The Trial of the Little Boys' New Toys

I. Trial

Since July 2025 a programme has been running in Ukraine under the name Test in Ukraine. It is not hidden. It is not unofficial. It is operated by the Ukrainian defence complex Brave1, has its own website, and invites international arms manufacturers to send their products to the front for testing. Drones, anti-drone weapons, naval drones, unmanned ground vehicles, AI systems, electronic warfare. The manufacturer supplies the goods, the Ukrainian military deploys them, and the measured results flow back to the manufacturer.

The language of all this is the language of product development. Faster iterations. Scalable platforms. Modular architectures. Performance under real conditions. These are the terms a corporation uses to describe a software beta phase. With the single difference that the test environment is not a virtual sandbox but a battlefield, and the test subjects are not contracted developers but soldiers who are paid to shoot — and who sometimes die before they can fill out the feedback form.

It is the gold standard of weapons development. No one else has this data. A manufacturer who tests a weapon in a NATO exercise gets data under controlled conditions. A manufacturer who tests a weapon in the Ukraine war gets data under the toughest conditions imaginable — against an active adversary who is developing countermeasures, against real-time electronic warfare, against real logistics, real weather, real fatigue. This is the one test that no exercise can simulate.

Those who develop weapons have historically had one problem: they never knew whether they worked. Only deployment brought the truth, and deployment often lay decades behind development. The tanks of the First World War were further developed in the thirties; their fitness was first tested in the Second World War. Jet engines became operational in the early forties; their first real stress test came in Korea. A whole generation of arms development passed between concept and truth. Today, weeks pass. What leaves a German factory on Wednesday is in deployment on Sunday, analysed on Monday, corrected on Thursday. The iteration cycles of the arms industry have reached the tempo of software development. What the sector needed for this was a war.

II. Optimisation

Test in Ukraine is not a binary test — does it work or doesn't it. It is a closed loop. The manufacturers train the Ukrainian military online in the operation of their equipment. The military deploys the equipment. It measures hit rate, failure rate, range, effect, behaviour under jamming, behaviour in the cold, behaviour in wet conditions, behaviour against the latest Russian countermeasures. It sends the measurements back to the manufacturers. The manufacturers iterate. They send the next version. The cycle begins again.

The Ukrainian soldiers are in this loop not merely users. They are unpaid co-developers. They report what is jamming on a drone, what is inaccurate on a sight, what is failing on a camouflage skin. They do not write engineering drawings. But they supply the data stream with which engineering drawings and products are optimised — battlefield-capable and shell-resistant are the industry's terms for it. Whoever designs at a desk can calculate the one and simulate the other. Whoever designs in the drone war gets both delivered first-hand, without having to count the dead who died for the delivery. That the iteration ends at the cemetery when it fails — this belongs to the cost structure of the programme without anyone putting it that way.

A small British firm, Occam Industries, sent its AI software for drones to Ukraine. The software worked. The hardware it ran on was not suited to the front. Brave1 brought the British firm together with Ukrainian manufacturers. The software now runs on drones built in Ukraine and is approaching the next engagement. This is how the value stream reads today. A start-up in London, a piece of code, a workshop in Zaporizhzhia, a drone, a hit, a dead person, a data return, an update. Seven stations, between which no human being is alive who would bear responsibility for the whole. Each is responsible for his own segment. For the whole no one is responsible — if everyone thinks of himself, all are taken care of.

III. End use

A weapon that lies in the warehouse is a balance-sheet position. A weapon that is consumed is a position that must be replaced. This is the simplest truth of the arms economy, and it is rarely spoken so directly. Whoever produces weapons has one primary interest: that the weapons produced are consumed. Peace is no business model for the arms industry. War is. A long attrition war, in which stocks have to be continuously replenished, is the ideal business model.

The end use of a weapon or weapons system is reached only when either the enemy's infrastructure has been damaged as severely and lastingly as possible, or as many so-called soft targets as possible have been eliminated. Both, preferably. Here too the usual cost-benefit calculation applies, as in normal business — evaluated through corresponding reconnaissance assets. War is not only the continuation of politics by other means, but also the continuation of business by special means towards the usual economic objectives. The quality and efficiency of products, the competitive situation, the distribution channels — these are here too essential factors for the return ultimately to be achieved.

