Helsing is worth $18 billion. Its largest confirmed government order is €269 million.
Quantum Systems is worth $8 billion. Its largest order ever is $246 million, signed with the German Army in December 2025.
General Atomics has raised no venture capital, gets almost no European coverage, and sits on €29.9 billion of tracked drone contracts. Close to nine-tenths of that is a single product line, the MQ-9 Reaper and its SeaGuardian cousins, sold mostly to governments outside Europe.
Hold those three facts together and you have the state of European military drones in 2026. The companies the press writes about and the companies that sign contracts are barely the same list, and the largest name on the contract list is an American firm selling a design that first flew in 2001.
A word on what follows. This is drawn from the record of who has actually been paid: defence-drone tenders and contract awards from European, US and NATO sources, 1,173 of them worth €105 billion, contract values converted to euros. It maps spending, not factories, and the first thing it shows is that Europe is still importing.
The order book and the cap table disagree
Rank the awards by value and the top of the table is uncomfortable for anyone who thinks Europe has rearmed itself:
€29.9B
€9.1B
€8.5B
€4.8B
€4.0B
€3.5B
€3.4B
€3.1B
Two American firms, General Atomics and Northrop Grumman, hold more contract value between them than every European supplier in the top ten combined. The Europeans who make the list got there through a handful of very large awards: Airbus Defence and Space books €9.1 billion across nine contracts, Poland's WB Group €4.0 billion across three, PGZ €3.4 billion from one.
None of the names that dominate the coverage come close. A venture valuation is a price for a decade of orders a company hopes to win; a contract is money that has already changed hands. Right now the two point in different directions, and the interesting question is which one is early.
A bubble, or a bet?
Read the table as proof the startups are overvalued and you are reading it the way the smart money refuses to.
Anduril is the sharpest case. It raised $5 billion at a $61 billion valuation in May and by late July was reportedly in talks near $100 billion, which would rival Northrop Grumman. It is private and publishes no audited accounts, but on the roughly $2 billion of revenue its investors have described, the May round already priced it at about thirty times sales, and the mooted round would be nearer fifty. Lockheed Martin, on $69 billion of real revenue, trades at under two times sales. Even Palantir, the market's favourite defence-software stock, sits in the low-to-mid teens. Anduril is priced above all of them, and its backers know exactly what they are doing.
They are buying a change in where the money goes. Ukraine has made cheap drones a strategic weapon rather than a gadget, defence budgets are climbing across NATO, and the bet is that the durable margin moves to whoever owns the autonomy software rather than the airframe. That last part is a wager, not a fact. The companies carrying the argument are hardware manufacturers today: Helsing builds HX-2 airframes, Quantum builds the Falke, Anduril builds airframes and rocket motors. Whether the software layer holds premium margins, or gets commoditised while the value drains to the cheapest factory, is the open question the valuations are answering yes to in advance.
The neat "plucky startups versus dinosaurs" story does not survive contact with the evidence, in either direction. The incumbents are moving. General Atomics and Anduril are the two finalists in the US Air Force's Collaborative Combat Aircraft programme, and both won production contracts in June 2026 for autonomous jets the Air Force plans to field by the thousand at tens of millions of dollars each. That is autonomy at the exquisite end, a different thing from the €400 quadcopters of the Ukrainian front, and the old-guard prime is winning it head to head with the venture darling. Where primes cannot build the culture, they buy it: Dassault bankrolls the French drone startup Harmattan, Leonardo runs the Baykar joint venture. The autonomy startups are as likely to be acquired as to displace.
And the bet rides on the war. Public defence shares have dropped on ceasefire headlines more than once, and the harder objection is operational: Ukraine's lesson so far is that a $500 airframe flown by a human, increasingly down an unjammable fibre-optic cable, does most of the killing, while software-piloted autonomy is a last-mile aid inside a jamming environment that breaks radio links in seconds. Whether that transfers to a peer war, and whether it rewards a $100 billion software house or a shed full of soldering irons, is unsettled. At a $100 billion mark there is also no prime large enough to buy Anduril, so the only way out is a public listing into the biggest defence company in American history. That, stripped of the noise, is the trade.
