Strategies and Visions for a possible European Digital Structure
by Fabio Gnassi
In today’s context, the issue of digital sovereignty plays a strategic role, directly impacting the ability of governments and institutions to manage data, infrastructure, and technology independently. According to Fausto Gernone, a competition economist specialising in digital ecosystems, Europe is currently facing a structural delay stemming from decades of neglect of issues that are now increasingly central to its technological and political self-determination
What is the aim of this report and how does it fit into the international context?
Contemporary society increasingly relies on digital technologies. In this scenario, Eurostack emerges as a response to the lack of autonomous digital infrastructure in Europe, highlighting the risks associated with dependence on foreign private companies.
One of the main issues is security. In times of crisis, states tend to utilise all available resources to protect their national interests. During the Covid-19 pandemic, for instance, many governments imposed export restrictions on essential goods such as masks and ventilators. In an era that economists describe as one of “profound uncertainty”, it cannot be ruled out that such restrictions may extend to technological infrastructure. Having indigenous technologies thus becomes crucial not only for strategic autonomy but also for ensuring collective security.
From a technical standpoint, the concentration of supply in the hands of a few companies creates bottlenecks. A striking example was the malfunction in August last year of a security system linked to Microsoft (specifically CrowdStrike), which caused severe disruptions to IT systems of critical infrastructures such as hospitals and airports worldwide.
Another central motivation behind the report concerns competitiveness. Digital technologies are regarded as “general-purpose technologies”, capable of influencing the entire economic fabric across sectors. Possessing a strong digital sector is essential for remaining competitive even in traditional industries, thanks to spillover effects—such as those seen in the interaction between the automotive industry and software development. An advanced digital infrastructure is now a necessary condition for overall competitiveness.
There is also the matter of democratic sovereignty. Digital technologies, which are a core component of ICT (Information and Communication Technologies), have become fundamental to the proper functioning of democracies. The management of technological infrastructure also implies control over information flows. The political support expressed by some American Big Tech CEOs for anti-European positions raises important questions: is it legitimate for the political cycle of a third country to influence the European democratic debate?
Lastly, the report addresses the issue of bargaining power. We are witnessing a shift from a rules-based international order to one based on power relations. In this context, relying on foreign private entities for essential services constitutes a critical vulnerability. One need only consider the role of Elon Musk and the repeated threat to withdraw Starlink support from Ukraine—an act that underscores how risky it is to depend on others’ technologies in an unstable geopolitical landscape.
Eurostack is thus presented as a tool for analysis and proposal—a concrete attempt to map Europe’s technological dependencies and to identify pragmatic solutions to overcome or at least mitigate them.
How did it originate, and why did you choose the “Stack” as your model? Is there a link with the Stack theorised by Benjamin Bratton?
The idea of using the “Stack” as the theoretical framework behind the report stems from the need to analyse Europe’s digital infrastructure systemically. Rather than treating individual components—such as semiconductors, cloud technologies, or software platforms—as isolated compartments, the Stack allows us to contextualise them as part of a complex ecosystem. This approach makes it possible to identify not only critical points—so-called bottlenecks—but also strategic investment opportunities.
The concept of the Stack is explicitly inspired by the work of Benjamin Bratton, who was among the first to propose a layered vision of global technologies. However, while his contribution operates on a theoretical and philosophical level, the approach adopted in Eurostack is oriented towards the definition of concrete economic and industrial policies. The goal is to provide practical tools to address Europe’s technological and industrial challenges.
It is important, however, to highlight the limitations inherent in the Stack metaphor. There is a risk of interpreting this structure as a rigid vertical supply chain, where each level depends solely on the one beneath it. In reality, digital technologies do not follow a strictly hierarchical order, but rather constitute a network of interconnected elements with complex and often unpredictable relationships. For this reason, in the report we opted to visually represent the components not only as layered strata but also as interlocking puzzle pieces, reflecting the dynamic and networked nature of the system.
The link with Bratton’s thinking is further strengthened by the contribution of Haroon Sheikh—professor of philosophy and one of the report’s authors—who helped popularise the concept of the “Digital Stack” in Europe.
During the early drafting phase of the report, other models were also considered, such as those traditionally used in European industrial policy, which tended to treat various domains—from semiconductors to artificial intelligence, and cloud computing—as separate sectors. However, we chose to move beyond this fragmented approach in favour of an integrated and systemic vision.
It is also worth noting that, alongside the Digital Stack model, our analysis draws heavily from the literature on “National Innovation Systems”, which emphasises the importance of considering the innovation system as a whole, rather than focusing on its individual components. Our intention is also to highlight the non-linear interconnections between the various elements of the Stack, emphasising the existence of feedback loops that can generate both virtuous cycles and vicious circles.
From a practical perspective, what are the strategies for achieving digital sovereignty?
The report puts forward several proposals, which are structured on two levels: on the one hand, targeted interventions for individual layers of the Stack; on the other, cross-cutting actions aimed at strengthening the European digital ecosystem as a whole.
