Take the trade on its own terms. Between February 2022 and March 2026 the United States committed $67.8 billion in defence articles and services to Ukraine, and the principal land adversary of the alliance has been fought to a standstill at a price in allied lives of zero. No conventional accounting of national interest makes that a bad purchase. The weakness of this policy has never been its strategy; it has been its financing.
The Congressional Research Service records the position without adjectives. The special inspector general reports that Congress "has not made significant new funding available for … the Ukraine response since April 2024", and the presidential drawdown ceiling has fallen back to $100 million a year. What is left is a machine for spending other people's money: under the Prioritised Ukraine Requirements List, allies had committed more than $4.8 billion by March 2026, of which about $4.2 billion had actually reached the US Treasury.
An alliance that funds deterrence by annual improvisation has not bought deterrence. It has rented it, and the landlord sets the term.
At Ankara in July the allies pledged €70 billion in military equipment, assistance and training for 2026 and undertook to sustain at least the equivalent in 2027. The declaration asks that the support be "equitable, predictable, and sustainable in the long-term", which is exactly right and exactly what it has never been. On the same weekend the arithmetic was less flattering: only five of thirty-two members are projected to reach the 3.5 per cent core-defence figure this year, against a European average of 2.53 per cent.
Meanwhile the thing being bought has changed. Hal Brands notes that development cycles for small drones have fallen "from years to weeks or days", that mid-range systems now reach logistics targets more than a hundred miles back, and that Russian refinery output is down about 15 per cent year on year even with global prices higher. Russia is losing something like 35,000 men a month for this.
That points at the reform worth making. Ukrainian officials said in January that the country produces more than half its own weapon systems and put its defence-industrial capacity at $55 billion for 2026. A capability that costs a fraction of the American equivalent and is being iterated weekly under fire is not charity to be topped up; it is a supplier to be contracted, licensed and paid on time.
None of which licenses complacency. Brands is right that Russia remains a resilient, learning organisation, that Ukraine's manpower problem is unsolved, and that the grid faces another winter with its air defences thinner than last. Drone superiority is a lead, not a moat, and leads close: the side that is behind on a technology it can manufacture cheaply does not stay behind for long. An advantage held for eighteen months is a reason to convert it into contracts now, not a reason to assume it will still be there when the next appropriation cycle gets round to noticing.
So the conservative case is not for more emotion or less. It is for a contract: multi-year appropriations rather than supplementals, allied contributions written into national budgets rather than announced at summits, and procurement placed where the marginal dollar buys the most capability — which is now, increasingly, in Ukraine itself. Fund it properly or stop pretending. The one thing that cannot be afforded is the current arrangement, which costs real money and buys a policy nobody would design.
