Internet Protocols, Power and the Rebirth of the Border

It began, as so many myths of the digital age do, with a promise: that this time things would be different. In the early decades of the Internet, engineers and researchers spoke a language of inevitability. Networks would route around damage, protocols would be neutral and connectivity would dissolve borders rather than harden them. Governance—if it existed at all—would be light, technical, and bottom-up. Politics, like friction, would be engineered away.

This faith was not naïve so much as historically conditioned. The Internet emerged during an extraordinary moment: the late Cold War and immediate post-Cold War years, when American power was unrivaled, liberal democracy appeared ascendant, and globalization promised integration rather than rivalry. The Internet was born into what political theorists would later recognize as a temporary suspension of power politics.

And yet, as Thomas Hobbes warned centuries earlier, peace is not the absence of power, but the temporary alignment of interests under a dominant force. The Internet’s early governance worked not because power disappeared, but because it receded into the background—unquestioned, asymmetrical, and largely benevolent. The multistakeholder model was built on that recession.

A world governed by engineers

In the beginning, Internet governance meant coordination, not control. The Internet’s pioneers were engineers and researchers rather than politicians, and they built the network in their own image. Accustomed to distributed systems where no single node holds sway, they engineered reliability through redundancy and peer review. The Internet Engineering Task Force (IETF), the primary forum for this technical class, operated under a simple mantra: "rough consensus and running code." This placed functionality above formal hierarchy. In this meritocratic era, legitimacy wasn't granted by decree; it was earned through what worked, debated in public, and validated by the community.

As the network expanded from a research experiment into a global utility, this culture of coordination began to scale. Domain names, IP addresses, and routing tables were first coordinated through informal agreements among universities and government labs. These arrangements later crystallized into institutions like the Internet Corporation for Assigned Names and Numbers (ICANN) and the Regional Internet Registries (RIRs). Even then, authority was earned, not imposed. Influence flowed to those who demonstrated technical competence, sound judgment, and a willingness to collaborate across borders. Governance was meritocratic rather than hierarchical; in principle and often in practice, anyone with expertise could shape the rules.

This was governance without sovereignty. There were rules but no ruler—structures without a final enforcer. Disputes were settled through persuasion and compromise, not coercion. The network itself imposed constraints: systems that ignored protocols simply failed to interoperate. Yet no political authority stood above it. The Internet functioned as a self-organizing ecosystem, ordered by shared norms and technical interdependence rather than fiat.

In those early days, the network resembled what Hannah Arendt called a space of appearance, where authority arises from participation and power from acting in concert. For Arendt, power was not violence but collective agency. Early Internet governance embodied that insight. No actor could compel compliance, yet cooperation endured. Authority was enacted through credible contributions, peer recognition, and consensus. It was dynamic, contingent, and rooted in shared engagement.

But Arendt also warned that power without institutions is fragile. It endures only so long as people believe in it and continue to act within its framework. The early Internet depended on trust, transparency, and the sense that participation mattered. There were no armies, courts, or binding contracts—only shared commitment. If faith in the system faltered, so would its authority.

This experiment also unfolded in a rare historical moment. Global tensions were shifting, and the United States—though structurally dominant—largely refrained from overt control. Engineers could govern because geopolitical stakes remained, for a time, peripheral to the network’s technical core. The multistakeholder model thrived on ingenuity, openness, and contingency. It was resilient, but never permanent—sustained by belief and restraint.

In retrospect, the Internet’s early governance showed that authority can emerge without sovereignty and power without coercion. It also confirmed Arendt’s warning: such power is delicate, dependent on collective recognition, and vulnerable when its supporting conditions change. The era of the Internet governed by engineers was a brief interlude—before politics and strategic rivalry pressed into the network’s core.

The hidden sovereign

Early Internet governance flourished under a largely unspoken condition: the structural dominance of the US. While bodies like the Internet Society, ICANN and the IETF celebrated borderless, meritocratic ideals, the network’s physical infrastructure and legal foundations were overwhelmingly American. Root DNS servers, core routing systems, backbone providers, and top-level domain administrators were largely based on US soil and subject to US law. The Internet appeared global, but its jurisdictional anchor was national.