Among the layer-specific proposals are strategic partnerships to ensure the supply of critical resources essential for the production of key technologies. At the same time, cross-cutting initiatives are proposed, such as the standardisation of compliance frameworks, simplification of procedures for public-private partnerships, development of regulated federated data platforms, and the establishment of open standards based on open-source principles.
Another central aspect of the report is the coordination of the many initiatives already existing in Europe, which are often fragmented or isolated. The goal is to channel them into a coherent strategic framework, embodied in the Eurostack concept.
A key element of the proposed strategy is the strategic use of public procurement. The idea is to leverage public demand—from various European agencies and institutions—to generate economies of scale in favour of technological solutions aligned with the principles of Eurostack: local, open, decentralised solutions, based on open-source protocols. In this way, the public sector can become a crucial driver in supporting the development of an autonomous European digital infrastructure.
Among the more ambitious cross-cutting proposals is the creation of a European sovereign fund. In the short term, this fund would address the funding gap that hinders many small tech enterprises, helping them to grow and overcome existing bottlenecks in the digital infrastructure. In the long term, the fund would support Member States in achieving strategic objectives related to technological sovereignty and national security. The idea is to begin with an initial capital of €10 billion, scaling up to €300 billion within ten years.
Finally, the report proposes to test the Eurostack concept in practice through the development of a few Minimum Viable Products (MVPs)—trial versions of digital applications and services capable of demonstrating its value. These would be functioning prototypes designed to showcase, on a smaller scale, the positive impact of a European digital infrastructure. Examples of MVPs could include a European digital wallet or federated platforms for the interoperable sharing of health data between Member States.
The rapid pace of technological advancement has made reflection on this topic even more urgent. What is your view on the relationship between nations and artificial intelligence?
In recent years, the progress of artificial intelligence has sparked a wave of enthusiasm among markets and the general public. However, many experts have pointed out that this enthusiasm may be disproportionate to the actual capabilities of current models, which are still far from maturity.
Nevertheless, it is undeniable that AI represents a crucial strategic asset. We are witnessing a profound transformation of the digital industrial landscape—a tectonic shift that will see the decline of certain technologies and the emergence of others. In response to this shift, major tech companies are adapting their business strategies to remain competitive.
Currently, developing AI models involves extremely high fixed costs: servers, energy consumption, intensive water usage, and highly specialised personnel. These are investments that only Big Tech can afford. This has created a barrier to entry that they exploit to their advantage, as emerging startups are often compelled to enter into partnerships that place them in a dependent position.
However, the recent case of the Chinese DeepSeek model has shown that it is possible to develop high-performing solutions at lower costs. The strong impact this had on financial markets is telling, as it suggests that a significant reduction in costs could undermine Big Tech’s dominant position and reduce dependence on them.
That said, the structural dependence on a few dominant actors remains a problem. The real challenge, especially for Europe, is to build an AI infrastructure that is more open, accessible, and inclusive, avoiding a scenario in which innovation is monopolised by industry giants. This is the key challenge we face today.
Given this connection, what is the strategy you propose for the development of artificial intelligence? Could you explain the concept of “composite learning”?
To address growing dependence on Big Tech, Eurostack proposes a strategy based on three key pillars.
The first involves building a European computational infrastructure. This means developing data centres and supercomputers within the European Union to ensure autonomous access to computing resources, without relying on American platforms. This step is fundamental for ensuring technological independence and digital sovereignty.
The second pillar is the diversification of AI training models through technologies such as composite learning. This approach seeks to overcome the centralisation caused by the high fixed costs of training AI models. The idea is to distribute the computational load across multiple entities, rather than concentrating it in a single infrastructure. Composite learning thus makes it possible to process large volumes of data in a decentralised way, promoting a more open and accessible model.
To make this possible, the report proposes the creation of a network of European data-sharing agreements: shared data pools accessible to public institutions, startups, and companies, thereby reducing dependence on datasets controlled by non-European multinationals.
The third element is the adoption of a clear and coherent European regulatory framework that supports decentralised models based on open-source software and limits abuses of dominant market positions by major American or Chinese platforms. There is a need for harmonised strategic industrial policies, geared towards transparency, fair competition, and the promotion of a European innovation ecosystem.
At the heart of this vision lies the idea of AI governance based on interoperability, decentralisation, and open-source principles. This is a complex, but necessary, challenge. As the report suggests, not pursuing this path would mean accepting far worse alternatives.
FAUSTO GERNONE
Fausto Gernone is a competition economist specialising in digital ecosystems. His work sits at the intersection of industrial organisation, information science, and business strategy, centred around a key idea: value creation is a collective phenomenon. For this reason, economic policies must adopt an ecosystem perspective, recognising the interconnections among its various elements.
Fausto is currently completing a PhD at the UCL Institute for Innovation and Public Purpose, under the supervision of Antonio Andreoni and Mariana Mazzucato. He has also conducted research at UC Berkeley, supervised by David Teece. His research analyses competition dynamics among complementary products in digital ecosystems, with a particular focus on coordination mechanisms such as industry standards, APIs, and open-access initiatives.
Before and during his doctorate, he gained several years of experience in industrial organisation, specialising in innovation policy, intellectual property regulation, standard-essential patents (SEPs), antitrust and coordination, mergers, and state aid.
