Image generated with AI
The cloud and artificial intelligence open up enormous possibilities: faster processes, better services, less complexity. But with every new technology, a different question arises: how do we harness innovation without creating new dependencies? For me, that is the essence of digital sovereignty.
Digital sovereignty is often confused with independence. Sometimes even with isolationism. In my view, neither of these is accurate. It is not about using only European technologies or foregoing global innovation. It is about being able to decide for ourselves which technologies we use, where data is processed, who has access to it, and how flexibly we can change our decisions at a later date. In short: it is about freedom of choice.
And that is precisely why it would be wrong to ignore the innovative strength of the major global technology providers. Hyperscalers have driven massive progress in cloud computing and AI in recent years. We can and should benefit from this pace of innovation. The relevant question is therefore not: Europe or hyperscalers? The question is how we can bring the two together: access to innovation whilst retaining the ability to act independently.
For me, the cloud and AI are key building blocks of the next phase of digitalisation in public administration. Cloud solutions provide the technical flexibility required by modern digital services. AI can ease the workload on staff, analyse large volumes of information and simplify processes.
However, things only really get exciting when we go beyond simply digitising existing workflows and instead fundamentally rethink administrative processes – from the citizens’ perspective rather than that of established structures. The cloud and AI open up new possibilities in this regard. However, to ensure that this potential can actually be realised, we need the right technological foundation.
AI, in particular, relies on powerful and reliably available computing capacity. Access to GPUs and other AI resources is therefore increasingly becoming a strategic issue. Anyone wishing to utilise AI effectively must therefore not only discuss applications and models, but also how the necessary infrastructure is provided and under what conditions it is available. Computing power thus becomes part of the question of how capable and competitive we will be in the digital world of the future.
Image generated with AI
Cloud and AI infrastructures require electricity. A great deal of electricity. This makes energy a key part of the debate on digital competitiveness. If we wish to build up the necessary data centre and AI capacities in Germany and Europe, it must be possible to operate them economically. This requires a reliable energy supply, competitive costs, available grid capacity and reasonable planning conditions. At this stage, technology policy and energy policy can no longer be viewed in isolation. Digital sovereignty does not arise solely from laws, standards or software. It requires physical infrastructure.
For public administration, security, data protection and resilience are, of course, key requirements. But even these are not enough on their own. A sovereign cloud must allow for the possibility of switching. Applications and data should not be permanently locked into a single technological environment. That is why open standards, interoperability and portability are so important.
Open source also has a place in this discussion. Open technologies can create transparency and reduce dependencies. But open source does not automatically guarantee autonomy. Ultimately, it depends on how a technology is operated, who takes responsibility, how security and support are organised, and whether it can be further developed in the long term. For me, therefore, the decisive question is not whether something is open source or proprietary. What matters is whether it preserves freedom of choice.
For me, there is another point that is absolutely essential to the debate on sovereignty: our own capacity for innovation. Europe has strong companies, start-ups and research teams. There is no shortage of ideas. The greater challenge is often scaling them up. How do we ensure that promising technological ideas develop into companies of international significance? This requires markets, capital, computing capacity, data and partnerships. And above all, it requires the ability to rapidly translate innovation into real-world applications.
The public sector, in particular, can play an important role here. Not only as a regulator, but also as a customer and user of new technologies. In doing so, we should not create an artificial divide between European providers and global technology companies. We should harness the innovative strength of the hyperscalers whilst ensuring that European companies have a fair chance to grow and scale their solutions. The two go hand in hand.
For me, therefore, the key question is not whether something is open source or proprietary. What matters is whether it preserves freedom of choice.
In the public sector in particular, the success of AI will not be measured solely by how efficient it is. Decisions must remain transparent. Responsibilities must be clear. Sensitive data requires special protection. And there is one further point that is often underestimated in the AI debate: data quality.
Even the best AI model is of little use if data is unavailable, unstructured or unreliable. That is why we must work together to further develop our infrastructure, data, governance and skills. Technology creates opportunities. Trust is built through the way we use it.