This position rested on a powerful innovation ecosystem. DARPA funded foundational research; universities such as Stanford University and Massachusetts Institute of Technology forged early protocols; Silicon Valley translated them into scalable technologies. The commercialization of TCP/IP, the development of routers and semiconductors, and the rise of cloud platforms all emerged from this nexus of public funding and private capital. The Internet’s architecture carried the imprint of American law, markets, and entrepreneurial culture.

Yet for decades Washington exercised restraint. It did not dictate protocol design, seize the DNS root, or micromanage multistakeholder institutions. Instead, it allowed the system to function as a global commons—while retaining the ultimate capacity to intervene. Sovereignty existed less in daily administration than in latent authority.

The dynamic resembles the theory of sovereignty articulated by Carl Schmitt’s concept of the sovereign of exception. In Schmittian terms, a sovereign is someone who decides on the exception. Power lies not in routine governance but in the ability to suspend it. During the Internet’s formative years, the US occupied that position. It rarely acted, but it could have. Its authority was omnipresent precisely because it remained mostly invisible.

The 2016 IANA stewardship transition illustrates this paradox. Oversight functions were formally transferred from US government supervision to the global multistakeholder community. Yet the transition was carefully managed and legally structured within US jurisdiction. Restraint signaled trustworthiness, but structural leverage persisted. Sovereignty was not abolished; it was deferred.

Latent power, however, shapes behavior even when unused. Other states understood the asymmetry. Data localization laws, national routing strategies, and sovereign cloud initiatives emerged partly in anticipation of potential unilateral action. The possibility of intervention—rather than its execution—reconfigured global expectations. Ambiguity itself became a tool of influence.

This tension complicates the legitimacy of early Internet governance. Its authority rested on shared participation and technical consensus, but also on the tacit acceptance of American restraint. The system’s stability depended on a rare alignment: concentrated capability, voluntary self-limitation, and collective belief. If that belief wavers—through geopolitical rivalry or technological centralization—the equilibrium frays.

The early Internet was thus governed not in the absence of sovereignty, but in its shadow. The hidden sovereign enabled cooperation by declining to rule directly, anchoring a distributed system to a concentrated reserve of power. Yet the very invisibility of that reserve made it both stabilizing and fragile. The Internet’s openness was real—but it was never fully detached from the structural authority that made it possible.

The power of possibility

The central challenge to global Internet governance today is not American overreach, but the architecture that makes overreach unnecessary—and, when convenient, effortless. The threat is structural and anticipatory. The world fears less what the US does than what it could do within a system that already embeds asymmetrical leverage. That duality—restraint alongside ready capacity—generates more anxiety than overt intervention ever could.

Power in international politics operates through anticipation as much as action. The Internet’s infrastructure, despite its decentralized rhetoric, is materially concentrated: root servers, hyperscale cloud platforms, undersea cables, semiconductor supply chains, and major content networks cluster within a small set of firms and jurisdictions, with the US at the apex. The ideology of an open, borderless network persists, but structural dependencies tell a different story.

Washington need not seize the DNS or nationalize infrastructure to shape global outcomes for instance. Influence is embedded in architecture: jurisdiction over key firms, extraterritorial enforcement of domestic law, sanctions regimes, export controls on advanced chips and AI models, and the policy alignment of globally dominant companies. The deeper source of leverage is optionality—the standing capacity to act. These tools need not be constantly deployed; their demonstrated availability generates anticipatory compliance. When rules are enforced occasionally but the system is built to support them constantly, power becomes a permanent part of the foundation rather than just a series of one-off events.

This is power as architecture. As Michel Foucault observed, modern power often operates through systems of constraint and dependence rather than visible decree. Influence functions like gravity: rarely announced but constantly shaping trajectories. States localize data and build sovereign cloud systems not only in response to past actions, but in anticipation of possible ones. Firms align with US policy because the cost of misalignment is potentially catastrophic. Compliance becomes preemptive.

Recent history illustrates the pattern. Sanctions enforcement has leveraged platform and cloud dependencies to compel global adherence with minimal direct intervention. After the disclosures by Edward Snowden, many governments accelerated data localization—not because coercion followed, but because structural reach had been revealed. Demonstrated capacity proved sufficient.