Statista's records hold this state of affairs in the dry tone of an audit. Since the outbreak of the Ukraine war, demand for armaments in Germany and internationally has risen sharply. European states are rearming and must replenish their own stocks that they have transferred to Ukraine. This brings major new contracts not only to German arms manufacturers. — Three sober sentences in which the entire economy of war stands. Double sale: once to Ukraine, once to the supplying countries as replenishment. From one market becomes a self-renewing market.

Rheinmetall, as the most important German beneficiary, has communicated its strategy openly: four factories in Ukraine, a potential of two to three billion euros in annual turnover from this location alone, and a total turnover that is to rise to ten billion euros. Production in Ukraine is not an act of solidarity. It is a site decision. The Ukrainian self-propelled howitzer Bohdana costs 2.5 million dollars — half the price of the French Caesar. The Ukrainian anti-tank system Stugna-P is exported for between twenty and fifty thousand dollars; the American Javelin costs between one hundred and seventy-five and one hundred and seventy-seven thousand. Ukrainian labour costs account for a factor of two to eight. Whoever has production done in Ukraine produces more cheaply. Whoever produces more cheaply sells more.

The economic logic carries itself. No one needs to be evil for the system to function exactly as it does. It functions because each actor rationally pursues his interests: the manufacturers want to sell, the governments want to equip, Ukraine wants to defend itself, the shareholders want returns, the politicians want to show that they can act. Each one of these wishes is legitimate. The interplay is a machine that produces, consumes, replaces, and again produces weapons. The machine runs.

IV. The little boys

Whoever follows the German defence debate of 2026 hears a language that for a moment sounds familiar, and then suddenly no longer does. It is the language that men in middle age and beyond speak when they have been given a new toy. Friedrich Merz in the Bundestag on Taurus deliveries. Roderich Kiesewetter on ARD on the necessity of countering Russian escalation. Boris Pistorius at the Security Conference on the war-ready Bundeswehr. It is not the language of the old diplomacy. It is not the language of security analysis. It is the language of the nursery, in which the words that once came from war are now being used for the first time by a generation that knows war only from television.

Rheinmetall CEO Armin Papperger himself supplied the vocabulary. He compared Ukrainian drone production to playing Lego and spoke of housewives with 3D printers. This is no slip of the tongue. It is the inner truth of an industry that cannot bear what it does, and therefore flees into infantile language. Whoever speaks of Lego does not have to think about what a Lego brick does when it strikes a human dwelling at a hundred and fifty kilometres per hour. Whoever speaks of housewives does not have to think about what it means when those housewives at 3D printers are themselves producing the next drone carriers for their dead sons.

The generation of the political class of 2026 has, in its majority, not served. It has no experience with ammunition, with weapons, with what a weapon physically does. It knows the vocabulary from briefings, from talk shows, from the PowerPoint slides of defence experts. It plays with terms like strategic depth, escalation dominance, deterrence effect, without these terms being measured against any experience of one's own. When Friedrich Merz speaks of the defence-readiness of Europe, it sounds like a child saying aloud for the first time a word from a textbook, because he wants to look clever in front of the class.

This is no indictment. It is an observation. A generation that grew up without service of its own has two possibilities of dealing with the return of war. It can become aware of its own incompetence and be cautious. Or it can replace the missing experience with loud language. The German political class has, in its majority, chosen the second possibility. The consequences are to be observed in the talk shows — and in Rheinmetall's balance sheet.

V. The certifiers

Whoever looks at the rearmament line of the generation of 2026 might take it for a men's affair. It is not. The class that is now rearming has become gender-neutral, and precisely in this lies its hardness. Two women supply the necessary complementation. They stand for two functions that the loud male chorus alone cannot fulfil.

The loud variant is Marie-Agnes Strack-Zimmermann. She was Chair of the Defence Committee in the Bundestag from December 2021 to June 2024. Since January 2025 she has been Chair of the new Security and Defence Subcommittee of the European Parliament. She is simultaneously a member of the Presidium of the Friends of the German Army (Förderkreis Deutsches Heer), which, alongside the Society for Security Policy and the German Society for Defence Technology, ranks among the most important lobby associations of the German arms industry. Members of the FKH are politicians, senior military officers, and most German companies in the security and defence sector.

