In the 1990s, the dominant story about globalisation was a story about the retreat of the state. As goods, capital and information began to move across borders at unprecedented speed, a generation of thinkers – Kenichi Ohmae announcing “the end of the nation state,” Thomas Friedman celebrating a flattening world – argued that dense economic interdependence would render coercion obsolete. States bound together by trade and finance would find war too costly and sovereignty too porous to defend. Power, in this telling, was dissolving into the frictionless flow of the market.

A quarter of a century later, something closer to the opposite has happened. Interdependence did not disappear, nor did it pacify. Instead it became a weapon. The same networks that were supposed to erode state power have turned out to be among the most potent instruments of coercion in the contemporary system. The claim of this essay is that globalisation did not abolish power politics; it relocated power – out of armies and borders and into the hubs of the networks that globalisation itself constructed.

The subtler theorists always suspected as much. When Robert Keohane and Joseph Nye published Power and Interdependence in 1977, they distinguished between symmetrical interdependence, which binds states equally, and asymmetrical interdependence, which does not. Where one party needs a relationship more than the other, the less dependent party acquires leverage: the ability to impose costs by threatening to withdraw. Interdependence, in other words, was never the antonym of power. It was a new medium for it. What the hyperglobalists missed was not the persistence of asymmetry but its architecture – the fact that in a networked economy, asymmetry is built into the physical and institutional plumbing.

This is the insight that the political scientists Henry Farrell and Abraham Newman crystallised in their theory of “weaponised interdependence.” Global networks – for payments, for shipping, for data, for supply chains – are not flat. They have a topology. A small number of nodes, through accidents of history and economies of scale, become hubs through which a disproportionate share of the world’s flows must pass. The dollar-clearing system, the SWIFT financial-messaging network, the handful of firms that fabricate the most advanced semiconductors: these are not neutral conduits but bottlenecks. And crucially, whichever state holds legal jurisdiction over a hub inherits a form of structural power that no one designed and no treaty granted.

Farrell and Newman identify two ways this power is exercised. The first is the “panopticon effect”: jurisdiction over a hub confers the ability to see, to surveil the flows that pass through it. The second, more dramatic, is the “chokepoint effect”: the ability to deny access altogether, to cut an adversary out of the network on which its economy depends.

The examples are now familiar, though we rarely name the pattern. When the United States and the European Union pressured SWIFT to disconnect Iranian banks in 2012, and again when Russian institutions were expelled after the 2022 invasion of Ukraine, they were exploiting a chokepoint. When Washington enforces secondary sanctions – punishing third-country firms for trading with a targeted state – it is monetising the centrality of the dollar, whose clearing runs through New York and therefore through American law. When the US restricts exports of advanced chipmaking equipment to China, it leans on the fact that a single Dutch company controls the lithography machines without which the most sophisticated chips cannot be made. The map of these chokepoints is, unsurprisingly, a map of pre-existing power: overwhelmingly American, secondarily European and Chinese.

Here the argument connects to a deeper theme in institutional economics. Institutions are never neutral. They encode the distribution of power that produced them, and then, through path dependence, reproduce it. The networks of globalisation were built as efficiency-maximising infrastructures under a liberal order that presented them as apolitical utilities. But control of a chokepoint is a source of what we might call infrastructural rent – a durable advantage extracted not from producing anything, but from sitting astride a flow that others cannot route around. The state that controls the plumbing writes the rules, and collects on them. Globalisation’s promise of a flat world was, in this sense, an illusion of surface. Beneath it lay a steep hierarchy of who could switch whom off.

The implications for weaker states are stark, and they bear directly on the question of self-determination. For most countries, integration into these networks is simultaneously the price of prosperity and the source of a new vulnerability. To join the global economy is to accept dependence on infrastructures one does not control and cannot exit without enormous cost. Formal sovereignty – the juridical equality of states – coexists with a profound material asymmetry in which a handful of jurisdictions can impose their preferences on the rest by manipulating the network. This is why we now see a scramble for exit: China building an alternative payments system, the EU legislating an “anti-coercion instrument,” states across the world pursuing “de-risking,” “friend-shoring” and “digital sovereignty.” These are attempts to escape the chokepoint – and their limited success so far is itself a measure of how concentrated the power has become.

The liberal hope was that interdependence would make the state obsolete; the realist fear was that interdependence was fragile and would collapse back into autarky. Weaponised interdependence describes a third and more unsettling reality: interdependence endures, but it has been captured. The network is at once the thing that binds states together and the weapon they wield against one another. If the twentieth century located power in territory and force, the twenty-first has quietly relocated it into the hubs of the world’s economic infrastructure – a redistribution that no one voted for, that maps onto old hierarchies, and that may prove more consequential for the shape of global order than any formal treaty of our age.