This logic unsettles traditional theories of governance. Power here is neither formally consented to nor openly resisted; it is internalized and anticipated. The Internet remains technically interoperable, but behavior is increasingly guided by a strategic reflex—hedging against possibility rather than reacting to fact. The multistakeholder model, built on neutrality and trust, strains under conditions where one actor retains unmatched latent capacity.

The result is a governance paradox. Coordination still functions, yet legitimacy erodes as actors orient themselves around potential intervention. In a networked world, the ability to dominate can matter more than domination itself. When behavior is shaped by contingency rather than command, governance becomes a drama embedded in infrastructure—a system steered less by rules than by the shadow of what might happen if they are suspended.

Anticipatory sovereignty

The response to latent power is rarely rebellion. It is hedging.

States are not fragmenting the Internet because the US has abused its dominance. They are fragmenting it because they cannot exclude the possibility that it might. This is not grievance politics but preemption—governance driven by uncertainty rather than outrage, by risk calculation rather than ideology.

Sovereignty is now asserted in advance of violation. Data localization laws illustrate the shift. Governments frame them as privacy safeguards, resilience measures, or protection against foreign surveillance. The language is technocratic but the logic is geopolitical. Data is localized not necessarily because harm has occurred, but because exposure creates vulnerability. The question is not what Washington has done, but what it could do under different political conditions.

National cloud strategies follow the same pattern. States invest in domestic infrastructure not to reject interconnection, but to ensure that dependence never becomes coercive. These policies are described as redundancy and autonomy, not decoupling. The aim is optionality: the capacity to exit if necessary. Sovereignty becomes insurance.

Even democratic allies now speak of “digital sovereignty.” In Europe, the term has moved to the center of policy discourse. This is not sovereignty as exclusion, but as risk management—the recognition that structural dependence, however cooperative, is strategic exposure.

This dynamic inverts the logic of Thomas Hobbes. Hobbes argued that in the absence of a guarantor of security, actors revert to self-help. Today, self-help returns not amid open conflict, but amid functional cooperation. Networks interconnect, standards are negotiated, platforms operate globally—yet trust has thinned beneath the surface.

The Internet was once imagined as post-Westphalian: coordination without borders, norms without sovereignty. Instead, sovereignty has reemerged in defensive form. Borders are drawn less with firewalls than with procurement rules, infrastructure investments, and regulatory frameworks that reassert control over dependencies.

“Anticipatory sovereignty” requires no bad faith. The US need not threaten intervention; institutions need not fail. The mere possibility that infrastructure, law, or platforms could be leveraged in crisis alters behavior now. Trust erodes through contingency, not betrayal.

This exposes a fault line in multistakeholder governance. The model assumes legitimacy flows from participation and that shared technical interests can offset geopolitical rivalry. Anticipatory sovereignty operates differently. It treats governance as latent risk. Cooperation continues—but hedged. Engagement persists—but provisionally.

As Hannah Arendt warned, power endures only while people act together in trust. When trust dissolves, power mutates. In the Internet’s case, it mutates into fragmentation driven less by authoritarian ambition than by democratic anxiety. States recalibrate not because they reject openness, but because they doubt its durability under stress.

The result is incremental reconfiguration. Each localization rule, procurement clause, or national standard appears reasonable in isolation. Together they produce a network that is more territorial, more brittle, and more cautious. No single fracture shatters the system. Instead, strategic friction accumulates. The Internet, once sustained by belief in restraint, is now organized around preparation for its absence.

From innocence to empire

For years, multistakeholder governance operated under political innocence: the belief that participation could substitute for power, that transparency could neutralize asymmetry, that legitimacy required no coercive anchor. This innocence was productive. It allowed the network to scale without diplomatic paralysis. Engineers, firms, civil society, and states coordinated without invoking sovereignty at every turn.

But innocence only lasts as long as power agrees to stay in the background. Bruno Latour warned that technology is never neutral; it simply hides the forces that sustain it. The Internet was shielded not by its own "ethos," but by a temporary geopolitical truce. US dominance was total but restrained, and no rival was yet strong enough to blink. That truce has expired.

