Everyone knows by now that Europe is rearming. That part is easy to write and carries almost nothing, because “investing in defence” survives any amount of vagueness. The interesting reading sits a little lower down, in the documents that have to name things: how much money, raised in what way, spent on what, through which legal route, under whose command. Those are harder to write and harder to take back, because somebody has to act on them. Read side by side, they sketch a shape, and the shape says more than the announcements do.
None of what follows is secret. It is all published, on the European Commission’s own sites, in the European Parliament’s own papers, in think-tank reports anyone can download. The work is just in putting the pieces next to each other, which is a thing anyone can do with a free afternoon, and which tends to show outlines that no single document sets out to show.
The guarantor steps back
It helps to start with the cause, because for once it is written down rather than guessed at. The Commission’s defence white paper, from March 2025, says it about as directly as an official document ever says anything: the United States, long the main guarantor of European security, now believes it is over-committed in Europe and means to step back. The European Parliament, in its resolution on the same paper, put it more bluntly still, recording that the US “appears to have ceased to be a reliable ally within NATO”. Whatever one thinks of the wording, this is the institutions describing their own situation, not an outsider reading it onto them. Seventy years of assuming the Americans would be there, and then the assumption quietly coming up for review.
What followed was fast, and the speed itself turns out to be part of the story. The scale gives a sense of the register: at the NATO summit in The Hague in June 2025, the allies pledged to spend 5% of national income a year on defence and security by 2035, more than double the old 2% mark, split into 3.5% for core military forces and 1.5% for related things like infrastructure and industry. Whatever else that number is, it is a commitment to spend on a scale Europe has not contemplated since the Cold War.
The fast road round Parliament
There are now two big money instruments, and the difference between how they were made is more telling than either one on its own.
The first is called SAFE: up to €150 billion that the Commission borrows on the financial markets and lends on to member states for buying weapons together. According to the Parliament’s own briefing, it was built on an emergency article of the EU treaties, the kind meant for crises, which needs only the member-state governments to agree and leaves the elected Parliament out of the decision. It is the same article that was used to set up the COVID recovery fund. And it was done quickly: agreed in May 2025, a few months after it was proposed.
The second instrument, EDIP, does much the same sort of thing but went the ordinary way, through the Parliament as well as the governments. That route took about two years, from a proposal in early 2024 to final approval in December 2025.
So the same destination, reached by two roads: one fast and quiet, going round the body whose job is to ask questions, and one slow and noisy, going through it. There is nothing sinister in preferring the fast road in a hurry, and the hurry is real. But it is worth noticing which corners get cut when an institution is in a rush, because the corner that tends to go first is scrutiny. The Parliament noticed too; the same briefing records the worry about being sidelined and about leaning too hard on emergency measures. And the two big instruments are not the whole of it: a defence “simplification” omnibus proposed in June 2025 to thin out the legal and administrative framework, and a military-mobility package in November 2025 to move troops and equipment across the continent faster, both pull in the same direction, toward speed and away from friction. Each single decision to take the fast road looks small enough not to fuss over, and because they come one at a time, the pattern is hard to see until it has quietly become the normal way of doing things, at which point it is much harder to undo than any one decision would have been to stop.
The fence and the open gate
There is a second tension hiding in the financing, and it is about the word “European”.
The €150 billion of jointly borrowed money comes with a string attached: at least 65% of what it buys, by value, has to come from inside the EU, its close neighbours, or Ukraine. Buy European, in other words. The newer EDIP programme sharpens the same idea, capping non-European components at 35% and barring parts from countries whose interests clash with the EU’s.
But that fence only goes round the shared pot. Alongside it, member states have been let off their normal spending limits so they can each spend up to 1.5% of national income a year on defence, which the Commission reckons could add up to something like €650 billion. That is the far bigger number, and it carries no such string at all. Each country decides on its own where the money goes, and nothing stops it going across the Atlantic. So the smaller, shared pot is the one fenced to European industry, and the much larger, national pot is left open. For all the talk of standing on Europe’s own feet, the bulk of the actual spending is left free to wander wherever each capital prefers.
Drones, and the wiring between them
Drop down one more level, to the shopping list, and the vagueness disappears entirely, because here the documents have to name things and put prices on them. The European Defence Fund’s 2025 plan spreads about a billion euros across roughly fifteen budgeted categories, broken down into named items. The single biggest slice, €192 million, goes to ground combat, and one line inside it reads “drone-based affordable mass munitions”, which is procurement-speak for cheap attack drones made in quantity. That is the lesson of Ukraine written straight into a budget. There is money for watching from space, for fighter aircraft, for a “naval combat cloud”, for defending undersea cables, for cyber.