This is no accidental multi-position. It is an institutional interlocking. The Chair of the parliamentary defence committee sits on the Presidium of the arms lobby. This is the classic revolving-door finding, in its most compact form. What the arms industry needs is not a politician who openly represents its interests — that would be corrupt. It needs a politician who represents its interests on moral grounds. Strack-Zimmermann supplies this. She speaks of the threat from Russia, of the duty of protection towards Ukraine, of the historical responsibility of Germany — and the result is that Rheinmetall builds more tanks. Whoever hears her speeches does not think of lobbying. He thinks of morality. This is the function.

A point that completes the picture: in January 2023, Strack-Zimmermann was one of three initiators of a cross-party Parliamentary Circle for Feminist Foreign Policy in the Bundestag, alongside Michelle Müntefering of the SPD and Agnieszka Brugger of the Greens. What looks like a contradiction to her rearmament role is none. When in 2024 the weekly ZEIT offered her a contested dialogue with Alice Schwarzer on sexual violence in war, she declined — according to Schwarzer with the reason that this was not really her topic. That is the operation in question. The feminist label is taken on, the feminist content is not. Whoever demands the delivery of cruise missiles no longer has to walk the world as a cold warrior. He — or rather she — can appear as a progressive feminist voice, without engaging in the more strenuous work of feminist politics. The left, which would have liked to speak out against more weapons, is thereby placed in the camp of backwardness. This is the rhetorical operation. It works.

The structural variant is Franziska Brantner. On 14 May 2026, the Green party leader gave a keynote speech at the University of Oxford. Her proposal is to bind the emerging new Bundeswehr so deeply into European command structures that a future German government can no longer direct it independently. Binding multinational structures should not be capable of being simply dissolved by a future government in Berlin. The Germany of 2026 must, she says, impose limits in advance on the unpredictable Germany of 2035.

Here a distinction is necessary which Brantner's proposal blurs. If Europe is to become a power to be taken seriously, then it does indeed need structural binding. A common defence that can be revoked after every national election in a member state is no common defence. Whoever says A — shared command structures, joint procurement, a headquarters of the navy in Gdańsk, as we have proposed in our security paper A Security Architecture with Precisely Defined Tit-for-Tat Strategies among other things — must say B — binding that is serious. Brexit was a decision that became possible with a narrow majority, without a two-thirds requirement, without a confirmatory vote, out of momentary discontent. If Europe wants to become a power to be taken seriously, it must afford protection against such decisions. If the Russian leadership changes tomorrow and another reformer like Gorbachev appears — does every defence structure then again become obsolete, and Europe returns to peace, joy, pancakes and the peace dividend, only to stand again without substance in the next conflict? The answer cannot be that every adjustment leads immediately to dissolution. Structural bindings are the precondition for adjustments to be possible without everything collapsing and far-reaching decisions being taken according to the prevailing situation and mood.

But the form of the binding is decisive. A democratically grounded binding — through referenda in each founding state, through a constitution with an exit procedure, with qualified majorities and waiting periods — is hard to dissolve but not impossible to dissolve. It is what Article 50 is for exit from the EU: a high hurdle, a clear procedure, no act of fate. This would be the binding Europe needs as a serious power. It is not what Brantner proposes.

Brantner binds not the citizens to a democratically ratified architecture, but wants to bind the future majority to the current policy. What speaks from her proposal is a paternalistic attitude. She mistrusts the people and those then in responsibility. She expects that the off-centre forces — AfD on the right, BSW or similar on the left — will no longer stand outside the governing majority in 2035, and wants to place a cuckoo's egg in their nest today, which they will no longer be able to throw out tomorrow. Instead of trusting herself with the democratic confrontation with these forces, she wants to prevent the confrontation before it takes place. This is not only undemocratic in its form. It is also politically unwise. A binding that is grounded only in mistrust of one's own electorate becomes unstable the moment the electorate recognises it as such. It produces precisely the resistance that it is meant to prevent. Brantner's cuckoo's egg will reveal its character the moment another bird sits in the nest. And then the whole construction will be put up for disposition — not because the binding is bad, but because it came about undemocratically.

Brantner's political binding and the Scholz-Merz special fund — rearmament on credit — bind the next generations irrevocably with an outright outrageous hubris of those responsible today. In essence, the most probable failure of the coming generations is being programmed today, as we set out in our essay The Structural Redistribution Between the Generations.