Now the mask is off. When China pushes new standards at the ITU, it isn’t "misunderstanding" the Internet—it is weaponizing its architecture. When Russia passes sovereign Internet laws it is asserting that the network is not a commons, but a territory. Even Europe’s rights-based regulations are, at their core, a geopolitical grab for leverage. Values without enforcement are just polite suggestions.

The real question, however, is the one no one wants to ask: What happens when the US stops practicing restraint?

No serious actor believes Washington will abruptly seize the Internet. That caricature misunderstands how power operates in networked systems. The concern is subtler and more corrosive. The US could reshape access through export controls, condition interoperability through security requirements, or redefine acceptable participation through law and sanctions. It could do so incrementally, legally, and in pursuit of legitimate objectives. It would not need to declare a new doctrine or dismantle existing institutions. The fact that it could is enough.

As this realization spreads, innocence dies. Participation becomes a strategy; consensus becomes a ceremony. The Internet community still meets in forums and drafts texts about "openness," but the real authority has migrated. It has moved to national security councils, trade negotiations, and corporate boardrooms. Governance hasn't vanished—it has been displaced.

Hannah Arendt warned that when authority becomes performative, power decays into coercion. What replaces the "engineer’s dream" is infrastructural power: control exercised through chokepoints and supply chains.

What we see is an Empire of Layers. It is a world of conditional inclusion and negotiated connectivity. In this world, multistakeholder institutions will survive, but only as guilds serving larger power systems. They will legitimize outcomes decided elsewhere.

The founding illusion was that coordination could replace power. The reality is that power never left the room. It just waited for the innocence to expire.

Beyond multistakeholderism: governance after innocence

If multistakeholder governance is no longer sufficient, the answer is not its negation. The goal is not a return to state monopoly, nor nostalgia for an era of trust rooted in historical contingency. The task is harder: to imagine governance where power is acknowledged, asymmetry managed, and restraint structured rather than assumed.

The first step is abandoning the fiction that the Internet can function as a commons without guardians. As Elinor Ostrom showed, successful commons endure not through goodwill but through clear rules, distributed authority, and credible enforcement. Applied to the Internet, this means moving from participatory innocence to accountable power. Certain actors—states, platforms, infrastructure providers—possess structural dominance. The challenge is not erasing it, but binding it.

One possible model is layered sovereignty. Technical bodies retain autonomy over protocols, but their decisions would operate within political and legal constraints that limit the exercise of power. Restraint would not be voluntary; it would be institutionalized. Powerful states would accept formal limits on how they leverage infrastructure, law, and companies for unilateral ends—a logic akin to international trade law, where sovereignty is retained but constrained to ensure predictability.

A second principle is diffusion of control. Infrastructure stewardship, legal jurisdiction, platform dominance, and standards influence should not converge in a single actor. Power becomes unaccountable when concentrated; freedom arises from constraint. This aligns with Montesquieu’s separation-of-powers insight: the Internet’s resilience may depend less on shared values than on engineered friction between centers of power.

Third, legitimacy must flow from responsibility as well as participation. Actors who shape outcomes should bear obligations proportional to their influence. Platforms governing speech would answer not just to users but to interoperable oversight regimes. States benefiting from global infrastructure would accept reciprocal exposure to its rules.

Finally, governance must prioritize predictability. In a networked world, the greatest threat is not exclusion, but uncertainty. Authority endures when power is exercised in rule-bound ways, as Max Weber observed. The tragedy of the current order is that power is contingent, optional, and opaque. Future governance must constrain power to sustain cooperation and investment.

This vision will disappoint purists. It lacks the romance of the early Internet and abandons the hope that technical coordination alone can transcend politics. But it does not foreclose alternative futures.

Rather, it engages reality directly: connectivity is political, and governance must accept that fact.

The age of innocence is over. The age of empires is emerging. Between them lies a narrow path: governance that accepts power as inevitable, yet refuses to leave it unaccountable. The Internet’s future will not hinge on whether states govern it, but on whether power—state, corporate, infrastructural—can be made predictable, reciprocal, and bound. That, not openness alone, is the next frontier of Internet governance.