Two threads run through the whole list rather than sitting in one place. One is artificial intelligence and machines that act on their own, woven through the drones and the sensors and the systems meant to tie everything together. The other is exactly that tying-together: a great deal of the money goes on getting separate national systems to talk to each other. That kind of spending is what happens when a single big partner can no longer be assumed to supply the glue. An alliance confident the Americans will provide the connecting wiring does not generally pay to build its own. That one is worth holding on to. It comes back when the shape gets clearer.
Names early, money later
The obvious next question is who pockets all this, and here the open record is honest about its own limits, which is worth being honest about in turn.
The breadth is real and visible. The latest round of fund projects involves more than 600 different organisations across the member states and Norway, and small companies make up well over a third of the participants. The rules push that way on purpose: a project needs partners from at least three different countries to qualify, so the money cannot simply pool in one national champion.
What is harder to see is where, inside all those consortia, the money actually settles. The Commission publishes promptly which companies are in which project, but not how much each of them takes home: the factsheets name the members of a consortium, not the split between them, and they leave out the affiliated entities and subcontractors entirely. A great many small firms taking part is not the same as a great many small firms getting paid, and the figures for who received how much in the most recent round are not settled yet; the independent trackers that piece them together say that will not be done until well into 2026. So the list of names arrives early and the money-trail arrives later and more quietly, which is its own small lesson about which facts a system makes easy to find. For now the honest answer is that the participation looks broad and the concentration is not yet readable, and the gap between those two is a thing to keep an eye on rather than a thing to conclude from.
The official line, and the other one
Money and weapons are the visible half. The harder half, and the more interesting one, is what shape all this is meant to fit into, because a pile of drones and a combat cloud are not a defence; they are parts waiting for an architecture. And here the official line and the working assumption seem to be pulling in different directions, sometimes in the same week.
The official line is steady and often repeated: NATO is the cornerstone, all of this is complementary, nobody is walking away from the alliance. In late January 2026 the NATO Secretary General, Mark Rutte, told the European Parliament that anyone who thinks Europe can defend itself without the United States should “keep on dreaming”. He was specific about the price of going it alone: not the agreed 5% of national income but 10%, and a home-grown nuclear capability costing billions, because the alternative is losing what he called the ultimate guarantor of European freedom, the US nuclear umbrella. That is the position as it appears on the news, and it is not posturing; it is largely true. America still supplies the bulk of the alliance’s actual fighting strength, something like 75,000 troops still stationed across Europe, and the Supreme Allied Commander Europe, NATO’s top military officer, has been an American without exception since the post was created in 1951.
And yet the official choir does not sing in unison. A few weeks later, at the Munich Security Conference in February 2026, the European Commission’s own president, Ursula von der Leyen, took the opposite tack, calling for Europe to bring its own mutual-defence clause to life and stating flatly that mutual defence is not optional for the EU but an obligation under its treaty. The head of the EU’s executive, pressing in public for European self-reliance days after the head of NATO told Europe to keep dreaming about it. Two senior figures, two halves of the same coin, held up at once. The think-tank world has a useful name for the resulting condition: “Schrödinger’s NATO”, an alliance in which the United States appears both committed to it and absent from it at the same time, formally inside while behaving as though it were halfway out the door. That is, more or less, the situation the news already shows, given a name.
The Americans, for their part, also made some noise. The US National Defense Strategy published on 23 January 2026 reordered priorities openly: defending the homeland and deterring China sit at the top, and everything else, Europe included, is secondary. The document tells European allies they will take primary responsibility for their own defence, with what it calls more limited US support. Read alongside the European white paper from 2025, the two sides are describing the same change from opposite ends: Washington announcing the step back, Brussels registering it.
The tells in the structure
If the official line were the whole truth, the spending and the institutional fidgeting would not make much sense. Nobody prepares for a thing they are confident will not happen. So the more revealing reading is to treat the words as the reassurance and the structure as the intention, and to ask what the structure is quietly getting ready for.
The clearest single tell sits at the very top of the command chain. There has been open, serious discussion of appointing a European to the role of Supreme Allied Commander Europe, the post an American has always held. There is no legal obstacle; the founding treaty never specified the commander’s nationality. On its own it is a personnel question. As a signal it is louder than that: nobody rehearses handing over the top military job without imagining a day when the current holder’s country is no longer the one in charge.
The second tell is nuclear, and it is the sharpest of all, because nuclear deterrence is exactly the thing the United States currently provides and exactly the thing a Europe-without-America would have to find for itself. Rutte named this himself, as the clinching argument for why going alone is folly: lose the Americans and the umbrella goes with them, and replacing it costs billions.
