The Death of Defense: From Shield to Sponge

2026-09-21

This article argues that the traditional concept of defense, relying on an impenetrable 'shield' to seal territory, has become obsolete due to the rise of cheap, precise mass weaponry. It proposes a shift from denying penetration to functioning under damage, likening modern defense to a 'sponge' that absorbs impacts while maintaining state functionality, emphasizing survivability over imperviousness.

Introduction: The Wall That Could No Longer Pay for Itself

For as long as states have existed, one image has sat at the heart of how they think about their own survival: the shield. The sovereign state has been imagined as a bounded thing that rings itself with a wall — physical, technological, or political — to bar an enemy from its space. Defense, in this classical sense, was fundamentally an act of sealing. It interposed an obstacle between a protected inside and a threatening outside, so that any attempt at penetration would be stopped at the border or close to it. The lineage of that idea runs the length of military history, from the Great Wall of China to the Maginot Line, from Cold War early warning arrays to the "smart fence" that was meant to hold the perimeter of the Gaza Strip. All of these rested on a single premise, one so basic it was rarely spoken aloud: that a technological and economic threshold existed at which the shield could, in principle, deliver something close to total imperviousness. Beneath that premise lay an even quieter assumption — that a sustainable ratio held between the cost of defending and the cost of attacking. The wall might be expensive, but it would not be expensive by an order of magnitude that could drain the treasury in a matter of days or weeks. The entire conceptual edifice of modern territorial security stood on that unstated ratio.

That ratio has broken. This article argues that the classical, monolithic conception of defense — the shield that seals the territory — has largely dissolved, and that its dissolution is not a temporary embarrassment awaiting the next engineering fix but a structural change in the cost relationship that underlies the whole idea of defending ground. The proximate cause is the arrival of cheap precise mass: drone swarms, inexpensive cruise missiles, loitering munitions, and a widening family of multidomain penetration means that cost a few thousand dollars per unit and can be fielded in numbers no interceptor magazine can match. This is not a quantitative shift that a parallel rise in the shield could offset. It is a reversal of the arithmetic on which sealing depended. The shield can still win nearly every individual engagement and lose the war, because it wins each engagement at a price its adversary sets and its treasury cannot bear.

The consequence is a redefinition of defense that is, I will argue, unavoidable. Defense is migrating from an axis of denying penetration to an axis of functioning under damage — from the shield that keeps the enemy out to the sponge that absorbs what gets in and keeps the state working. The metaphor is deliberate. A shield is judged by whether anything passes it; a sponge is judged by how much it can soak up before it fails. The move from one to the other is the move from imperviousness to survivability, and it is the central claim of this article that this move is neither a defeat nor a choice but the only realistic organizing principle left to states that face cheap precise mass.

This is not, to be clear, the end of defense as such. It is the end of one form of it, and the distinction matters, because the debate tends to collapse into two irreconcilable camps. One announces flatly that "territorial defense is impossible" (Al Habtoor, 2026) and proclaims the disappearance of the "sanctuary" — the protected rear from which a state can act, produce, and command in security (Lange, 2026). A second holds that defense is not dead but reinvented, and that in the current era the defender's advantage at the front is in fact strengthening (Watling, 2023). This article contends that both camps are right at the same time, and that the apparent contradiction dissolves once one sees its central theoretical claim: the offense-defense balance has bifurcated. It has become defense-dominant for holding ground and advancing across land, and simultaneously offense-dominant for reaching into the civilian and military deep rear. These two regimes do not contradict one another; they coexist, and both flow from a single technological substrate — cheap precise mass that punishes exposed movement at the front while saturating point defense in the depth. Defense dies as uniform sealing and is reborn as a distributed fabric of discriminating survival.

The larger conclusion, to which the argument returns at the end, is unsettling. If even Israel — holder of one of the world's most advanced multilayered air defense arrays, backed by a superpower and its stockpiles — is pushed past the economic and operational limit of sealing, then weak and middle powers confront the new equation with almost no answer at all. The Israeli case is not an exception that proves the rule. It is the upper bound of feasibility: if sealing collapses there, it has already collapsed everywhere below it. The implication is not defensive despair but the need to organize national security, from the outset, around a logic of survival rather than the dream of the wall.