The loitering munition number that looks small
Across every tracked contract, loitering munitions come to €3.3 billion over seventeen deals, about three percent of the total and a fraction of fixed-wing drones at €58.4 billion. For the weapon everyone writes about, that is still strikingly little, and the true figure is higher again, for three reasons.
Values go unpublished. MBDA signed its first One Way Effector contract with France's DGA on 22 January 2026 and disclosed no value, quantity or schedule. A contract with no published figure adds nothing to a value ranking however large it is.
Frameworks book their floor, not their ceiling. In February 2026 Germany's budget committee approved its first one-way-attack-drone contracts: €268 million to Stark Defence for 2,200 Virtus munitions and €269 million to Helsing for HX-2, inside framework agreements ceilinged near €4.3 billion. The record shows the €268 million signed, not the €4.3 billion that might follow.
And most of it never reaches a tender portal. That €3.3 billion already counts the search-and-strike systems, like Poland's €2.1 billion GLADIUS, that are easy to log as ordinary aircraft. What it cannot count is Ukraine's own war economy, millions of FPV and one-way-attack drones a year built in hundreds of workshops and volunteer lines, almost entirely off the books. The tracked market is the formal tip of a much larger iceberg, and anyone sizing the real thing from tender data alone will be out by an order of magnitude.
What each country is actually buying
Germany
Helsing builds the HX-2, a GPS-independent strike drone, at a site in southern Germany it says can turn out more than 1,000 a month. Treat that as the manufacturer's nameplate figure, not verified output, and set it against Ukraine's hundreds of thousands of FPVs a month before calling it mass production. The Ukrainian order is disputed: Bloomberg reported a pause after reliability problems, WELT cited HX-2 accuracy near 35 percent, and Helsing rejects the pause and claims orders above 1,000 units. Note that 35 percent terminal accuracy for a radio-linked drone in a heavily jammed sector may be closer to par than to scandal; the metric is a battlefield condition as much as a verdict on one vendor.
Quantum Systems has the tidiest order-to-hype ratio of the German field. Its $1.2 billion round in July 2026 doubled its valuation to $8 billion, but the number that counts is the December 2025 German Army award: 520 Falke systems for 2026, with options for roughly 500 more. Add a $36 million Romanian deal through the EU's SAFE loan facility and a $15 million US Army order, and its signed book is a fraction of its price.
France
France's €11.8 billion across 17 contracts looks like the boldest buyer in Europe until you open it. €8 billion of it is a single four-nation programme, Eurodrone, booked in full to France as lead nation and not yet flown. Strip that out and France has spent about €3.8 billion, roughly €1.7 billion of it on American Reapers. Its own median drone contract is €250 million, still concentrated in a few big awards, but a long way short of the headline.
The most interesting French buy is the smallest. Harmattan AI, a two-year-old startup backed by a $200 million Dassault-led round, delivered 1,000 of its 1.8kg Sonora reconnaissance drones to the Army within six months, then won an order for 5,000 more. These are €1,000 soldier's-eye quadcopters, not Reaper replacements, and the point is the clock: six months from signature to delivery, from a firm a prime chose to fund rather than fight.
United Kingdom
Britain runs France in reverse. 158 tracked contracts worth €3.2 billion, a €21 million average and a median close to zero: many small buys, and a lot of awards published without a value at all. The pattern is dispersal, not concentration. Its biggest recent move is Tekever, selected in July to replace Watchkeeper with the AR5, a sole-source deal worth up to £400 million over a decade that starts with six aircraft and is due to be signed by 31 August. A ten-year headline, six aircraft up front: the British habit in one contract.
