For seventy years, the rising cost of air defence effectors was tolerable because the targets were themselves expensive, crewed, and scarce. The cheap attritable drone inverted that equation, and the industry is now scrambling back down a cost curve it spent decades climbing. The technically hardest problem — hitting the target — was substantially solved a generation ago. The economically hardest problem — hitting it affordably, thousands of times — is only now being treated as the primary design requirement rather than a footnote.
The economics of air defence began at the bottom of the curve. A heavy anti-aircraft shell in the Second World War cost, in relative terms, pocket change — the problem was that almost none of them hit anything. The commonly cited figure for German heavy flak is 16,000 rounds of 88mm ammunition expended per Allied bomber destroyed, though serious estimates vary widely: analyses of Luftwaffe records give figures closer to 4,000 rounds per kill in 1941–42, falling toward 1,600 by 1944 as radar direction improved, and contemporary Luftwaffe calculations credited the improved Flak 41 with roughly half the consumption of the standard Flak 36. Whichever figure one prefers, the structure of the problem is the same: the ammunition was cheap and the defence was ruinously expensive at the same time, because the cost that matters is not the cost of the effector but the cost of the kill. The proximity fuze demonstrated what closing that gap looks like. Against the V-1 campaign in the summer of 1944, the proportion of incoming missiles destroyed by the coastal gun belt rose from 17 per cent in June to roughly 74 per cent by late August — a transformation achieved by the VT fuze working in combination with the SCR-584 gun-laying radar and electronic predictors. The first great lesson of air defence economics was thus established early, and so was a second one this series returns to later: the revolution was as much in the fire-control chain as in the shell.
The missile era pushed the effector cost upward by several orders of magnitude, and for decades this was rational. A Nike or Hawk round cost vastly more than a shell, but it engaged targets — crewed jet bombers, later strike aircraft — that cost more still, carried irreplaceable crews, and existed in inventories of hundreds rather than tens of thousands. The cost-exchange ratio, measured properly as interceptor cost against target value plus prevented damage, remained comfortably favourable even as unit prices climbed. Hit-to-kill technology, matured through PAC-3 in the 1990s and 2000s, represented the summit of this logic: extraordinary terminal precision at extraordinary unit cost, justified because the target was a ballistic missile whose warhead — potentially unconventional — made almost any interceptor price defensible. A PAC-3 MSE round has run at roughly four million US dollars flyaway in recent US budget cycles, with the latest FY2027 request showing unit cost rising toward five million — and FMS case values per round running considerably higher still. That distinction, between flyaway and FMS case cost, is one this site has repeatedly insisted upon, most recently in our PAC-3 ACE cost model, because conflating the two figures is the single most common error in public interceptor economics.
The inversion arrived quietly and then all at once. Loitering munitions and attritable one-way attack drones had appeared in various forms through the 2010s, but the Shahed-136 campaign against Ukraine from late 2022 demonstrated the new arithmetic at scale. The airframe’s true cost remains genuinely uncertain and illustrates why concept-stage and adversary pricing alike deserve scepticism: leaked contract data indicate Iran sold early kits to Russia at $193,000–370,000 apiece in 2022, while localised Russian production at Alabuga is now credibly estimated around $70,000 per unit, and the widely circulated $20,000–50,000 figures stem largely from early-war component inspections. Even at the highest of these figures, the drone absorbed interceptors costing ten to a hundred times as much. The Red Sea engagements from 2023 onward made the same point in the naval domain, with multi-million-dollar standard missiles expended against targets worth a small fraction of their price. The defence could win every engagement and lose the exchange. Magazine depth, not intercept probability, became the operative variable — a theme this site has tracked across the interceptor cost database since its inception.
The industry’s response has taken four forms, all of them, in essence, a retreat back down its own cost curve. The first is rediscovery of the gun: the Gepard’s second career in Ukraine is the emblematic case, a system retired by its originating army as obsolete proving to be among the most cost-effective counter-drone platforms in the war, at an estimated cost per engagement burst on the order of $5,000 — a rounding error against any missile. Programmable airburst ammunition, modernised AAA, and systems in the NOMADS class extend the same logic. The second is the cheap guided effector: laser-guided rockets in the APKWS class, and now the purpose-built low-cost interceptor cluster launched at Farnborough this July — PAC-3 ACE, Counter Mass Interceptor, Vorexon, Red Kite — which we surveyed as an industry-wide admission that the cost curve, not intercept performance, is the binding constraint. As we argued in the ACE analysis, these systems improve magazine depth and cost-per-kill against the drone and cruise missile tier; they do not, and do not claim to, address the upper tier, which is why they complement rather than substitute for the exquisite interceptors above them. The third form is the newest and, in Ukraine, already the most productive: the interceptor drone. Ukrainian interceptor drones costing on the order of $2,500 apiece were credited with roughly 70 per cent of the record 1,704 Shaheds downed in January 2026 — a cost-exchange ratio no gun or missile approaches, achieved by meeting the attritable threat with an attritable defence. The fourth is directed energy, which promises a near-zero marginal cost per shot and which we examined at length in our survey of Western laser programmes — with the standing caveat that power, beam quality, weather sensitivity, and above all the gap between demonstrator and deployed system have humbled directed-energy promises for fifty years.
What has actually changed, then, is not the technology so much as the design philosophy. For decades, cost was an output of the design process: requirements specified performance, and the price was whatever the performance cost. The Farnborough cluster represents — at least rhetorically — the inversion of that process: cost as an input, a hard ceiling around which performance is engineered. Whether the announced price points survive contact with production is precisely the question this site’s standing scepticism toward concept-stage pricing exists to hold open. The historical record of defence programmes converging downward toward announced target costs is not encouraging.
The deeper point is structural. An air defence architecture is now, unavoidably, a portfolio problem: high-cost interceptors for high-value threats, cheap effectors for cheap threats, and a command layer capable of assigning the right effector to the right track — the subject of the fourth article in this series. A nation that buys only the exquisite tier purchases the ability to lose expensively; a nation that buys only the cheap tier leaves its upper airspace open. The cost curve does not merely constrain the effector — it dictates the architecture. For Norway, whose geographic duality places Finnmark under ballistic threat from the Kola direction while the south and west face the cruise missile and drone tier, the portfolio logic is not optional. No single system class, at any price, addresses both requirements — a point this site has made before and will make again.
AI-assisted article.
