Low-cost drones and attrition: the industrial lesson the civilian world has not yet grasped

The rise of low-cost drones has reshuffled part of contemporary military doctrine. Its real lesson, for anyone interested in industry, is less tactical than economic.

An openly accepted attrition logic

Unlike costly equipment designed to last, the low-cost drone is designed to be consumed: lost in large numbers, replaced quickly, improved through closely spaced iterations rather than long development cycles.

What this means industrially

• Production lines capable of high rates on standardised components
• An organisational tolerance for loss, which changes the usual quality logic
• Product iteration cycles measured in weeks, not years

Transfer to the civilian world remains partial

The civilian world still largely reasons in terms of durability and maximum quality per unit. The attrition logic (produce fast, accept loss, iterate on field feedback rather than on initial perfection) remains culturally foreign to most organisations, including in tech.

The example that makes the logic tangible

Ukraine illustrates this shift on an industrial scale. FPV drones appeared there in large numbers in 2023, to make up for the shortage of artillery shells. In January 2026, Ukraine's National Security and Defence Council stated that its industry's capacity makes it possible to produce more than 8 million FPV drones a year, across more than 160 manufacturers. This volume comes with a high individual failure rate, accepted rather than fought: according to RUSI (February 2025), between 60 and 80% of Ukrainian FPVs fail to reach their target, depending on the sector of the front and the pilots' skill. I detail this industrial trajectory in my drone war panorama 2022-2026. This production rate forced a military overhaul; see how Ukraine reorganised its army around drones.

This choice would be unthinkable in most civilian production lines, where the target defect rate is counted in fractions of a percent, not in tens. The real question a civilian industrial leader should ask is not "can I accept this level of loss", but "would my market forgive me for iterating this fast if I agreed to pay the price in unit quality". The answer is rarely yes across the board, but it deserves to be asked deliberately rather than dismissed by reflex.

Ukrainian production figures, and their limits

Ukrainian figures circulate widely, and they do not all measure the same thing. Three quantities need to be distinguished.

• Production capacity: what the lines could turn out if they were funded and supplied. This is the figure of 8 million FPVs a year announced by the National Security and Defence Council.
• Production ordered and delivered: in June 2025, Ukraine's Defence Minister, Rustem Umerov, announced a target of more than 4 million drones delivered to the front over the year, all types combined.
• The effect achieved: the number of drones that actually reach their target, much lower, as the RUSI estimate shows.

These three figures are statements or estimates, not audited data. They remain consistent on the essentials: Ukraine thinks in millions of consumable units a year, and its constraint is no longer design but supply. RUSI also estimates, in a November 2025 report, that China supplies around 80% of the global small drone market, complete systems and components included.

Unit cost versus target cost: the attrition equation

Attrition is only sustainable if the weapon lost costs much less than what it destroys. The same RUSI report gives a few orders of magnitude, with an important caveat.

System	Cost given by RUSI	What it changes
FPV drone	Under 500 dollars for some models	Losing most of the units remains bearable
Heavier reconnaissance or bomber drone	Several thousand dollars	Already less disposable, better protected
SPIKE NLOS anti-tank missile	250,000 dollars	Every shot has to count

The caveat comes from RUSI itself: FPV production costs often include neither labour, nor the ground station, nor the pilot's goggles, nor the explosive charge. And the equation deteriorates when the target protects itself: the report cites a case, in July 2025, where some sixty drones were reportedly needed to destroy a single reinforced tank. Even in that case, the whole remains cheaper than a guided missile. The February 2025 report noted for its part that tactical drones then accounted for 60 to 70% of Russian systems damaged or destroyed.

What Europe and France take from it

For a European defence industry, the lesson is not primarily about the drone; it is about the chain. Mass-producing a consumable object requires components available in volume, replaceable suppliers and short qualification lead times. This is precisely where dependence on Chinese components becomes a sovereignty issue, and where qualification cycles inherited from equipment designed to last become a brake. I analyse the French situation in the blind spots of the French defence industrial base on military drones.

Transferring attrition to the civilian world: where it works, where it does not

The attrition logic transfers well wherever the failure of one unit costs little and teaches a lot: software prototypes, marketing message tests, page variants, small batches of parts intended for trials. In these contexts, multiplying cheap attempts and quickly discarding what does not work is often more profitable than making every version reliable before showing it.

It does not transfer where a faulty unit puts a person or data at risk: medical devices, aeronautics, systems security, personal data processing. The right trade-off is therefore not "attrition or quality", but knowing, component by component, which of the two regimes applies.

The lesson of low-cost drones is not only military. It is a reminder that iteration speed can sometimes matter more than unit robustness, a trade-off that few civilian organisations yet dare to make consciously.

Further reading: industrial sovereignty: what the war economy really demands of defence SMEs and AI and nuclear deterrence: the problem is not the decision, it is the patch delay

Frequently asked questions

How many drones does Ukraine produce per year?

In January 2026, Ukraine's National Security and Defence Council announced a capacity of more than 8 million FPV drones a year. This is a capacity, not audited production: in 2025, the stated target was more than 4 million drones delivered to the front.

Why do so many FPV drones fail to reach their target?

Electronic jamming, weather and pilot skill are the main causes. In February 2025, RUSI estimated that between 60 and 80% of Ukrainian FPVs failed to reach their target, depending on the sector of the front.

What does attrition mean for a civilian company?

Accepting the loss of cheap units in order to learn faster, where a failure puts nobody at risk. This applies to prototypes or tests, not to products whose failure has serious consequences.