Two numbers define how much longer Ukraine can keep fighting: how many interceptor missiles it has left to defend its skies, and how many of its soldiers are actually dying to hold the line. Both are contested, both are politically massaged, and both look worse than the headline versions suggest once you do the math.

The interceptor math

President Volodymyr Zelenskiy asked US President Donald Trump for an emergency package of 300 Patriot interceptors before winter, he told Axios in an interview in late July. But with the US stockpile down to a mere 759 PAC-3 interceptors due to the Iran war, Trump said no. “We need them ourselves.” Within a week he had also back tracked on the licenses pledge. Ukraine is not going to get any more air defence ammo before the anticipated Russian winter barrage is expected to start once the mercury starts falling.

But even 300 interceptor missiles won’t change the calculus. Ukraine burns through an estimated 60-70 PAC-3 interceptors in a routine month and 150-180 in an intensive bombardment, according to Defense Express, which put Ukraine's bare minimum requirement at 60 interceptors a month even before the start of the current brutal tit-for-tat missile war that is now rapidly escalating in recent weeks. this summer's escalation - and at 122-244 a month if Ukrainian crews used interceptors at the two-per-target rate the US Army itself considers standard; due to the shortage the AAFU) has reportedly been ratioing themselves to one interceptor per inbound ballistic missile and are using them in manual mode in order to prevent them chasing lower-value missiles like drones that can be brought down by cheaper means, according to Defense Express.

Andrius Kubilius, the EU's commissioner for defence and space, put a similar range on it in blunter terms in March: Ukraine used about 700 Patriot interceptors over the winter, or roughly 175 a month, Ukrainska Pravda reported.

But Zelenskiy's request for 300 PAK-3s remains insufficient against all the variables in play: at a heavy-month rate of 150-180, 300 interceptors covers 1.7 to 2 months while the routine rate of 60-70, stretches that out to four-and-a-bit months.

Russia has increased its fire rate to hit targets but it also intentionally trying to exhaust Ukraine’s remaining stockpile ahead of winter. July was Russia's heaviest bombardment month of the war so far – double that of June, which was already a record month. Moscow fired around four dozen ballistic missiles at Kyiv in a single night on July 18, part of a week that saw 1,450 attack drones, 1,640 guided bombs and 99 missiles launched, and Ukraine's own air defences intercepted just one of 27 incoming ballistic missiles during an August 1 barrage.

Supply side constrains

Lockheed Martin built 620 PAC-3 MSE interceptors in the whole of 2025 - a record year for the company, and still barely 52 a month – less than Ukraine’s needs in a “normal” month, shared across the US Army's own stockpile, training reserves and every other Patriot operator in the world, not just Ukraine. Zelenskiy's top-up it is roughly half of an entire year's global production run.

The Pentagon and Lockheed Martin have since signed a framework to expand output to 2,000 missiles a year, but that ramp-up is not expected to complete until around 2033 and Ukraine needs more ammo now.

Zelenskiy has separately said Ukraine needs at least seven Patriot systems to shield its key cities and has called for 25 in more ambitious moments to protect all its large cities; but no official breakdown exists of how many of Ukraine's 24 oblasts, or how many individual cities within them, a given battery count would actually cover.

The 60-70/150-180 baseline assumes the intensity of Russia's ballistic missile campaign holds roughly steady. It has not. Russia fired about 126 ballistic missiles at Ukraine in July - Iskander-M, North Korean KN-23s, and S-300/S-400 systems used in a ground-attack role - the highest monthly total of 2026, according to the Oko Gora monitoring project, reported by Militarnyi on August 2. That is up from 96 in June and 86 in May, following a bumpier run earlier in the year (74 in January, 116 in February, 39 in March, 50 in April) - July is both the sharpest month-on-month rise since February and the highest total recorded so far in 2026.

Counting the dead

Separately there is a lot of confusion over the number of casualties the AFU have been taking. The number is important as Ukraine is suffering from an acute manpower shortage, but the number of Ukraine’s killed and wounded servicemen remains a closely guarded state secrete.

Zelenskiy's own casualty figures do not form a coherent series, and the gap became impossible to ignore at the end of July. He told CBS 46,000 Ukrainian soldiers killed in February 2025, then France 2 TV station 55,000 in February 2026, then, in a Fox News interview recorded during his Washington visit for Senator Lindsey Graham's funeral, "about 50,000 soldiers killed" on July 31 - a five-month fall of 5,000. RT gleefully wrote up the discrepancy under the headline "Zelensky's troop death toll moves backward", and the reversal drew open mockery on Ukrainian social media, where commentators noted that dead soldiers do not un-die.

KIA and WIA numbers are a political hot potato and Zelenskiy's own numbers don’t quiet add up. Compare February 2025 to February 2026 and the implied rate is 750 killed a month. Compare February 2026 to July 2026 and the rate is negative - which is not possible. The only way to get a plausible, positive monthly figure out of his statements is to skip the February 2026 spike entirely and compare February 2025 (46,000) directly to July 2026 (50,000): a rise of 4,000 over roughly 18 months, or 222 a month – a number several analhysts have cited.

That estimate tallies if calculated backward from the reported number of Russians killed, reported by the Ukrainian Defence Ministry and the attacker to defender kill ratios which run from 3:1 up to 8:1 in the first half of 2026, according to a CSIS study. That gives and estimate range of between 100 and 300 Ukrainians KIA per day – roughly the same as the 222 estimate above. But even these estimates are likely to be low.

Independent estimates put the real toll far higher. A CSIS brief published on July 1, "Russian Blood and Treasure" (Seth Jones and Riley McCabe), calculated that Ukraine suffered between 525,000 and 625,000 total casualties and between 125,000 and 150,000 deaths between February 2022 and June 2026 - two-and-a-half to three times Zelenskiy's own public figures. That is consistent with IntelliNews own reporting a year earlier, when a big investigation by French daily Le Monde into Ukraine's overflowing military cemeteries concluded Kyiv's official toll was almost certainly understated. The new national memorial complexes under construction in Kyiv and Lviv for fallen soldiers were running out of space due to the swelling numbers of graves.

On the Russian side of the ledger, the same CSIS study found monthly casualty rates of 30,000-34,000 in 2026 - which, once split by CSIS's own estimate of the killed share of that total, works out closer to 8,700-9,800 dead a month, not the far lower confirmed-only figures that circulate from named-casualty trackers such as Mediazona and the BBC's Russian service, whose jointly verified list of identified Russian dead passed 236,000 in early August - a real but necessarily incomplete count, since it depends on obituaries, family posts and local reporting catching up with each death.

CSIS attributes the rising death ratio creeping up to as high as 8:1 as largely due to Ukraine's deepening use of AI-assisted drones in a war that continues to favour the defender over the attacker. Applied to Russia's actual current death rate rather than to Zelenskiy's own contested figures, that 8:1 ratio implies Ukraine is losing somewhere in the region of 1,100-1,225 soldiers a month - a modelled estimate, not a reported one, but one that fits a war in which both sides have every incentive to make their own dead look like someone else's problem.