Poland
Poland is the buyer coverage keeps underrating. €8.1 billion tracked across 20 contracts, €5.8 billion of it in the last two years, and both of Europe's two largest single 2026 awards. First, the €3.4 billion "San" counter-drone wall, an 18-battery system from PGZ with Advanced Protection Systems and Norway's Kongsberg. Second, WB Group's €2.1 billion GLADIUS. Both ran through SAFE, the EU's €150 billion defence loan instrument that is quietly becoming the machine that turns European intent into signatures.
Italy, Turkey and the point about incumbents
The neatest structural shift of the year barely left the trade press. Baykar bought Piaggio Aerospace in June 2025 and stood up LBA Systems with Leonardo, a venture that reached full operations at Farnborough in July 2026. TB2 assembly moves to Villanova d'Albenga, Kizilelma work to Grottaglie. Read it one way and Turkey now manufactures inside Europe; read it another and Leonardo, a legacy prime, just bought itself a fast-drone line. Both are true, and together they are the clearest sign that the incumbents are not sitting still.
The money is in stopping drones
Counter-drone systems account for €15.9 billion across 202 contracts, roughly five times the loitering-munition category, and this is the one segment where European primes clearly lead. Poland's €3.4 billion San wall is the largest counter-drone award anywhere, and it took a Polish prime, a Polish sensor house and a Norwegian effector maker to win it. Rheinmetall's Skyranger 30 has quietly become the continental standard, ordered by the Netherlands in a $1.35 billion deal, then Belgium, Denmark and Austria. Our earlier piece on the €30 billion counter-UAS market sizes the demand; that €30 billion is a forecast of the market, against the €15.9 billion of contracts booked so far. Either way, the winning formula in 2026 is a German gun turret with a Norwegian command layer and a French effector, beating anything a pure-play startup has fielded.
Eurodrone is the warning
The four-nation MALE programme still carries the €7.1 billion contract OCCAR signed in February 2022 for 20 systems and 60 aircraft, shared between Airbus, Leonardo and Dassault. Its critical design review closed only in late 2025, first flight has slipped to 2028, and service entry is not expected before 2031, against a 2016 plan that promised 2025. Airbus and Dassault have fallen out publicly over cost recovery, and France has hinted it may thin its order.
The fair defence is that Eurodrone exists precisely to end Europe's dependence on the American Reaper, that a MALE platform certified to fly in civil airspace is a harder thing than an attritable quadcopter, and that four governments, not the primes, chose the committee that runs it. The fair prosecution is the comparison the programme cannot dodge: the US ran its Collaborative Combat Aircraft from design contracts in 2024 to flying prototypes in 2025 to production awards in 2026, on a broadly comparable autonomous-combat-aircraft class, and expects to field before Eurodrone has flown. Same problem, half the time. Europe can build sovereign capability or it can build it slowly; on this programme it is doing both.
What the pattern is worth to you
Three things follow.
Where you sell decides what company you must be. France concentrates its money in a few enormous programme awards; Britain spreads it across scores of small ones. Winning the first is a decade-long bid campaign, winning the second is a pipeline of £20 million decisions, and the disclosure gap flatters the contrast, so read the medians, not just the means.
Size the market from tender portals and you will misjudge it in both directions, undercounting strike systems and Ukraine-bound volume, overcounting anything a big MALE programme touches. The categories that publish are not the categories where the war is being fought.
And watch SAFE. Poland's two mega-awards and Romania's Quantum buy all run through it. Understanding SAFE eligibility will close deals that better products lose.
Europe is spending like a continent that has finally woken up. But the top of its own contract table is still American, its fastest drone factory is a startup a French fighter-maker had to fund, and its flagship sovereign programme will not fly for two more years. The rearmament is real. Whose flag ends up on it is the part still being decided.
Figures are Drone Consult's tracking of public defence-drone tenders and awards as of 29 July 2026; company valuations are the dollar figures of their funding rounds, contract values converted to euros at rates on that date. Awards with undisclosed values are counted but add nothing to the totals, which is one reason the strike-system numbers understate real spending. Drone Consult sells contract intelligence built on this data.