The argument proceeds in six movements. It begins with the shield's economic dilemma — the "wrong side of the cost curve" — the deep driver of the change. It then examines the mechanism of saturation and the way it hollows out even very high interception rates, traces the loss of the rear and the end of the sanctuary, and surveys the rapidly evolving fabric of counter-unmanned-aerial-system (C-UAS) technologies, whose very limits establish the claim, along with the related question of "drone walls." It turns next to the Israeli case as an empirical laboratory, from 7 October 2023 to the war of June 2025. Finally, it builds the theoretical frame of the redefinition, confronts the poles of the debate, and draws out the recommendations that follow.

A word of caution belongs at the outset, because some of the most striking evidence is also the most recent and least settled. The figures from 2025 — the scale of the Iranian salvos, and above all the reported crisis in interceptor magazines — remain fluid, contested between official denials and foreign reporting. Where the argument leans on them, it flags the uncertainty, and it distinguishes carefully between the concentrated engagement of June 2025 and the separate, smaller incidents that preceded it.

The Shield Dilemma: The Wrong Side of the Cost Curve

The force driving the death of the classical shield is not, at bottom, technological. It is economic. The sealing conception assumed, if only implicitly, that the cost of interception would bear a reasonable ratio to the cost of the threat it stopped. That assumption has collapsed, and it has collapsed not at the margins but at the core. Its sharpest expression is what may be called the shield dilemma. When an interceptor costing more than a million dollars is fired at a drone costing roughly thirty-five thousand, each successful interception generates a "value loss" of more than six hundred thousand dollars (Rumbaugh, 2024). Tactical success translates directly into economic-strategic failure. The shield wins the engagement and loses the economic war, because the attacker, not the defender, dictates the pace of expenditure and the point at which the magazine breaks. This is a dilemma in the precise sense of the word: every available course of action carries an unbearable cost. If the shield intercepts, it bleeds money; if it withholds interceptors to conserve them, it abandons its target to be struck. There is no cheap exit from the trap so long as the asymmetry between the price of the interceptor and the price of the threat holds.

Observers have described the United States — the richest military power in the world — as sitting "on the wrong side of the cost curve" (Jensen et al., 2025). The phrase does not mean that the shield is technically ineffective; it means that technical effectiveness is bought at an unsustainable price. The cost curve is simply the ratio between the investment the attacker must make and the investment the defender must make to neutralize it. When that ratio tilts consistently toward the attacker — when every dollar he spends forces the defender to spend ten, or a hundred, or a thousand — the inevitable outcome is magazine depletion and the economic exhaustion of the defending side, whatever the momentary interception rate happens to be. This is a genuine historical inversion. For a long era the cost curve favored the defender: a fortification, a wall, or a warning array was expensive to build but relatively cheap to maintain, while breaching it demanded of the attacker an enormous investment of manpower and materiel. Cheap precise mass has reversed the ratio outright. The cheap thing is now the attacker; the expensive thing is the shield.

A dimension of this too often neglected is magazine depth. An advanced interceptor is not only costly to buy but slow to produce. Production of the PAC-3 interceptor, the backbone of the U.S. array against ballistic missiles, has run at roughly fifty-two missiles per month at about $4.97 million per unit (CNAS, 2025). That rate reflects a substantial industrial base and a complex supply chain of precision components, and it is nonetheless dwarfed by the consumption rate of an intense engagement, in which dozens of interceptors may be expended in a single night. The gap between the rate of production and the rate of consumption is precisely the seam along which the shield tears. It is not a technological failure but a failure of the supply and value chain. A defense system may perform flawlessly in every engineering sense and still fail strategically because its magazines empty faster than they can be refilled. The long replenishment time turns any protracted engagement into a race that the attacker, producing his means cheaply and quickly, will very often win.

