement, the UK Competition and Markets Authority has been tasked with promoting growth, for which its head was replaced by an ex-boss of Amazon UK; the UK government gets its advice on industrial policy from an ex-Microsoft executive, and a secretary of state warned that Big Tech companies need to be dealt with like nation states. And the UK announced it would “mainline AI into the veins” of the nation – all of which is happening under a Labour government. This shift is not confined to the UK: The CEO of Deutsche Telekom called for a European DOGE, the leader of the German Liberal, but now more accurately described as libertarian, Party considers Musk and the anarcho-capitalist, chainsaw-wielding Argentinian president Milei examples of how to cut back public spending, and the Austrian far-right party also wants a “DOGE Europe”.
At the same time, welfare is cut, asylum rules are stretched to and across the boundaries of fundamental rights and climate policies get rolled back: These topics are nowhere to be found in Brussels and many national governments. The key role of the state appears to become to provide an army, while companies are “unleashed” to do everything else.
Tech sovereignty comes in different flavours
There is too much dependence on the goodwill of US politicians and companies, nobody seems to doubt this. This includes public infrastructure like hospitals, schools, public administration as well as private companies and individuals using technology in factories and in the form of digital platforms. But again, what is the key concern of policymakers: dependence on companies somewhere else in the world, or the fact that these companies have become complicit in implementing a nationalist, authoritarian and fascist agenda? The answer to this question guides the type of sovereignty to work towards.
Regulatory sovereignty: governments should be able to enforce rules
Democratic governments should be able to make and enforce laws on companies that provide goods and services to citizens. While this may sound trivial, this was already tricky before Trump because especially digital companies can move around their headquarters and can hide their internal workings from regulators and the public. Now, the European Commission is facing the challenge of how not to irritate a choleric bully when e.g. determining the fines and remedies for non-compliance with the Digital Markets Act. At the same time, regulations like the AI Act and the GDPR are being attacked for allegedly stifling innovation. However, this rhetoric is not only coming from the US; also European companies are keen to push back regulations that are designed to mitigate the harms of tech. Arguably, regulatory sovereignty should empower citizens and consumers vis-á-vis companies.
“Buy European” sovereignty: governments should buy services from businesses based on where they sit
The current EuroStack debate mainly focuses on encouraging everyone, in particular public institutions, to buy from European companies. Not very surprisingly, this idea finds broad support among European companies. It is worth noting that there is no single EuroStack proposal, but at least two – one by Francesca Bria and another one by Cristina Caffarra, but I will not go into the differences here. Using public procurement to pave the way out of Big Tech dependency is very sensible as the EU, national and regional governments have a huge purchasing power that currently mostly benefits Big Tech. Even where they go for “sovereign” options, for example in the cloud, they mostly pay for a local company to sell on a Big Tech product, such as a German ministry that signed a 4-year contract for hundreds of millions with AWS in December 2024 and another company that offers Microsoft cloud products with a 15% mark-up.
To “buy European” certainly makes it easier to enforce laws on EU soil. But it is likely to play into the hands of big European companies that operate on principles that are not fundamentally different from those of the Big Tech companies from which sovereignty is desired. And the strong push for military investment is likely to shape the kind of technology in focus. Buzzwords like “AI” and “quantum” feature prominently both in the EuroStack proposals and in military tech. VC funding for military tech was up by a quarter in Europe in 2024 (with generally falling VC investments), producing “success” stories like that of Helsing “which is developing AI for use on the battlefield” and valued at about €5bn. We know that tech solutionism is harmful, but a push for more military tech will translate into more efficient, more automated killing systems that are an even more direct threat to people’s lives. The fact that more of them might be developed in the EU does not make them less deadly, and it does not reduce the need for strong oversight and public debate and engagement.
While many like to portray the EuroStack as a progressive agenda with conservative appeal, it contains little to prevent it from being captured for a nationalist agenda.
While many like to portray the EuroStack as a progressive agenda with conservative appeal, it contains little to prevent it from being captured for a nationalist agenda. In fact, Cristina Caffarra frames her proposal as a “patriotic effort” and defends this while portraying the EU as colonised by the US. This is coherent with the, in my view, cynical celebration of more “talent” that can be attracted to EU institutions and companies because it fails to recognise that people flee from an increasingly authoritarian regime and deserve a safe haven, irrespective of their value to the European economy. This nationalist framing is uncomfortably consistent with Elon Musk’s attack on empathy. So how do we know if the EuroStack is not becoming a nationalist agenda that would even work under an authoritarian shift, and that currently has some progressive elements sprinkled over it?
