Rate, not prototype: reading Delair as Europe’s drone industrialisation test

I. The mechanism: rate, not range

For fifteen years the competitive question in small unmanned aircraft was performance per airframe — endurance, payload, range, autonomy. Ukraine ended that framing. When an attritable aircraft has a service life measured in sorties rather than years, the binding constraint moves from the design office to the production line and to the supply chain that feeds it. The decisive quantity becomes units per day at an accepted quality level, sustained while the electronic-warfare environment forces a change of guidance, radio or seeker every few months.

This reverses the logic of European aerospace industrialisation. A conventional programme optimises for a certified, frozen configuration produced slowly at high unit value. An attritable-systems line optimises for a re-qualifiable configuration produced fast at low unit value — and its hardest engineering is not aerodynamic. It is industrial: fixturing, test automation, component substitution, incoming inspection, configuration management across variants that ship in the same week.

Read against that mechanism, Delair's publicly visible moves stop looking like opportunism and start looking like a single coherent bet: keep the flight-proven civil platforms as the qualification and cash base, take over floor space that was built for automotive-grade serial manufacturing rather than aerospace-grade craft production, and state a daily rate as the objective rather than a unit price.

II. The identity: what the register actually says

Delair's own communication describes a company created in 2011 by engineers out of the Toulouse aerospace basin, and that origin date is consistent across its materials. The legal person that trades today is younger: the SIREN 877 643 486 corresponds to an entity registered in 2019, of the SAS form, and the company's published legal notice gives a share capital of €188,997.85 and an address at Labège, on the eastern edge of Toulouse.

We flag the gap rather than smoothing it, because it is exactly the kind of detail that later gets mis-cited: a 2011 founding narrative and a 2019 legal entity are both true statements about different objects — the business and the company — and a reader who conflates them will draw wrong conclusions about continuity of contracts, accounts and shareholding. We describe the founding as 2011 and the current registered entity as created in 2019, and we do not reconstruct the corporate path between them, because the reorganisation deeds are not in the openable record we consulted.

On ownership we have to be blunter still. Three claims circulate widely enough that we tested them specifically: that Delair belongs to a "Groupe Chiron" led by Michael Kluehr; that Tikehau Capital holds it; and that Ace Capital Partners holds it. In this pass, none of the three was supported by any document we could open. We therefore state no shareholder for Delair at all. An unsourced cap table is worse than a missing one, because it is repeatable.

III. The Grenoble acquisition, and the one we refuse to print

The company's inorganic history includes a genuine, documented Alpine link: the acquisition of Squadrone System, a Grenoble rotary-wing and indoor-inventory drone company, which was reported at the time as employing 22 people. That transaction matters to the argument of this dossier for a structural reason rather than a sentimental one: it moved rotary-wing and confined-space autonomy competence out of the Grenoble research basin and into a Toulouse industrial vehicle, at a moment when the market still believed the growth would be in industrial inspection.

A second acquisition is frequently attributed to Delair — the Toulouse automated-inspection company Donecle. We could not open a single document supporting it. It is not printed here as fact, and no part of this analysis rests on it.

IV. The product ladder: from survey aircraft to loitering munitions

Four platform families are publicly identifiable, and the ladder between them is the argument.

The UX11 is the civil-survey fixed-wing: a small, hand-launched photogrammetry aircraft aimed at surveyors and corridor inspection. Its published endurance is one of the disagreements we carry rather than resolve — 59 minutes in one Delair-associated specification and 80 minutes in another. Both are company-side numbers. We print the pair, because a 35%¹ spread in stated endurance is not a rounding difference; it is either a configuration difference the publisher failed to state or a documentation error, and either way a procurement officer needs to see it before signing.

The DT26 is the long-endurance fixed-wing that carried the company into beyond-visual-line-of-sight operations, the regulatory territory where a drone stops being a camera on a stick and becomes an aircraft in someone's airspace. The DT46 is the current tactical reconnaissance platform, and it is the pivot: the same airframe logic, now specified against military users.

Then the step that changes the company's category. Colibri and Damoclès are publicly named loitering munitions. A reconnaissance aircraft that fails is a lost sensor; a munition that fails is a safety and legal event. Crossing that line obliges a manufacturer to add warhead-adjacent handling, arming-chain discipline, and a level of traceability that civil survey work never demanded — while going faster, not slower.

