What’s the difference between a sovereign cloud and a sovereign AI cloud? This isn’t the opening of a nerdy joke but an issue that’s galvanizing governments worldwide – particularly in Europe.
Let’s be clear from the outset. The answer to the question is, of course, nothing. Sovereignty means complete control and ownership of the data journey, both from a technical and human capital perspective, in a place that is your own.
AI is simply another computing tool which forms part of the sovereignty package. The real questions are: what is spurring governments into action? How can nations best equip themselves to achieve sovereignty in relation to AI?
Europe’s existential challenge
There has been a well-publicized narrative for some time now that Europe needs digital sovereignty and reduced reliance on US or Chinese hyperscalers. The recent popularization of generative AI has only increased the intensity of this focus.
Countries now see it as an imperative to safeguard innovation and data for the benefit of their citizens. They also want to mitigate the risk of it getting into the wrong hands and being used against them in, for example, a cyber attack.
The need to reduce Europe’s dependency on overseas suppliers is highlighted in the ‘Future of European competitiveness‘ report. It refers to this situation as ‘an existential challenge’.
It’s why the European AI Office – designed to support the implementation and enforcement of the EU AI Act – was created earlier this year. This is also why the EU’s AI Innovation Strategy was developed. Intended as a platform to improve the region’s use of AI on a global scale, it has a public and private investment package of around €4 billion through 2027 dedicated to generative AI.
This desire for AI sovereignty is clear at the national level. Germany’s Sovereign Tech Fund donates money to initiatives that promote cloud sovereignty. Elsewhere, the French government’s Deeptech plan and substantial investment in innovative technologies have significantly strengthened the country’s AI ecosystem. The UK is investing more in AI safety than any other country globally.
AI cyclone that is gathering speed
Nations are in an arms race to equip themselves with the best AI capabilities to protect their interests and achieve sovereignty. However, the rapid advance of the technology is creating a kind of AI cyclone that is gathering speed with governments trying to keep up.
Yet governments are hardly renowned for speed of movement, irrespective of what proactive measures they put in place. This means we have two elements almost diametrically opposite in how they behave. It’s creating a tense and almost clandestine scenario where governments are rushing to fund projects, AI-focused start-ups and introduce regulation. While admirable, this is not going to bring AI sovereignty to life.
The issue was highlighted recently by a collective of technology companies who, in an open letter, say that fragmented regulation in Europe is hindering AI opportunities. In it, the signatories stress that Europe has become less competitive and less innovative than other regions. It also risks falling further behind in the AI era due to inconsistent regulatory decision-making.
Accelerate achieving sovereignty
This need not be the case. Europe can take steps to accelerate its objective of achieving AI sovereignty. It should embrace openness, particularly when it comes to technology. While it may sound counterintuitive in the context of keeping activity within borders, the region needs greater collaboration to avoid ending up as over-regulated silos where companies are limited in development.
Collaboration is needed on several levels. Firstly, AI needs data to be effective. Having a bigger and more diverse data set from across Europe will enable the training and re-training of models to deliver the outcomes we’d all like to see. This won’t work for everything (some data needs to remain sovereign). However, in many cases, data sharing can safely advance AI capabilities without compromise in areas like, for example, international travel, education or sustainability initiatives.
Secondly, some European countries have a GDP smaller than the AI budget of some hyperscalers. On their own, they’re simply not going to keep up, which comprises the collective. In terms of funding, simply putting money on the table and hoping start-ups and private companies come up with the answers is short-sighted.
Governments also need to make their money work harder. AI needs repeat action, so Europe needs a project plan that reflects this and doesn’t blitz all the money early or in only a few companies. This also means investment in infrastructure and the capacity to build hardware. What’s the point in going to extreme lengths to achieve sovereignty only to rely on overseas companies for chips and components?
Openness is the way forward
The only way to achieve sovereignty is through openness. This means open-source software, open data and open protocols. Each EU country needs to join a treaty to create a national alliance to share and collaborate. It will overcome many of the challenges Europe faces in achieving AI sovereignty, especially around cost.
Open software enables nations and companies to create data pipelines that satisfy geographical needs while complying with local regulations. It also enables the freedom to build solutions while mitigating any cloud or solution lock-in. This will fuel faster innovation and create a sovereign Europe we all want to see.
Aiven is a data and AI platform that combines open-choice services to rapidly stream, store and serve data across major cloud providers — simply and securely.
