
Europe likes to imagine that technology can be governed the way a motorway is governed: establish the rules, install the signs, fine the reckless and let everyone get on with their journey. Artificial intelligence is making that fantasy increasingly difficult. The arrival of Chinese models such as DeepSeek has exposed a problem Europe has been postponing: in a technological world splitting into rival ecosystems, sovereignty is no longer simply about where data is stored. It is about who controls the machinery that processes it, the assumptions embedded in it and the infrastructure on which it depends.
The European Union has built an impressive regulatory architecture around digital life. GDPR established data protection as a political principle rather than a bureaucratic nuisance. The AI Act extends that instinct into artificial intelligence, attempting to classify risks and impose obligations before the technology becomes another uncontrollable marketplace. Data sovereignty adds another layer, insisting that Europe should not become merely the geographical location where other powers harvest information.
The difficulty is that regulation cannot manufacture technological independence. DeepSeek is important not merely because it is Chinese, but because it arrived as another reminder that the assumptions of Western technological superiority are not permanent. Europe can regulate a Chinese AI system operating within its borders. It can demand transparency, impose safeguards and restrict particular uses. But it cannot regulate the geopolitical ecosystem from which the technology originates. Nor can it regulate the strategic ambitions of Washington or Beijing.
This is where Europe's balancing act becomes almost absurd. Brussels wants the benefits of American and Chinese innovation while remaining sufficiently independent of both. It wants open digital markets without becoming dependent on foreign platforms. It wants privacy without suffocating experimentation. It wants strategic autonomy without paying the enormous price of rebuilding every layer of the technological stack.
And then there is the Splinternet. The internet was once sold as the great borderless marketplace of ideas. That vision now looks increasingly like a historical curiosity. China has its own heavily controlled digital universe. The United States remains dominated by enormous private technology companies. Europe is attempting to create a third model in which markets remain open but digital power is constrained by public rules. The result could become a genuinely European alternative or simply a heavily regulated island surrounded by technological giants.
The danger is confusing regulation with power. Europe is very good at writing rules for companies it does not own. It is less good at producing the chips, cloud infrastructure, foundational models and platforms that determine how those rules operate in practice. A continent that regulates everyone else's technology but depends on everyone else's technology is not sovereign. It is an excellent customer.
Yet abandoning regulation would be an even greater mistake. GDPR and the AI Act embody something valuable: the idea that technological progress should remain subordinate to citizens rather than the reverse. The answer is not to surrender those principles because the world has become competitive. It is to connect regulation with investment, research, infrastructure and European technological capacity.
Europe therefore faces a choice that is less glamorous than the language of digital sovereignty suggests. It can become a rule-maker with little leverage, a dependent consumer of competing technological empires, or begin building enough technological muscle to make its rules consequential.
The Splinternet is not coming. It is already here. Europe's challenge is deciding whether it wants merely to regulate the new borders or possess enough power to have a say in where those borders are drawn.