V. Ukraine, Boussens and the hundred-a-day claim

Delair has publicly delivered systems to Ukraine. That single fact does more analytical work than any brochure, because Ukraine is the only qualification environment in the world that tests the attributes described in section I simultaneously: attrition rate, electronic-warfare pressure, and iteration speed. A European manufacturer with fielded deliveries there holds evidence its competitors can only model.

The industrial answer is at Boussens, in the Haute-Garonne, where the company has taken over premises previously operated by Schaeffler and announced an expansion of production, publicly framed around an objective of roughly 100 drones or interceptors per day. Note what kind of building that is. Schaeffler is an automotive-grade serial manufacturer; the floor, the utilities and the logistics geometry of such a site are designed for continuous flow, which is precisely what an attritable-systems line needs and precisely what a legacy aerospace workshop is not organised to give. Reusing it is faster than building it, and it is also a labour-market statement in a basin whose automotive employment has been contracting.

We treat the hundred-a-day figure as a stated objective, not as a measured output. No document we opened reports an achieved daily rate, a certified line capacity, or a delivered volume. The distinction is the whole difference between a capability and a press release, and we will not collapse it.

VI. The state layer: France 2030 and VOLTAIR

Delair appears in the French public-support architecture through France 2030 and the VOLTAIR programme, the state's vehicle for accelerating uncrewed-systems capability. This is the structural point: in Europe, attritable-systems capacity is not being financed by defence procurement alone but by industrial-policy instruments designed for dual-use scale-up, which is why the same company can hold an ISO 9001 quality certification, a civil BVLOS operating record and a munitions product line at the same time.

What we do not print is any amount. No grant value, no contract value, no order book figure for Delair was available in an openable document in this pass. Readers who have seen numbers attached to these programmes elsewhere should treat them as unverified until the awarding body publishes them.

VII. The numbers that disagree, printed as they are

This company is a case study in why a single-source figure should never be quoted with confidence. For headcount we found three incompatible values in the openable record — 84 (2024), 217 and 399 (March 2026). For 2025 revenue we found one openable figure — about €50m, reported on 8 July 2026; a higher projection circulates behind a paywall we could not open, so it is not carried here. On the regulatory side, the release date of a C6 class-marked configuration is given inconsistently across the sources that mention it.

We do not average these, and we do not pick the flattering one. A four-way headcount spread of that magnitude for a company of this size usually means the figures describe different objects — the SAS alone, the group, full-time equivalents, or a target — and none of the publishers said which. Until Delair or its filings state a scope, the honest number is the range with its provenance attached.

VIII. What we refuse to print

• Any shareholder. The Chiron/Kluehr, Tikehau and Ace Capital claims are unsupported in the openable record.
• The Donecle acquisition. No supporting document.
• Blue UAS status. We found no entry evidencing Delair on the US Defense Innovation Unit's cleared list, so the company is not described here as Blue UAS approved.
• Any contract, grant or order-book amount, French or Ukrainian.
• Achieved production rate, per-site headcount, unit prices, warhead characteristics, or customer identities beyond what the company itself has announced.

IX. The analytical conclusion

Delair is best read not as a drone brand but as a test of whether European industrial policy can convert a research-grade competence into rate. Every element of its public position points the same way: a civil qualification base that gives it flight-hours no start-up has, a Grenoble autonomy acquisition that gave it competence its Toulouse peers lacked, a war-tested product line that gives it feedback its European competitors cannot buy, and a repurposed automotive plant that gives it floor geometry aerospace could not lend it. The bet is coherent.

The unresolved risk is equally visible, and it is not aerodynamic. It is governance and disclosure. A company whose headcount is reported four different ways, whose revenue is reported two ways, whose shareholding is asserted three ways without a document, and whose flagship industrial claim is a daily rate no third party has measured, is carrying an information deficit that scales badly. In defence procurement, an information deficit is a pricing event. The most valuable thing Delair could publish next is not a new airframe. It is a scope-defined headcount, an audited revenue line, a shareholding table, and one verified output number from Boussens.

Reporting standard: every figure above is attributed to a document we opened on the stated date. Contradictions are printed side by side and not adjudicated. Absences are named as absences. Nothing in this dossier is estimated, interpolated, or inferred from a competitor.