Free and open-source sovereignty: governments should buy services based on whether they contribute to an open-source stack/ecosystem
Another way to use public procurement to move away from Big Tech is to embrace open-source principles. What makes this appealing is that there is a generally fairly big community within Europe (even though it is also partly dependent on Big Tech inputs). In addition to public procurement, MEPs like Alexandra Geese are calling for a €10bn European Sovereign Tech Fund modelled on the German Sovereign Tech Agency. This is presented as part of the EuroStack proposal, but it is worth noting that the approach is fundamentally different from “buying European”: The Sovereign Tech Agency does not focus on European projects, but puts money into international open-source infrastructure.
Two main questions are relevant for assessing the current potential of open-source sovereignty:
- First, governments e.g. in Germany and France have long been using the rhetoric of “support open-source”, but government agencies have not put money behind it and continued to buy from Big Tech instead. Will this change? Processes are shaped to slow down such a shift: Procurement rules have been shaped by proprietary tech. Besides, open-source principles are less aligned with the military mission the EU is keen to embark on.
- Second, is open-source sufficient to reduce dependencies? Big Tech already builds closed solutions based on more open infrastructures, thereby monetising products it does not even have the cost of providing. Hence, additional measures to curb power are necessary. Besides, efforts will be needed to prevent open-source from getting even more captured by the Big Tech and the AI industry – this capture has become more severe as e.g. the AI Act turned open-source into a fairly meaningless concept by exempting an overly broad range of so-called open-source AI. What is more, open-source is not necessarily safe from autocratic takeover. That brings us to the last interpretation of tech sovereignty.
Democratic sovereignty: governments should play a role in providing key parts of technological infrastructure
A different approach to sovereignty is to emphasise democratic empowerment and the need to stay within planetary boundaries, as pushed by Cecilia Rikap and others. Public procurement and open-source infrastructures play a key role here, too. But the role of the state and democratic legitimacy goes beyond giving out money, but the aim is to shape technologies and markets to be aligned with ecological internationalism. This means having publicly governed infrastructure which requires the European Commission to go beyond industrial policy and establish e.g. an oversight institution that is independent from national governments and companies.
Government institutions would then need to procure services through this infrastructure, such as a cloud marketplace, to provide the minimum demand to turn it into a viable alternative. And, crucially, they would need to end their existing contracts with Big Tech companies.
Where we end depends not on how we start, but on what is seen as the problem
The different proposals include regulatory sovereignty and are compatible with similar actions like the strategic use of public procurement. European cloud alternatives have long been discussed, and there is momentum to explore real action e.g. in the Netherlands. It might seem easier to start with certain levels of functionality that can be sufficient to serve more local, less digitally sophisticated, but even more sensitive services such as schools, hospitals and public administration.
Nonetheless, the type of sovereignty to ask for makes a big difference to how to proceed after the first steps. Especially “buy European” sovereignty follows a geopolitical framing and is easy to capture for nationalist interests which, in turn, are highly compatible with interests of European big business that are anti-regulation in a way not too dissimilar to Big Tech. Instead, the key concern needs to be dependency on powerful companies that align with fascism and that shape the lives of people across the globe almost unchecked. Only then, the solution can be democratic empowerment that allows for actions that do not almost automatically translate into “more AI” and “more arms”.
To conclude, sovereignty might not be the complete answer to everything. A feminist view would include shaping dependencies, highlighting that sovereignty does not always translate well (“digital sovereignty in Greek is translated as digital dominance (ψηφιακή κυριαρχία).”). Realistically, we need to acknowledge where we stand and what can be done: While European countries are not immune from US tendencies, the authoritarian shift is not a law of nature. To turn back the tide, we need to be active: to deconstruct the surge of military and tech solutionist views and defend what we have, like basic social welfare and human rights enshrined in asylum law. And, if there is still energy left, keep reminding everyone of the importance of democratic empowerment, of the urgency of climate action and of radically needs-based digital policy.