In a single month, according to Ukraine’s Ministry of Defence, Russia employed more than 9,000 strike drones and missiles against Ukrainian territory, and Ukrainian air defence reported intercepting or otherwise neutralising 5,142 attack UAVs and 216 missiles. In the same month the UN Human Rights Monitoring Mission recorded more civilians killed or injured than in any month since March 2022 — 437 killed and 2,610 injured — and attributed 38 per cent of those casualties to long-range missiles and drones. Those two records, one from a belligerent and one from an independent monitor, describe the same campaign from opposite ends: enormous volume, high reported attrition, and severe residual effect. Europe’s rearmament debate has moved beyond the question of whether inexpensive drones can substitute for missiles. The more consequential problem is how different classes of weapon should work together over the duration of a campaign. Long-range one-way attack drones can provide volume, persistence and an economical means of complicating air defence; cruise and ballistic missiles provide different combinations of speed, survivability, payload and time-to-target; inexpensive defensive effectors can preserve scarce missile interceptors; and high-performance air and missile defence remains indispensable against threats that cheaper layers cannot reliably defeat. The requirement that follows is not a larger inventory of any single weapon class. It is an architecture in which sensors, targeting, launchers, mission data, effectors, stockpiles, industrial replacement and command arrangements are designed around the same assumptions about what must be struck or defended, for how long, and at what sustainable rate. Whether Europe is building that architecture, or a set of programmes that will later have to be reconciled, is the question the public record can now be asked to answer.
The evidence used here is the disclosed record and nothing else. On the operational side, the monthly and single-night accounts published by Ukraine’s Ministry of Defence, the president’s own statements, and the independent verification of the UN Human Rights Monitoring Mission. On the programme side, NATO’s multinational capability register and the remarks made at the 2026 Defence Industry Forum; the joint statement of the twelve states behind the Deep Precision Strike Capability Investment Initiative; the British Defence Investment Plan and the Ministry of Defence’s programme announcements and procurement notices for Brakestop, Nightfall and PrSM; the German federal government’s articulation of its three-pillar approach and the defence ministry’s own record of the Taurus Neo decision; and the Franco-Ukrainian joint declaration of 14 July 2026. On the industrial side, OCCAR’s contract amendments and firing records for SAMP/T NG and Aster, MBDA’s 2025 annual results, Diehl Defence’s capacity announcements, RTX’s GEM-T contract and Lockheed Martin’s PAC-3 MSE production disclosures. On the financial side, Regulations 2025/1106 and 2025/2643 in their operative text, the Council’s SAFE implementing decisions, and the Commission’s disbursements under the €90 billion Ukraine Support Loan. This report does not estimate stockpiles, does not infer classified planning, and does not offer a cost-per-target figure, because the public record does not support one. It asks a narrower question: at each stage between political demand and a replenished magazine, what does the disclosed record actually prove.