This is exactly what we are discussing in the European Economic and Social Committee (EESC) while considering the new European Commission proposal – the Cloud and AI Development Act (CADA).
Read more The grandson of Garastas revealed the coach’s wish spoken before death
Experts agree on the problem, but not on the solution. Joseph Jarnecki, an expert from the analytical center RUSI, assesses Europe’s technological dependence as a strategic vulnerability. However, Mikołaj Barczentewicz from the International Center for Law & Economics warns that technological sovereignty should not turn into technological isolation – a fully European chain of the most advanced cloud and AI technologies is not realistic today.
An even more sensitive issue is defense. The “Financial Times” wrote about European defense officials’ concerns that too rapid US technology restrictions might weaken rather than strengthen European military capabilities and NATO interoperability.
Thus, a formula emerges: strategic autonomy must not mean strategic autarky. Europe’s goal should not be to disconnect from American technologies. The goal is to have enough alternatives so that Europe can choose when to use them, rather than being forced to do so because it has no other choice.
This is a matter of freedom of strategic choice.
Two data centers alone are not enough
What is Europe trying to do? This is exactly the context in which CADA appears.
The European cloud market is growing rapidly, but the share of EU providers has decreased from 29% in 2017 to about 15% in 2022. More than 70% of the market is controlled by three major non-EU cloud service providers.
This does not mean that foreign technologies are bad. The problem is concentration and dependence. Therefore, the Commission proposes to significantly increase Europe’s computing capacity over the next year, accelerate the development of data centers, and create a common cloud sovereignty system. The more sensitive the state function, the more important the requirements for jurisdiction, data control, and operational autonomy would be.
This principle of risk-based sovereignty seems rational to me. Not everything requires maximum sovereignty. However, for critical state, infrastructure, and defense functions, Europe must have a real alternative.
At the same time, it is important not to confuse European computing infrastructure with European technological sovereignty. We can build three times more data centers in Europe, but what good is that if non-European processors, platforms, software, and cloud services continue to dominate inside them?
Therefore, the EESC proposes to use one of Europe’s greatest levers more actively – public procurement, giving more importance to European added value and at the same time opening more opportunities for European small and medium-sized enterprises.
The logic is similar to discussions about the European defense industry: if Europe wants to have its own technology industry, it is not enough to subsidize it. It needs to be given a market.
AI strategy is also an energy strategy
Data centers do not live “in the cloud.” They need a lot of electricity, as well as networks, land, water, and telecommunications infrastructure.
The EESC warns that the ambition to significantly increase data center capacity may remain on paper if electricity production does not keep pace with their demand. Faster grid connection does not create more electricity.
Read more Unseen success: a Lithuanian found a treasure in Latvia
Therefore, Europe’s AI strategy inevitably becomes an energy strategy. This is especially relevant for Lithuania if we want part of the new European computing infrastructure to appear in our region. This will require not declarations, but electricity, network capacity, fast permits, and appropriate telecommunications infrastructure.
Resilience – not just reliable technology
Another risk raised by the EESC is “silent erosion.” The more functions we transfer to AI, the greater the likelihood that people and institutions will gradually lose the ability to perform them independently or verify the results provided by AI.
Therefore, it is important to assess not only the supplier’s reliability but also whether the institution retains the ability to perform or restore a critical function if AI or cloud service is lost. True resilience means not only having reliable technology but also being able to operate when it does not work.
What does this mean for Lithuania?
First, it is an opportunity to attract part of the new European computing infrastructure. We must be prepared in terms of electricity production, networks, permits, and telecommunications infrastructure.
Second, we should more systematically assess our digital strategic dependencies. Which critical state functions depend on a single supplier? Where is the data stored and what jurisdiction applies to it? How many alternatives do we have? What would happen if a critical service suddenly became unavailable? These are no longer IT department issues but national resilience issues.
Third, it is an opportunity for Lithuania’s technology business. If EU public procurement more strongly supports European added value and SMEs, we should think not only about how to become users of new infrastructure but also about which part of the future European cloud and AI ecosystem we can be creators and suppliers of.
The CADA discussion reminds us of a broader lesson Europe is now learning in several areas. For decades, systems were optimized for cost and efficiency. Russia’s war against Ukraine has shown that the cheapest supplier is not necessarily the cheapest when geopolitical risk is factored into the cost.
We have already learned this lesson in energy. We are learning it now in the defense industry. In cloud, data, and AI infrastructure, we still have the opportunity to learn it before a crisis.
But it is important not to go to extremes. Technological sovereignty is not the ability to make everything ourselves. Sovereignty is freedom of choice: the ability to cooperate with the US because it is the best choice, not because we have no other.
Therefore, the CADA discussion is really not just about “the cloud.” It is about a much more important question: will Europe have enough of its own technological capacities and alternatives to control its strategic choices when it really needs to?
