The internet is breaking up – and Europe is the only bloc that knows what to do about it

The global internet is fragmenting. Not through accident or technical failure, but through deliberate sovereign infrastructure choices made by states that have concluded the open internet serves other people's interests better than their own. Four geopolitical shocks – mass surveillance revelations ( Greenwald, 2014 ), the weaponisation of digital sanctions, China’s demonstration of a viable self-contained digital ecosystem ( Bradford, 2023 ), and pandemic-era platform dependence – have accelerated a divergence that was always latent in the infrastructure. The result is a new digital map: China builds inward, Russia fortifies, India non-aligns, the United States securitises. Each is responding to the same structural condition with the tools available to it. Europe's response looks weaker by comparison. No Great Firewall. No domestic hyperscaler (a technology company – Google, Amazon, Microsoft, Meta – that operates global cloud computing infrastructure at massive scale). Russia rehearses full disconnection; Europe has no such option. China built its digital ecosystem over three decades of state-directed investment; Europe missed that window entirely. Critics on both left and right treat this as a failure of ambition or nerve. They are wrong. Europe's regulatory sovereignty model is the most sophisticated and most exportable response to internet fragmentation that any democratic entity has yet produced – and European policymakers have yet to fully grasp what they have built. Why infrastructure sovereignty is a dead end for democracies The infrastructural sovereignty model – build your own stack, control your own routing, own your own platforms – works under specific conditions: authoritarian governance, massive state investment capacity, and decades of deliberate isolation. China's model required all three. Russia's attempt at replication has repeatedly faltered (Bradford, 2023) because democratic accountability, market openness, and infrastructural autarky pull in opposite directions. A liberal economy on a closed network bleeds closure into the economy itself. This is the trap that European voices calling for "digital sovereignty" through infrastructure sometimes miss. Building a European hyperscaler to compete with AWS or Azure means catching up on someone else's terms, at enormous public cost, in a race that started thirty years ago. The EU's own assessments acknowledge this gap without resolving it. Regulatory sovereignty operates differently. It requires setting the terms on which infrastructure operates within your jurisdiction – and making those terms consequential enough that global actors comply rather than exit. Europe has demonstrated, through GDPR enforcement alone, that this works. American and Chinese platforms restructured global data practices to comply with European law. No European hyperscaler was required. Three tools that actually work DNS4EU tends to get described as a resilience measure – a European alternative to American and Chinese DNS resolvers. That framing undersells it. A European public DNS infrastructure enforcing European legal standards at the resolution layer functions as a sovereignty instrument. What is unlawful in Europe becomes unreachable in Europe, without dependence on foreign platform compliance. More importantly, partner states in the Global South – currently choosing between American and Chinese DNS infrastructure by default – can adopt, adapt, and align with European legal standards rather than others. DNS4EU has the potential to anchor a democratic internet governance bloc well beyond Europe's borders. The Digital Markets Act's interoperability mandates are similarly underread. In a fragmenting internet, the gravest practical risk for users and businesses is mutual incompatibility – a world in which data, services, and communications stop crossing digital borders altogether. Interoperability requirements on gatekeeper platforms create friction against that worst outcome. Europe stands alone among major regulatory actors in building interoperability into law at scale, and that matters enormously as other blocs push toward closure. NIS2 and the broader European approach to critical infrastructure security address the coercive dimension of fragmentation – the use of infrastructure dependencies as geopolitical leverage – through mandatory resilience standards and incident reporting obligations that apply to all actors regardless of origin. Where the United States responds to infrastructure security concerns by banning vendors and classifying supply chains (see, e.g., Executive Order 13873, 2019), Europe has built a framework harder to weaponise and easier to harmonise with allies. What Europe needs to understand about what it has – and what it is doing to it Mueller (2017) argues that fragmentation is a political project, not a technical inevitability. DeNardis (2020) shows that control over infrastructure has become inseparable from contests over political and economic power. Patrick (2026) demonstrates that middle powers exercising real influence in a multipolar world do so through credibility, problem-solving orientation, and the capacity to fill vacuums in international cooperation – precisely what Europe's regulatory model offers. The problem is that Europe is now actively dismantling that credibility at the moment it matters most. In November 2025, the Commission unveiled its Digital Omnibus – framed as regulatory simplification, but proposing fundamental changes to GDPR and the AI Act before the latter had even entered full force. Core GDPR concepts were redefined to benefit AI training – in effect loosening data protection requirements that had taken years to establish. High-risk AI obligations were delayed and weakened. The DSA and DMA were flagged for a "digital fitness check" whose scope remained conveniently open-ended. Legal scholars and data protection authorities widely characterised these as substantive rollbacks, not technical simplifications. This happened not in a political vacuum but under sustained US pressure: the Trump administration had designated EU digital regulation as an unfair trade barrier , and the State Department had imposed visa restrictions on European regulators for enforcing the DSA against American platforms. The message from Washington was explicit: back down or face consequences. Some of Europe backed down. That is the self-sabotage – not Europe's failure to build a hyperscaler, but its willingness to hollow out the one instrument of digital power it actually possesses, under pressure from the very platforms its regulations were designed to constrain. The stakes extend well beyond Europe. Democratic mid-sized states across Southeast Asia, Latin America, and sub-Saharan Africa are watching how Europe navigates this moment. Brazil, India, Kenya and others have developed data governance frameworks that draw on European models precisely because those models appeared durable, enforceable, and independent of US corporate interests. A Europe that weakens GDPR to accommodate AI training pipelines, delays AI Act enforcement under American pressure, and treats its own regulatory achievements as negotiating chips in trade talks is not a model worth emulating. It is a cautionary tale. Europe has built the right framework. The argument here is simple: hold the line, enforce it, and export it – before the window closes.