A dramatic illustration surfaced in the report that "Patriot was empty" at certain points in recent fighting — the system had exhausted its immediate stocks (CNAS, 2025). The U.S. Navy's defensive campaign against Houthi attacks in the Red Sea told the same story from a different arena: expenditure exceeding a billion dollars while interceptor stocks were drawn down to what were described as dangerously low levels (CNAS, 2025). The naval case is especially instructive, because it unfolded in an arena long regarded as the home turf of the technologically superior power. Advanced warships carrying some of the most sophisticated air defense systems in the world were compelled to spend million-dollar interceptors on cheap drones and cruise missiles launched by an armed militia. When a superpower discovers that a few months of medium-intensity fighting can empty its magazines against a non-state adversary, the paradigm on which the shield rests — interception as the primary response — is plainly not sustainable against cheap mass. The problem is not the enemy's strength. It is his cheapness.

Intellectual honesty requires stating the caveats, because the shield dilemma is not an absolute mathematical decree. As the survey below will show, directed-energy technologies promise to break the cost curve at the point of destruction, and cheap kinetic solutions seek to level it in part. Yet even if the cost of the shot itself collapses to a few dollars, the cost of the system, the limits of range, the required dwell time, and the constraints of weather remain. The economic dilemma changes form; it does not vanish. It migrates from the interceptor-cost axis to the system-and-coverage-cost axis, and there it endures. A defensive fabric that provides reasonable coverage across space remains orders of magnitude more expensive than the threat it is meant to stop.

One serious counter-framing deserves to be engaged directly. Mitchell (2026) argues that the entire "cost per intercept" debate is misleading, because an interceptor's value is measured not by its price but by the "value denied" — the harm it blocks. A million-dollar interceptor that stops a strike on a nuclear reactor, a hospital, or a critical infrastructure node is unquestionably a sound bargain; no one measures the worth of a police officer or a firefighter against the damage averted by their salary. This is correct, and it is an important corrective to a crude economic accounting, a reminder that defense is not a business seeking profit but a service preventing catastrophe. And yet it does not refute the shield dilemma; it sharpens it. Precisely because the expensive interceptor must be reserved for high value, it cannot deliver uniform sealing across the whole space. It becomes a scarce resource to be allocated sparingly, and selective allocation is itself an abandonment of the monolithic sealing conception in favor of one built on prioritization and survival. Mitchell's argument does not rescue the classical shield; it describes the new defense, in which a protected and expensive resource is directed at high-value targets while the rest of the space is left to absorption. Value denied is a logic of prioritization, not of sealing.

The Saturation Mechanism: When Ninety Percent Is Not Enough

If the shield dilemma is the economic force undermining sealing, saturation is its operational execution. Saturation is the strategy by which the attacker launches, within a given interval, a mass of means that exceeds the shield's capacity to process and intercept, so that even a highly effective shield is forced to leave "leakers" — means that slip through the net. The logic of saturation exposes the deepest asymmetry of the era. The shield is required to be very nearly perfect; the attacker needs only a handful of leakers to achieve his aim. This is an asymmetry of success requirements, not merely of costs, and it is worth dwelling on because it survives even the most optimistic economic assumptions. Even a shield that is economically perfect — with an infinite stock of costless interceptors — remains subject to physical limits on its processing rate, its number of launchers, and the duration of each interception, all of which cap the number of means it can engage at once. The attacker need only cross that ceiling.

Russian penetration salvos against Ukraine supplied a sharp empirical illustration. In one night a combined salvo of seven hundred twenty-nine means was launched: six hundred fifty-six drones, thirty-three ballistic missiles, twenty-seven cruise missiles, and eight hypersonic missiles (CFR, 2024). The composition itself reveals a sophisticated strategy. The cheap, numerous drones serve as a sponge in the offensive sense — they saturate the system, flood the sensors, disperse attention, and force the shield to expend interceptors — and under the cover of the saturation they create, the fast and lethal means, ballistic and hypersonic, penetrate to their targets. The interception data expose the heart of the problem. Drone interception in that engagement ran at roughly 91.7 percent, a seemingly impressive figure, while interception of ballistic and hypersonic means plunged to around 26.8 percent (CFR, 2024). Precisely the most expensive, fastest, and most lethal means are the ones the shield struggles hardest to stop, and they deplete the stock of expensive interceptors fastest. The combination is devastating in its simplicity: the cheap saturates, the expensive penetrates.

