When Canadian Prime Minister Mark Carney took the stage in Davos on January 20 this year, he spoke not as the custodian of a fading rules-based order, but rather as a realist confronting its collapse. The postwar architecture, he warned, no longer commands trust. Economic interdependence has been repurposed as coercion; alliances have grown transactional; and global challenges are met with unilateral reflexes rather than collective resolve. Above all, Carney cautioned against retreating into memory. Nostalgia, he said, is not a strategy.
That warning should be taken seriously by anyone assuming that the open internet will somehow endure on its own.
The open internet — based on bottom-up coordination, decentralization and open standards — was never merely a technical achievement. It was a geopolitical project, sustained by a specific historical alignment between openness, interoperability and a broader liberal order. But that alignment has fractured: as the global political foundations weaken, so too does the network built upon them. What is eroding is not only architecture, but also legitimacy and trust.
Invoking the early internet as proof of inevitable openness is a category error because the conditions that sustained it no longer exist. In today’s environment, nostalgia does not preserve openness; it obscures its vulnerability. The open internet will not survive by default; its survival depends on it being actively defended.
A Fragmenting World, a Fragmenting Internet
Currently, geopolitics is reorganizing around control. The United States is retreating from multilateral commitments while seeking to leverage its technological dominance, while China continues to export a model of digital authoritarianism that integrates infrastructure, governance and political authority. Russia treats information space as a domain of permanent contestation, and even democracies are now justifying digital protectionism through the language of security, resilience and industrial policy.
The result is a self-reinforcing cycle. States fragment the network to protect themselves, which results in accelerating the very instability that they fear. Interoperability declines, costs rise and cooperation becomes harder precisely when global coordination is most necessary.
This moment would have been familiar to Thucydides. In the Melian Dialogue, international order is reduced to its bare essence: the strong do what they can, the weak suffer what they must, he argues. Today, this logic manifests in the pursuit of digital sovereignty. States increasingly treat dependence on foreign technology as a form of strategic vulnerability. Fragmentation, then, becomes a way to reclaim control — to choose the costs of isolation over the risks of subordination. The open internet is sacrificed on the altar of the security dilemma: in trying to make themselves safer through control, states collectively make the system more volatile.
Yet Thucydides’ deeper lesson is that such breakdowns are not inevitable. They are the consequence of choices — above all, the choice to abandon stewardship and mistake inevitability.
The Open Internet as Strategic Infrastructure
The shift from an open internet to a fragmented one represents a move from a global commons to a collection of digital islands.
Historically, the internet enabled permissionless innovation: developers could build, scale and collaborate across borders without navigating sovereign gatekeeping. Fragmentation replaces that openness with compliance moats, favouring firms large enough to absorb regulatory friction. As a result, innovation shifts from a meritocracy of ideas to a contest of geopolitical backing.
The implications for technologies such as artificial intelligence (AI) are especially severe. Democratic and accountable AI depends on diverse data sets, global scrutiny and shared safety norms. A balkanized internet risks a digital Tower of Babel: systems trained on narrow, censored or politically aligned data, insulated from external audit. Accountability erodes as power concentrates within state-aligned technology ecosystems.
Fragmentation also imposes a time tax on global survival. Climate modelling, pandemic response and disaster coordination depend on rapid cross-border data exchange. Firewalls, localization mandates and incompatible standards risk turning automated systems into manual bottlenecks. In fast-moving crises, delay becomes a systemic failure.
As such, the internet is no longer soft power — it is hard infrastructure, akin to shipping lanes or energy grids. When that infrastructure fragments, global cooperation itself begins to fail. Protecting the open internet is therefore not a niche digital concern; it is a prerequisite for maintaining the structural integrity of the international system.
Why Middle Powers Matter Now
The central problem of this era is not the absence of power, but rather the collapse of legitimate stewardship. The United States is retreating from maintaining shared systems, while China offers an alternative built on control, hierarchy and political obedience. In response, many governments have embraced digital sovereignty as reflex and refuge. The result is a vacuum in which the open internet survives as memory rather than strategy.
Middle powers cannot fill that vacuum by mimicking great powers or by defending openness as a moral preference. Their leverage lies elsewhere: in making the open internet indispensable.
The task is not to preserve openness as an ideal, but to re-engineer it as the most functional, interoperable and economically rational system to which one can belong. That shift requires moving from rhetoric to infrastructure and from values to incentives.
First, middle powers should form a standing coalition of interoperability that operates independently of geopolitical alignment. Membership would be based not on ideology but on behaviour: adherence to open standards, cross-border data connectivity, multi-stakeholder governance and due process in digital regulation. Access to shared data flows, digital services and research collaboration would be conditional.
States that diverge from these commitments would face tangible costs: loss of automatic data-transfer adequacy; exclusion from joint AI testing sandboxes and research funding pools; higher compliance requirements for firms seeking to operate across coalition markets; and ineligibility for coalition-backed digital infrastructure financing. Openness would cease to be a norm and become an advantage, while non-alignment would translate into slower market access, higher regulatory friction and reduced investment.
Second, middle powers should make the open internet administratively boring. Fragmentation thrives on friction, so openness should thrive on ease. Mutual recognition of compliance, shared audits, and common sandboxes for AI, data and platform governance would dramatically lower the cost of operating across jurisdictions. The open internet would become the default not because it is virtuous, but because it is efficient. Control-heavy models would reveal themselves as slow, expensive and brittle.
Third, rather than confronting digital sovereignty head-on, middle powers should build exit ramps from it. Sovereignty-compatible but interoperable systems — such as federated identity, modular cloud governance and open-source public digital infrastructure — can preserve domestic authority without breaking the network. This directly undercuts the claim that security requires isolation.
Fourth, middle powers should turn openness into a strategic dependency. By hosting neutral governance bodies, funding critical open-source components and convening interoperability testing, they can make themselves unavoidable nodes in the digital system. Influence would then flow not from dominance, but from indispensability.
Finally, fragmentation must be made visible — and costly. For instance, a coalition of middle powers could adopt a joint digital trade framework that includes an automatic review mechanism: when a participating economy introduces new data localization rules or platform licensing barriers, an independent panel quantifies the added compliance costs, duplicated infrastructure spending and emissions impact, and those findings trigger predefined consequences — such as suspension from shared research programs, exclusion from streamlined digital customs procedures, or loss of eligibility for coalition-backed infrastructure financing. At the same time, jurisdictions that reduce barriers gain expedited certification, preferential procurement access and priority participation in cross-border AI testing environments. Fragmentation would therefore generate tangible commercial and financial disadvantages, while interoperability would produce immediate, visible benefits.
The internet itself offers the lesson. Coordination can be a strategic asset; shared infrastructure can distribute power; trust, once embedded in systems, scales faster than control. Middle powers face a choice: defend openness as principle and watch it erode, or rebuild it as infrastructure and make it difficult to abandon.
