THREE DECADES ago, a German history professor listed 210 proposed explanations for the fall of the Roman Empire. The remarkable array included such fanciful causes as Bolshevism, public baths, hedonism, the pressure of terrorism and, most famously, lead poisoning.

The last explanation has been discredited. It is highly unlikely that lead water pipes caused the empire to collapse in a tumult of brain damage, gout and madness. Yet exploded theories can point towards important truths. Like classical Rome, America’s empire today depends less on pomp than on plumbing. Instead of roads, aqueducts and seaports, it relies on pipelines conveying financial flows and torrents of data, as well as vast distributed supply chains. These look like a global public infrastructure, but can readily be redirected to private national strategic advantage. America’s domination of obscure, seeming technical structures is generating its own forms of hubristic folly among imperial administrators, who have begun to think that there is nothing they cannot do with it.

Far beneath the boastful speeches, petty insults and spiteful feuds; the fights over NATO and Russian influence operations; the subterranean conduits of empire are failing. If America’s empire is indeed headed towards expiry, many future historians will blame the obvious problems: the rise of China, overextension in the Middle East, the defection of allies. Yet some might trace the beginnings of decay back to two more quotidian crises: America’s botched decision to sanction the Chinese telecommunications giant ZTE Corporation, and Europe’s creation of an apparently innocuous technical arrangement—the “Special Purpose Vehicle”—purportedly to facilitate humanitarian exchange with Iran.

TO CHART the empires of antiquity you simply mapped the cities, rivers and roads through which power flowed. Today’s international economy has a far more complex topography that encompasses both the physical and the digital. Crucial resources—money, information and components—pass through an intricate global tangle of conduits. These started to coalesce into their current configuration a quarter of a century ago, at a moment when the world economy appeared to be secure from empire’s grasp.

The early 1990s seemed the beginning of a new and simpler world where the law of kings had been displaced by the lex mercatoria. The collapse of the Soviet Union and its satellite states created a temporary delirium of liberalism, in which it appeared that bitter international disputes over the ordering of the economy were giving way to the complete victory of the free market. The historian Quinn Slobodian describes how neoliberals had long desired the victory of “dominium”—a global order based on property rights, free global and national exchange—over “imperium”—the power of national states to control property, block financial flows, and harness markets for strategic ends. In the 1990s, it appeared that imperium was being beaten back inside the bounds of national borders. States found themselves increasingly unable to interfere with business investments for fear of capital flight. As national power retreated, a new global order of easy trade and financial and intellectual exchange took its place, underpinned by multilateral institutions such as the World Trade Organization.

In this new era, businesses not only sought to grow and bestride the world, but to remake it in their own image. The great banks and financial corporations aggressively built international capital markets. E-commerce and social media companies constructed global platforms that seemed intended to rebuild national markets and societies on the basis of a deracinated universal liberalism. Sophisticated manufacturing firms like Apple invented new ways of doing business, where the parent company mostly provided intellectual capital, putting out physical production to other firms. All that was solid melted into sophisticated design concepts and supply management software.

This new world, like its self-appointed prophet Thomas Friedman, described itself through a garble of metaphors for flatness, openness and all-conquering market forces. U.S. policymakers believed that illiberal states could be tamed by free information and economic openness, welcoming China into the World Trade Organization. The Internet was by definition untamable: Bill Clinton famously described China’s efforts to control it as like nailing Jell-O to the wall. When the Arab Spring blossomed, it seemed a decisive demonstration of the power of social media to topple illiberal states and replace them with a less oppressive order.

These vast new flows of money, information and manufacturing components required the construction of a correspondingly vast infrastructure. Global finance relied on a complex system of institutions to clear transactions and facilitate messaging and communications between different financial institutions. The Internet was built on routing systems, physical “pipelines” and redundant information storage facilities to move and house data. Complex supply chains needed equally complex networks that drew together a myriad of assemblers, suppliers and sub-suppliers.

These infrastructural systems at first appeared to be technical, dull and apolitical. The businesses and private organizations that built them weren’t particularly concerned with geopolitics. They were rationalizers, who worked carefully and systematically to reduce the myriad complex interactions of the new world into an economically efficient order. This new order made visible what had previously been obscure, and centralized and made tangible what had previously been diffuse and difficult to grasp. By mapping and organizing the world for business, however, it accidentally made it more vulnerable to state power.

Financial institutions wanted to communicate with other financial institutions so that they could send and receive money. This led them to abandon inefficient institution-to-institution communications and to converge on a common solution: the financial messaging system maintained by the Society for Worldwide Interbank Financial Telecommunication (SWIFT) consortium, based in Belgium. Similarly, banks wanted to make transactions in the globally dominant currency, the U.S. dollar. This led them to rely on the U.S. dollar clearing system, which is dominated by a small number of American institutions. Internet communications are in theory supposed to be decentralized by design. In practice, the physical infrastructure, for a variety of efficiency reasons, tended to channel global flows through a small number of central data cables and switch points. Global supply chains, pursuing Adam Smith’s economic logic of specialization, often relied on a tiny number of key suppliers for crucial and difficult to manufacture components, so that a fire in a single factory could bring entire industries to a near standstill.

This did not just transform economies; it transformed international security. A global economy meant that states’ economies were interdependent with each other. This made certain kinds of large-scale coercion less likely—the Chinese and U.S. economies became so intimately entwined that one could not hurt the other without hurting itself. Direct military conflict over Taiwan became more costly as supply chains crisscrossed the strait. However, troubling new vulnerabilities to monitoring, disruption and abuse opened up, as states relied ever more on shared financial, informational and production networks.

Crucially, underneath the emollient rhetoric, this new world was anything but flat. The centralizing tendencies meant that the new infrastructure of global networks was asymmetric: some nodes and connections were far more important than others. These asymmetries mapped onto existing asymmetries in the global economy: the channels and key control points were overwhelmingly located in advanced industrialized states. What this meant was that a few states—most prominently the United States—had the latent ability to transform the global economic infrastructures that had been built to facilitate international exchange into an architecture of global power and information gathering. By seizing control of the key chokepoints in the infrastructure, they could turn this vast infrastructure into a machinery of domination. In the 1990s, business and private actors had beaten back the state to build a vast dominium of their own, which seemed to exclude governments from direct influence. Yet they built it unconsciously so as to allow imperium to come in through the back door.

AFTER THE attacks of September 11, 2001, the United States increasingly turned the global network into a weapon against threats and adversaries. This strategic transformation came about half by accident, in haphazard steps and quarter-steps. Yet its consequences were enormous. Over the last eighteen years, America has quietly created an occulted imperium, quite distinct from the sprawling system of military and naval facilities through which America exerts physical force, based not on high politics, but on control of apparently apolitical plumbing facilities.

This began with the SWIFT financial messaging network. The United States saw SWIFT as a “Rosetta Stone,” which it could use to decipher complex financial relationships amongst terrorists and state adversaries. As Juan Zarate, a former Treasury official saw it at the time in Treasury’s War, “Access to SWIFT data would give the U.S. government a method of uncovering never-before-seen financial links, information that could unlock important clues to the next plot or allow an entire support network to be exposed and disrupted.” The U.S. government quietly obliged SWIFT to provide it with extensive access to financial transaction data, breaking European law (due to SWIFT being based in Belgium) for fear of harsher U.S. penalties. When Europeans discovered this (thanks to a New York Times front page exposé), the United States was able to persuade the European Union (EU) to agree that the United States would continue to have access to the data, so long as it shared security information with European authorities too.