The sharpest formulation of the saturation asymmetry was cast in the Israeli-Iranian context — the observation that "even ninety percent interception is not enough against a ballistic attack on a populated area" (CFR, 2024). The claim is starkly simple and yet fatal to the sealing conception. If a hundred missiles are fired at a city and even ninety percent are intercepted, ten get through, and any one of them may exact dozens of dead and strike critical infrastructure. Sealing assumed that a high interception rate equaled security. Saturation refutes that assumption, teaching that the governing variable is not the relative percentage but the absolute number of leakers, which grows as the mass grows, independent of the percentage. A comparison across Iran's salvos on Israel illustrates the variance directly. The April 2024 attack achieved roughly ninety-nine percent interception; the October 2024 attack registered noticeably greater leakage (CFR, 2024). The variance is itself the point. The shield cannot consistently guarantee the perfection demanded of it, whereas the attacker can consistently guarantee leakers simply by adding mass. In this equation sealing is not an attainable goal but a receding horizon, a target that retreats as one approaches it.

The saturation mechanism clarifies why improvement in interception rates, impressive as it may be, does not solve the underlying problem. Improvement occurs along the shield's performance axis — accuracy, range, response speed. The attacker operates along the axis of quantity and cost, where his advantage is close to unlimited. Every improvement in interception is offset by an increase in mass, and every increase in mass is far cheaper than any improvement in interception. This is not a symmetrical arms race in which both sides advance at a similar pace; it is a race in which the defender climbs a steep and expensive slope while the attacker slides cheaply downhill. Saturation is not a fault to be corrected in the next upgrade. It is a structural property of the relationship between cheap mass and an expensive shield.

The Loss of the Rear: The End of the Sanctuary

The deepest implication of the death of the shield is the loss of the rear as protected space. The founding premise of modern national security was the existence of a rear — a deep space, distant from the front, where civilian, industrial, and command centers of gravity sat in relative security. Distance was, in the simplest sense, a form of defense. To strike the rear, the attacker had to cross the front, overcome the defenses, and move sufficient force across a great distance — a task that was expensive, slow, and visible. The age of cheap multidomain penetration dismantles this premise at its root. As it has been tersely put, "distance is no longer the protection it once was" (Stoil, 2025). When a cheap loitering munition can cross hundreds of kilometers, and when drones can be smuggled, assembled, and launched from within the adversary's own territory — as demonstrated by operations in which drones were flown from trucks positioned deep inside enemy territory — depth loses its defensive force. It ceases to be a barrier and becomes merely a distance to be crossed cheaply.

The image proposed for the new vulnerability is that of spider webs: a thin, near-invisible mesh of threats woven into the rear from every direction (Bondar, 2025). The image captures the shift from a concentrated, identified threat that can be stopped at a line to a distributed and elusive threat that percolates through the porosity of territorial space. Where sealing assumed a single front to be defended, the spider-web reality presents a space vulnerable at every point, with no clear front and no secure rear. The most far-reaching consequence concerns the vulnerability of the military force in its own rear. War games simulating a conflict over Taiwan have found that up to ninety percent of expected aircraft losses would occur on the ground — at bases, before takeoff — rather than in air combat (Bondar, 2025). This inverts the premise of air superiority. The most advanced aircraft, the product of hundreds of millions of dollars of investment, is worthless if it can be destroyed while parked, by a cheap drone launched from a few kilometers away. The vast investment in platform quality is annulled by the cheapness of the means of destroying it in a vulnerable state.

