Efforts towards sovereign AI data centers in Europe (and elsewhere) show that the issue of IT infrastructure is now perceived differently by companies, political decision-makers and politicians. Due to the growing importance of enterprise AI, computing capacity is no longer viewed as “just” IT infrastructure, but as strategic infrastructure such as energy or telecommunications.
AI infrastructure as a strategic factor
While having its own infrastructure is just one element of a digital sovereignty strategy, European politicians and policymakers have argued that without domestic data centers, Europe risks becoming dependent on US and Chinese providers of key AI technologies.
This concern is shared by some industry leaders – particularly in finance and regulated industries – who are increasingly looking to AI infrastructure as a foundation for economic security. Specifically, they argue that sovereign data centers enable companies to comply with strict European data protection and AI governance regulations. They say a locally operated infrastructure ensures data remains subject to European law, reducing the risk of being subject to foreign jurisdictions and increasing trust between customers and regulators.
Security and compliance requirements
European leaders also see AI infrastructure as a hedge against geopolitical risks. They believe that reliance on external providers leads to greater vulnerability, whether through legal risks, supply chain disruptions or political tensions.
As Christian Klein, CEO of SAP SE, noted at SAP Sapphire in Madrid last month, many European customers operate in the public sector or other highly regulated industries. “Geopolitical risks are a growing concern,” he said. “What if sanctions suddenly block the flow of data across borders? Or if the latest large language models (LLMs) cannot be used in certain regions?”
Christine Lagarde, President of the European Central Bank, also raised this concern in her November 2025 speech entitled “The transformative power of AI: Europe’s moment to act” highlighted that in critical areas such as data centers and computing power, Europe must “avoid dependence on individual components, the failure of which leads to the failure of the entire system”.
Proponents of a sovereign AI infrastructure also believe it can support broader economic growth. Data centers often support startup, research and industrial application ecosystems, allowing Europe to benefit more from AI infrastructure.
From a technical perspective, proximity is also important. Locally located data centers reduce latency and improve performance for AI applications, especially those that require real-time processing or integration into industrial systems.
But despite these perceived benefits, many European business leaders have called on policymakers to take a more moderate, nuanced approach to dealing with sovereign data. What worries them is not data sovereignty per se, but the way in which it is implemented – primarily through the accelerated construction of new, domestically controlled AI data centers. They emphasize that data residency (location) is just one element of the four cornerstones of a data sovereignty strategy, which also includes legal sovereignty (sovereign control), operational sovereignty (independent operation) and technical sovereignty (data control).
In conversations with policymakers, European executives from various industries warn that reducing reliance on U.S. technology too quickly is unrealistic. This reflects a structural reality: Europe remains heavily dependent on non-European providers of cloud infrastructure, chips and AI platforms.
According to a study by Swiss cloud provider Proton, around 75% of listed European companies rely on US technology services (mainly Microsoft and Google) for critical infrastructure, including email, cloud and software. However, switching providers quickly could result in operational disruptions without offering real alternatives.
Obstacles and concerns
Even the most ardent advocates of a sovereign AI infrastructure admit that there are major practical hurdles to overcome in order to build massive AI data centers in Europe, including in the area of energy. AI data centers are extremely energy intensive, and Europe is already dealing with network bottlenecks, high electricity prices and long approval times.
Without significant investment in energy systems, some European business leaders have warned, new data center projects could be delayed, prove more expensive than planned or be canceled.
Another concern is that sovereignty-focused policies that focus on infrastructure could distort markets. Critics warn that infrastructure subsidies could flow to less competitive domestic providers, leading to slower innovation and a misdirection of resources into politically motivated rather than economically viable projects.
There is therefore a risk that sovereignty will become industrial policy for its own sake, rather than a driver of efficiency or innovation. But perhaps the most important criticism is that the focus on infrastructure distracts from a more pressing issue: the use of AI.
Europe has historically been hesitant to adopt digital technologies. Some executives, such as SAP's Klein, believe it's more important to advance the use of AI across industries, noting that infrastructure alone won't drive productivity gains. If sovereignty overemphasizes infrastructure, this could lead to slower adoption of AI through additional complexity and costs. Klein emphasized that it is a mistake to focus primarily on infrastructure if the development of AI applications and software suffers as a result.
Europe, he recently said, should prioritize investing in applied AI and software solutions rather than infrastructure. At the World Economic Forum in Davos at the beginning of the year, top managers of large European companies such as Capgemini and Ericsson also warned against an approach that was too protectionist. Excluding or restricting global suppliers would increase prices, slow technology adoption and weaken competitiveness.
The company view
From a business perspective, AI is quickly becoming a universally applicable technology, and the costs of AI infrastructure have a direct impact on productivity. If European AI infrastructure is more expensive, European companies risk falling behind global competitors.
While data residency and the other elements of digital sovereignty are essential for some companies operating in sensitive and highly regulated areas, the debate on sovereignty in Europe risks oversimplifying a fundamentally global industry. Or as Henna Virkkunen, head of technology at the European Commission, put it: “No one can be competitive alone.”
However, because AI development depends on globally integrated supply chains including semiconductors, software and skilled workers, a fully localized infrastructure may not be feasible or desirable.
Instead of building additional infrastructure to support AI development, Europe should focus on its true competitive advantage, which could lie in its treasure trove of operational data – a resource that is often difficult to access due to overly restrictive regulations and data access rules. The call for reform is therefore becoming louder among managers across Europe.
Simplifying and standardizing data access rules would help European companies leverage this resource and compete more effectively with international competitors as they embark on the next phase of AI support – the journey to autonomous companies.