The striking thing is that the conversation he meant as a warning is already under way. At the Ile-Longue submarine base on 2 March 2026, the French president set out a doctrine of “forward deterrence”, extending the idea of French nuclear protection to its neighbours and announcing the first increase in France’s warhead count since 1992. He had already raised aloud, earlier in the year, whether the survival of France’s closest partners might be bound up with France’s own vital interests, which in the careful language of nuclear policy is a very large question to ask at all. This did not come from nowhere: France and the United Kingdom had signed a declaration in July 2025 pledging to deepen their nuclear coordination, and there is now open talk of formalising nuclear discussion among Europe’s main military powers, the so-called E5. Nobody builds a shared nuclear conversation as a hobby. It is the kind of thing that only becomes worth the political pain if the American umbrella can no longer simply be assumed, which is to say: the very scenario Rutte called dreaming is the one the nuclear talk is quietly costing out.
The third tell is the one already sitting in the procurement. The money going into a European combat cloud, a European reconnaissance constellation in space, the wiring to make separate national armies talk to one another, all of that is the material version of the same hedge. These are the connective and strategic functions the United States has supplied for decades. Buying them in-house is, in budget form, exactly what preparing to operate without the American backbone looks like. The white paper can call NATO the cornerstone in its text while the annexes quietly fund the means to stand without it. Both things are true at once, which is what Schrödinger’s NATO means in practice.
And then there are the shocks that keep nudging the hedge from prudent option toward working assumption. The clearest was the early-2026 pressure on Denmark over Greenland, with Washington threatening tariffs and pointedly declining to rule out force against the territory of a fellow NATO member. When the supposed guarantor of the alliance starts leaning on a member of it, the question of how to manage without that guarantor stops being theoretical for everyone watching.
A NATO that can run on European hands
It would be easy, from all this, to leap to “Europe is building its own army” or “Europe is leaving NATO”. Neither is what the documents show, and the most recent and most detailed reading of the shape, a June 2026 policy paper from the European Council on Foreign Relations, is worth following because it is unusually clear-eyed about what will and will not work. It frames the task within a window of five to seven years, the period in which European generals reckon the risk of Russian aggression is highest, which is the clock the whole effort is running against.
What will not work, on its own, is a tidy grand design. A single integrated EU army, or an EU nuclear deterrent, puts the cart before the horse: the EU has never been given the command structures, the operational authority, or the mandate to run a large military operation, and its members are not about to hand those over. Decisions about sending people to die remain national, and will stay national. Nor can the EU’s own mutual-defence clause simply be swapped in for NATO’s Article 5; the machinery behind it does not exist. So the futuristic versions, the ones that make the headlines, are the parts least likely to happen.
What the paper sketches instead is more modest and more plausible, and it is the shape worth watching for. It has three layers, each built from something that already exists rather than from a grand new institution. NATO’s command structure stays as the operational backbone, only reorganised into sub-regional coalitions of European countries that share a threat and can act together quickly. The EU supplies what it is actually good at: the money, the industrial coordination, the pan-European legitimacy to hold the effort together. And flexible coalitions of the willing provide the speed, the ability to act when the formal institutions stall, for instance if triggering Article 5 gets delayed or blocked. The whole arrangement is designed, in the paper’s own summing-up, to let Europeans defend themselves with America where possible, with less America where necessary, and without America if it comes to that.
That formula is the shape, stated plainly. Not separation from NATO but a NATO that can run on European hands if the American ones let go. Not a new superstructure but a re-wiring of the existing one so it keeps working through a delay, a blockage, or an absence at the top. The drones and the combat cloud and the talk of a European commander and the careful French sentence about partners’ survival are not separate stories. They are the same preparation seen from different angles: the build-up of a defence that can stand on its own the day it has to, while saying, truthfully and for now, that it hopes not to have to.
Where the reading runs out
The budgets say what is being paid for, not whether it will ever be built, or work, or arrive in time; defence spending has a long history of money committed and kit not delivered. The headline €800 billion is mostly not EU money at all but hoped-for national spending that no one is actually obliged to make, so it may be the most quotable and least certain figure in the whole business.
The shape, too, is a reading and not a certainty. The same facts could be told as a story of loyal burden-sharing inside a healthy alliance, the Europeans paying their way, exactly as Washington has demanded for decades, with the hedging amounting to no more than sensible insurance. That telling is available, and parts of it are surely right. Which story turns out to be the true one depends on things that have not happened yet: whether the American troops stay, whether Article 5 is ever tested, whether the coalitions hold when frightened. The documents fix the preparations, not the outcome.
And none of it can reach the one thing most worth knowing, which sits furthest from any invoice or org chart: whether all this makes Europe safer, or only better armed, and whether a continent rehearsing how to fight without its old protector is steadying a dangerous moment or helping it along. A budget can say what a thing costs. A policy paper can sketch a command structure. Neither can say whether the whole endeavour was wise, and anyone claiming otherwise is doing the speeches’ trick again, this time with footnotes instead of fine words.