The gap in preparedness between the major powers is stark, and it reveals the divide between those who have internalized the change and those who have not. China has built more than four hundred hardened aircraft shelters, while the United States across the Indo-Pacific has had only some twenty-two (Bondar, 2025). This gap of nearly twenty to one is not merely budgetary but conceptual: it shows which side is already planning to fight in a world where the rear is no sanctuary. The gap has ignited a central doctrinal debate: is the answer to vulnerability hardening — building shelters and point defense so that the platform survives the strike — or dispersal and agile basing — spreading the force across many rotating bases so as to deny the attacker any tempting concentration of targets? The debate is itself a symptom of the death of sealing. Neither side of it proposes to defend the base against penetration; their shared assumption is that the strike will arrive. Once the belief that penetration can be prevented breaks down, only two absorption strategies remain: harden the target so it survives, or disperse it so the strike destroys only a small part of the whole. Both are strategies of survival, not sealing, and the argument between them is one within the survival paradigm rather than against it.

The same trend sharpens the changing nature of offensive action. Poznansky and Sand (2025) argue that the new deep-penetration operations — the "showcase raids" of drones into enemy territory, aimed at psychological and cognitive effect no less than physical destruction — resemble cyber operations more than they resemble classical air strikes. They are elusive, deniable, operate from the shadows, and blur the line between war and peace. The conceptual implication is deep. If penetration resembles cyber, then defense against it should resemble cyber defense — which no one any longer imagines as an impervious wall, but as an array of resilience, detection, damage isolation, and recovery. This led Poznansky and Sand to recommend shifting the center of gravity of defense to the "left of launch" — to preventing the threat at the stage of its formation, before it takes off, through intelligence, preemptive action, and disruption of assembly and supply chains. The logic of left-of-launch is an implicit admission of the failure of right-of-launch. Once post-launch interception becomes too expensive and too unreliable, defense is pushed back in time and space, into the adversary's own territory, to the stage at which the threat is still cheap to neutralize because it has not yet materialized. This too is an abandonment of territorial sealing, but in the offensive direction. The best defense becomes preemptive attack, and the border loses its standing as the place where defense occurs.

The C-UAS Fabric: Promises and Their Limits

Against the threat of cheap precise mass, a rich and rapidly evolving fabric of counter-unmanned-aerial-system technologies has developed. Its limits — not merely its capabilities — are what establish the claim of the death of sealing. The claim is not that there is no technological answer to the drone; there are many answers, some quite effective. The claim is that none of them, and no combination of them, converges on the sealing once promised. Each layer solves one problem and opens a new vulnerability, and the sum is not a wall but a discriminating, layered fabric of absorption. The very need for multiple layers is itself testimony to the absence of a single impervious solution.

The cheapest and most widespread layer is electronic warfare — jamming the drone's control and navigation links, severing it from its operator and from satellite positioning. Its economic efficiency is impressive; jamming has been estimated to bring down some ten thousand drones per month in the Ukrainian theater (Bondar, 2025). It is the only cheap answer that can meet mass at a comparable cost, because it consumes no physical munition per interception. Yet its very success accelerated the development of countermeasures designed to undo it. The most prominent is the fiber-optic drone, tethered to its operator by a thin physical cable that pays out in flight and renders the drone wholly immune to jamming — there is no wireless link to jam — at ranges of fifteen to thirty kilometers (Bondar, 2025). In parallel, edge autonomy and "pixel lock" have developed: the drone's ability to lock onto its target through local image processing on a cheap onboard processor and complete the strike even with no link to the operator at all. These two developments hollow out electronic warfare. The cheapest and most central layer of defense is neutralized by a far cheaper upgrade of the threat — a distilled example of the inverse cost curve, in which the attacker's countermeasure costs less than the defender's measure it defeats.

