International Relations is still governed by a 17th-century cartography of the mind. It assumes that authority sits inside boxes called territories, that sovereignty is absolute within a line on a map, and that geography is a fixed stage on which states perform. This assumption, call it the Westphalian ontology of absolute space, organized the discipline’s core theories: realism’s balance of power, geopolitics’ fixation on chokepoints and buffer zones, even liberal institutionalism’s talk of “interstate” cooperation. It survived two world wars, the nuclear age, and the internet. It is not surviving artificial intelligence.
The claim of this essay is simple and, I think, unavoidable: the strategic geometry of the 21st-century is being produced by relations among systems, not by the possession of territory. Space itself has become relational, continuously generated by the interaction of algorithms, sensors, orbital assets, supply chains, and autonomous platforms, rather than a static container that pre-exists politics. The discipline needs its own Einsteinian correction: just as physics abandoned Newton’s absolute space for a relational spacetime, understood through the principle of relativity, defined by matter and energy, IR must abandon absolute territorial space for a geometry defined by technological and infrastructural relations. This is not a metaphor. It is a claim about what now actually determines security, deterrence, and power.
Four Symptoms of a Deeper Crisis
Four cases, treated separately, look like unrelated headlines. Read together, they are symptoms of the same theoretical failure.
Greenland. Washington’s renewed insistence on the island’s strategic indispensability cannot be explained by landmass, population, or resource extraction alone. Greenland’s value is positional: it sits at the intersection of missile-warning geometry, Arctic shipping lanes opening under ice retreat, undersea cable routes, and North American ballistic missile defense architecture. Its importance is an artifact of the relationships it anchors, not of the territory it comprises.
Taiwan. Reducing Taiwan to a contested island misses almost everything that makes it central to world politics: it is the fabrication hub for the semiconductors that run every AI system on earth, a link in the first island chain of maritime containment, and a node in overlapping alliance and supply architectures. Sovereignty disputes over the territory are real, but the strategic stakes are relational, rooted in what flows through Taiwan, not merely in who flies a flag over it.
Hormuz. The strait matters because of what it connects, which are energy markets, insurance and shipping networks and financial signalling, not because of who owns its shorelines. A closure would not simply interrupt tanker traffic; it would simultaneously destabilize several global networks that have no territorial existence at all.
Ukraine. This is the clearest case of all. The war is fought as much through commercial satellite constellations, AI-assisted targeting, drone swarms, cyber operations, and sanctions architecture as through the movement of front lines. Territory is contested, but its strategic meaning is generated by the technological relations layered over it. The battlefield became relational before IR theory noticed.
The Ocean Is Becoming Transparent
The clearest laboratory for relational space is the one domain absolute-space theory always treated as an empty void: the ocean. Submarine deterrence has rested for seventy years on a simple bet, that the sea conceals more than sensors can reveal. That bet is now being called.
Submarine-launched autonomous underwater vehicles are already operational rather than speculative: navies have begun deploying AUVs through standard torpedo tubes for reconnaissance, mine-hunting, and seabed mapping without ever surfacing the host submarine, and new seabed-resident drones are demonstrated to loiter and listen for months at a time, turning stretches of ocean floor into persistent passive sonar networks. Anti-submarine systems increasingly fuse acoustic, magnetic, and oceanographic data through machine learning to pick out signatures that previously vanished into background noise, while low-probability-of-intercept acoustic and satellite relay links allow these systems to share what they detect in near-real time. None of this is a single weapon system; it is the construction of an undersea information ecology.
At the same time, the sensor that has always been absent underwater, the GPS, is being replaced rather than merely missed. Quantum inertial sensors, gravimeters, and optical atomic clocks are being trialled on operational and testbed submarines specifically to deliver positioning accuracy without any external signal, immune to jamming or spoofing by design. Surface and airborne trials of the same underlying technology are already treated as a template for undersea deployment. When these systems mature, the ocean’s oldest strategic asset, its opacity, gives way to something closer to a navigable, mapped, contestable network layer, no less relational than the electromagnetic spectrum or the orbital domain above it.
The implication is blunt. A ballistic-missile submarine’s survivability will increasingly depend not on how well it hides in physical space but on its position within a contested information architecture, on whose sensors, whose AI classifiers, and whose relay networks reach it first. Deterrence is migrating from geography to network topology. The most secure vessel in the ocean may soon be the one best integrated into a relational counter-detection architecture, not the one furthest from any coastline.