Those who today bear responsibility, and those who voted for them, live by running down the substance that earlier generations built up, without renewing it. At the same time, they burden future generations in a way that takes away from these all room for manoeuvre. This is the actual point of the political class of 2026. They consume what they inherited and bind in chains what is yet unborn. Both in one move. Both without the faintest trace of consciousness that this is the opposite of responsibility.

The two women supply different functions. Strack-Zimmermann supplies the loud, morally cloaked certification of rearmament. Brantner supplies the structural fixing. Together they yield the female complementation of what the little boys loudly perform. That women take on these functions is not an act of emancipation. It is the closing of a gap. Whoever wants to play along in the political class of 2026 must take on the language of weapons, irrespective of gender. The bridge between the generations thereby becomes a bridge without a way back.

VI. Erhard's sentence

In Experiences for the Future, Ludwig Erhard's manuscript composed in 1976 and published posthumously in 2024, stands a sentence that now takes on an entirely new meaning: We must not leave to the United States the principal burden for the security of Europe. Erhard meant this in a very specific way. He meant: Europe must have its own strategic conception, its own economic basis for it, its own political independence. He did not mean: we must buy more weapons, take out more special funds, produce more rearmament rhetoric.

Erhard knew why he wrote the sentence. As Federal Chancellor he had experienced the structural dependence of the Federal Republic on Washington — the fact that Adenauer owed his power to the Western tie, that the Federal Republic was never really sovereign until 1990, that the American guarantee was simultaneously protection and cage. He saw in 1976 what would become consensus twenty years later: that this constellation does not hold forever. And he wrote down what should be done before it collapses.

What should have been done has not been done by Germany in fifty years. It has not developed its own strategic conception. It has not built up its own technological basis for its security. It has delegated its energy supply first to Russia and then to America. It has no digital infrastructure of its own, no space programme of world rank, no semiconductor industry of its own. It has delegated everything and noticed only when the supplier — Trump — raised the price that it can no longer do anything by itself.

What today's debate takes for a European responsibility is not Erhard's conception. It is its caricature. Erhard would have said: let us build a country that can think for itself, develop for itself, decide for itself. Today's class says: let us buy more munitions. Erhard would have said: let us create the conditions under which we are no longer at anyone's mercy. Today's class says: let us fix the structures so that no one can ever again leave the West. Erhard would have said: let us become free. Today's class says: let us be finally bound.

It is the exact inversion. Erhard's security policy would have been liberation. Today's is self-binding. Erhard's security policy would have built up its own substance. Today's buys American and Ukrainian substance and labels it the solution. Erhard's security policy would have been an adult answer to the end of the Cold War. Today's is the answer of a class that never had to grow up because someone else always took the burden off its shoulders. Now no one takes anything off it anymore. And it notices that it does not know how it works.

VII. What remains

A small class of men and women in middle age is putting the new toy through its trials. It does so with the money of the taxpayers, with the lives of foreign soldiers, with the balance sheet of Rheinmetall, with the structures that the next generation must no longer be allowed to dissolve. It does so without experience of its own, without conception of its own, without responsibility of its own for what will stand at the end.

The machine runs. The orders grow. The share prices rise. The drones find their end use, the replacement need arises, the next sales cycle begins. The talk shows fill with the vocabulary that makes the matter bearable — defence-readiness, alliance loyalty, deterrence, war-ready. The books of Erhard lie in the storerooms of the city libraries, because no one was borrowing them after two years. Whoever had the courage to read them today would find that all that is happening now has been foretold.

It is no longer necessary to be outraged. Outrage would be the noise that the machine needs in order to keep running — it shows the machine that it is having an effect. What the machine cannot bear is the quiet description of what it is. A small class, with small toys, with large consequences that others will bear. A class that does not know that it is a class, and therefore believes it is acting out of responsibility while it is acting out of habit. A generation that passes its burden to the next because it itself received none. A democracy that insures itself against its own future, because it no longer trusts the present.

Erhard knew that this would come. He wrote it down in 1976. No one read it. Now it stands in a book that lies in the Nuremberg library's storeroom, and that, after fifty years of postponement, twenty-five years of being forgotten, and two years of not being asked for, waits for someone to order it.

Someone is just ordering it.