Directed energy offers the most radical promise of breaking the shield dilemma. The high-energy laser destroys its target at a marginal cost of a few dollars per shot — roughly two to five dollars for Israel's Iron Beam, and about thirteen dollars for Britain's DragonFire — with the HELIOS system installed on U.S. Navy ships (Rumbaugh, 2024). If the cost per shot collapses to a few dollars, the wrong side of the cost curve appears, at last, to be broken, letting the shield fire more cheaply than the attacker. Small wonder that directed energy is sometimes presented as the silver bullet that will restore sealing. But a central caution belongs in any discussion of these figures. The "cost per shot" claims refer to marginal energy alone — the electricity consumed for a single shot — and ignore the cost of the system, its development, installation, and maintenance, and its operational limits. An operational laser system costs tens to hundreds of millions of dollars, and covering a broad space requires many of them. More important, the laser is unmistakably a point-defense weapon. Its range is limited; its performance degrades severely in weather, as fog, dust, rain, and cloud scatter and weaken the beam; and it requires a dwell time of seconds on each target to deliver energy sufficient to destroy it. That requirement critically limits the rate at which it can engage multiple targets: against a dense swarm penetrating simultaneously, a single laser can engage only one target at a time, and the swarm saturates it just as it saturates a kinetic interceptor. High-power microwave systems — such as Epirus's Leonidas and the THOR system — offer a complementary advantage, striking an entire swarm at once through a broad electromagnetic pulse that disrupts the electronics of many drones simultaneously and so solves the dwell-time problem. But microwave systems carry their own severe limits of short range and of the risk that their emissions will harm friendly electronics nearby, including the defense systems themselves and adjacent civilian infrastructure. Directed energy, then, is not the silver bullet that restores sealing but another point-defense layer within the overall absorption fabric — economically efficient at the point of destruction, yet limited in space, weather, and rate. It significantly improves the shield's position without abolishing the fundamental split between possible point defense and impossible spatial sealing.

Between cheap electronic warfare and the expensive laser, a space of cheap kinetic solutions has developed, meant to shrink the shield dilemma without breaking it outright: guided APKWS munitions at twenty to thirty-five thousand dollars, the Coyote interceptor at roughly one hundred twenty-five thousand dollars, and interceptor drones at just one to five thousand dollars (Rumbaugh, 2024). The cheap interceptor drones in particular seek to level the cost curve by matching the price of the interceptor to the price of the threat. The approach is promising precisely because it recognizes that the answer to cheap mass cannot be expensive quality but must be counter cheap mass.

Yet all of these layers — cheap and expensive, kinetic and directed — are subject to one shared limit, the hard floor beneath any C-UAS array: detection. One cannot intercept what one has not detected, and the quality of interception can never exceed the quality of the detection that feeds it. Detecting small, slow, low-flying drones is a fundamentally hard problem. They present a minute radar cross-section and a low thermal signature, and they move against ground clutter that hampers their discrimination. A single sensor delivers a defeat rate of only about eighty-five percent, leaving one leaker in six (Bondar, 2025). The figure is critical because it sets an upper ceiling on the entire array: if detection misses one in six, no interceptor, however effective, can stop what is not seen. To overcome the single-sensor limit, dense multilayered sensor networks have developed, combining radar, optics, thermal sensing, and acoustics — such as the "Sky Fortress" acoustic network of more than ten thousand sensors, which covers a broad space and minimizes dead zones (Bondar, 2025). The very need for a network of ten thousand sensors to provide reasonable coverage shows how far the sealing conception has drifted from reality. The new defense is not one wall but a dense, expensive, distributed fabric of sensing that, for all its scale, still leaves leakers. Detection is the hard floor above which every other layer hovers, and so long as detection is imperfect — and it is never perfect — sealing is unattainable by construction.

Drone Walls and the Defense of Territory

The natural political response to the drone threat is the call to build a drone wall — a continuous territorial array that seals the state's or the union's border against penetration. The call expresses a longing for classical sealing translated into the terms of the new technology: just as the wall once stopped armies, the drone wall will stop the swarms. But economic reality frustrates the longing at its root. As it has been put plainly, "no affordable drone-wall technology exists" (Dailey/RAND, 2025). This does not mean drones cannot be brought down — they can, a