From Absolute Space to Relational Space
Physics offers the precise analogy IR has never taken seriously. Newton treated space as an empty, “absolute” backdrop against which bodies moved. Einstein showed that space has no existence independent of the relations among mass and energy that constitute it. He called it “relativistic” space-time, in an interstellar or extraterrestrial context, which fuses spatial dimensions and time into a single four-dimensional continuum where geometry changes dynamically with velocity and gravity. Further, a third, “relational” space notion views space and time not as independent containers, but as a network of spatial relations between physical objects and events. This relational notion renders the absolute/Newtonian notion of space untenable or obsolete. IR’s Westphalian imagination is Newtonian: states are bodies, borders are coordinates, and security is measured by distance across a fixed field. What the cases above show is that distance no longer reliably predicts vulnerability, and territorial possession no longer reliably predicts strategic value.
A relational theory of geopolitical space starts from the opposite premise: space has no independent strategic meaning. Strategic space is continuously produced by dynamic relations among technologies, infrastructures, algorithms, and institutions. Several corollaries follow, each with direct theoretical consequences:
· Distance is measured in latency, not kilometers. Hypersonic delivery times and machine decision cycles compress what used to be geographic buffers into single-digit minutes or less, regardless of how many kilometers separate two capitals.
· Borders are algorithmic and infrastructural before they are territorial. Export controls on semiconductors, satellite bandwidth allocation, and undersea cable routing now do more to define the reach of state power than customs posts do.
· Sovereignty becomes positional rather than absolute. A state’s real autonomy depends on where it sits within chip supply chains, AI training infrastructure, cloud computing, and satellite constellations, not merely on the inviolability of its borders.
· Power is the capacity to organize, disrupt, or reshape relationships across domains, not simply the capacity to hold ground.
This does not make territory irrelevant. It makes territory derivative: land and water continue to matter, but their strategic weight is now generated by the relational architectures that pass through, above, and beneath them, rather than by their extent or possession.
What This Demands of Every Major Theory
If the ontology of space changes, every established research program in IR has unfinished business:
Realism must trade territorial balancing for relational balancing: modelling power as a function of network position (chip fabrication share, satellite bandwidth, undersea sensor coverage) alongside traditional military capability.
Liberal institutionalism must treat institutions as dynamic technological infrastructures (standard-setting bodies, satellite-spectrum regimes, AI-safety frameworks), rather than simply as fora for interstate bargaining.
Constructivism must reckon with algorithmic systems as co-constitutive actors in strategic reality: AI targeting systems and classifiers do not merely serve human strategic narratives, they increasingly generate the threat perceptions those narratives describe.
Geopolitics must evolve into what is best called algorithmic geopolitics, in which strategic advantage tracks control over informational relationships (chip design, model weights, orbital slots, undersea cable routes) at least as much as control over physical frontiers.
Security studies must shift from territorial defense to relational resilience. The most exposed states of the coming decades may not be the ones with the weakest militaries, but the ones occupying the most fragile positions in global technological networks, dependent on a single foundry, a single satellite operator, a single subsea cable landing station.
An Unfinished Research Agenda
This reframing opens questions the discipline has barely begun to ask with the seriousness they deserve. What does sovereignty mean when AI systems continuously reconfigure the strategic relationships a state depends on? What does deterrence mean once autonomous systems compress decision windows below any plausible threshold of meaningful human control? How should balance-of-power theory incorporate algorithmic and orbital infrastructure as a distinct category of capability, rather than folding it into vague talk of “technology”? What would neutrality even mean for a state embedded in transnational AI supply chains it cannot fully see, let alone control? And can territorial integrity remain international order’s organizing principle once strategic influence flows as much through relational networks as through recognized boundaries?
These are not thought experiments. They are already being answered, badly, by practice rather than theory, in the Arctic, in the Taiwan Strait, in the Persian Gulf, in the trenches and skies of Ukraine, and now beneath the sea.
A Manifesto for the Discipline
The danger facing International Relations is not that Westphalian concepts are wholly wrong. It is that they remain partially right, right enough to keep being used, wrong enough to keep misleading. Territory still matters. But its strategic significance is now generated by relational architectures that neither begin nor end at any border.
Physics did not discard Newton because Newtonian mechanics stopped working; it moved beyond Newton because new phenomena demanded a deeper ontology of space itself. International Relations stands at the same threshold. It can continue narrating 21st-century competition in 17th-century categories, treating each new case, such as Greenland, Taiwan, Hormuz, Ukraine, the militarized seabed, as an exception to be explained away. Or it can recognize that these are not exceptions at all, but the leading edge of a transformation in what strategic space is.
The next great debate in the field should not be another round of realism versus liberalism versus constructivism. It should be more fundamental: what conception of space is adequate to an age in which artificial intelligence, autonomous machines, and algorithmic infrastructure increasingly determine the strategic geometry of international order? Until the discipline answers that question directly, it will keep interpreting tomorrow’s world with yesterday’s map, and mistaking the accuracy of the paper for the shape of the territory.
