The Planes Arrive in 2028

Must catch up. That's what you write when you've decided the lag is a failure of will — that somewhere in Brussels there is a person who has not yet been sufficiently alarmed, and that alarming them is the intervention. It's a comfortable diagnosis, because it implies a cheap fix: better briefings, louder scientists, one more emergency summit with a good photograph at the end.

The lag is not a failure of will. Everyone involved is alarmed. The lag is a set of clock mismatches — structural, describable, and one of them has a delivery date you can look up.

i · the delivery date is the whole argument

A fire near Bordeaux moved through more than 420 square kilometres in under four days this week. The Madrid region is having the worst fire season in its recorded history. That's the physical clock: hectares per hour, a front that outruns the evacuation order.

Here is the institutional clock. In March, the European Commission published a new wildfire strategy — prevention, preparedness, response, recovery — built around a permanent rescEU firefighting fleet: twelve aircraft and five helicopters, roughly €600 million, stationed across France, Spain, Portugal, Italy, Greece and Croatia. The first helicopter reached Romania in January. The first aeroplanes are expected in early 2028.

Read that again with the fire in the frame. The considered European answer to a record fire season lands the better part of two years after this one ends, into a season nobody can specify yet, procured against a threat model written when the 2025 numbers were still being tallied.

None of this is a scandal. Every step is defensible in isolation. Amphibious water-bombers are a niche airframe with a supply chain that was allowed to go cold and has to be restarted; procurement at EU scale runs through budget instruments measured in seven-year blocks; aircraft take years to build. The 2026 seasonal posture — 777 firefighters pre-positioned across fourteen countries, twenty-two planes and five helicopters on standby, a new regional station opening in Cyprus — is genuinely the largest mobilisation of its kind Europe has assembled, and it was assembled by people doing their jobs well.

The uncomfortable part: this is what competence looks like when the clock is wrong. There is no villain to fire. Hold onto that, though, because "no villain" and "no choice" are not the same finding, and the distance between them is where this goes.

ii · every institution is a fossil of its last emergency

The pattern is older than the climate file, and the clean cases are the ones that worked.

January 31, 1953. A North Sea storm surge breaks the Dutch defences; 1,836 people drown in a night. The Netherlands responds with the Delta Works — one of the most impressive pieces of civil engineering any democracy has ever executed, specified against a storm expected once in ten thousand years. It takes roughly four decades to finish.

Same storm, English side: over three hundred dead, and a decision to defend London. The Thames Barrier becomes operational in October 1982. Twenty-nine years from the flood to the first closure.

These are the success stories. This is the fast version.

But be careful what those two actually measure. A sea barrier takes decades because sea barriers take decades. That is a construction schedule wearing a governance costume, and if you read it as the speed of democratic adaptation you will draw the wrong conclusion — roughly the conclusion the "must catch up" genre draws every time.

Strip out everything that has to be poured or welded, and the institutional response times in this story are fast. Pedrógão Grande, June 2017: sixty-six dead in Portugal, most of them in cars on a road that became a flue. Within about a year there is a new integrated fire-management agency, restructured command, real money. Mati, Attica, July 2018: a hundred and two dead in Greece. Reform follows there too. The 2026 seasonal posture was assembled inside a single planning cycle. Twelve to twenty-four months, repeatedly, across different states and legal systems.

So the cadence claim has to be narrower than "institutions are slow" — and the narrower version is worse news. Doctrine, staffing and coordination re-specify roughly on the threat's clock. Capital hardware cannot. It re-specifies on the build clock, which runs five to fifteen years, which means every aircraft that lands is an answer to the world as it was when the order was signed. You cannot run a decade-long build cycle against a twelve-month threat cycle, and the institution never finishes catching up, because "up" keeps moving.

And in 2023 Greece nonetheless records the largest single wildfire in EU history. In 2025, on the EFFIS figures, the Union loses over a million hectares — the worst season on record. In 2026 Portugal, France, Spain and Greece are all asking for help again.

Every one of those reforms was real. None of them was wrong. All of them were specified against the fire that had already happened. The fast reforms bought a season. The slow money bought an airframe for a decade nobody has modelled yet.

iii · yesterday's coherence is today's fuel

The Americans ran the extreme version of this experiment, and the result is worth staring at, because it shows the lag doing something worse than arriving late.

August 1910: the Big Blowup burns roughly three million acres across Idaho and Montana in two days and kills eighty-odd firefighters. The young Forest Service, fighting for its existence, draws the obvious conclusion — fire is the enemy — and by 1935 has codified it as the 10 a.m. policy: every reported fire to be controlled by ten o'clock the morning after it's spotted. Total suppression, as doctrine.

It worked. That's the problem. It worked for decades, in landscapes that had burned lightly and often for millennia, and the accumulated result was a continent of forest carrying three to ten times the fuel load it evolved to carry. The policy that was a coherent answer to 1910 became, by slow compounding, the primary manufacturer of the megafire.

That's the mechanism worth naming, and it's more than institutional slowness. A structure doesn't have to change to become incoherent — the environment can move underneath it while every rule, budget line and command hierarchy stays exactly where it was. Coherence isn't a property a system owns. It's a relation between a structure and the world it's acting on. When the world shifts, yesterday's fit becomes today's distortion without anyone doing anything wrong, and the distortion doesn't just sit there. It accrues. It becomes the fuel.

Europe's version of the 10 a.m. policy has a file number. It is the Common Agricultural Policy — half a century of payment structure written to answer post-war food insecurity, which it did, spectacularly, and which in the doing made smallholder grazing and terrace-farming on marginal Mediterranean land uneconomic. The depopulation that followed is usually filed under demography, as though people simply drifted off. They drifted under a price signal. Scrub and unmanaged fuel filled in behind them, where goats and olive terraces used to keep the landscape discontinuous.

Nobody drafted a fuel-accumulation programme. A fuel-accumulation programme is what fifty years of correctly-executed agricultural policy produced, on a continent whose climate walked out from under the policy while it was still running.

iv · prevention doesn't photograph

So why does the money still go mostly to the response side?

Because the political return on a water-bomber is immediate, visible, and attributable. There is footage. A minister can stand in front of it. When it drops on a burning ridge the causal chain from public spending to public good is one link long and legible from a helicopter.

The return on prevention is a fire that does not happen. There is no footage of that. Controlled burning, fuel-break maintenance, landscape-scale mosaic management, the unglamorous work of paying people to live in and tend marginal countryside — these produce an absence, in someone else's electoral term, and they produce it while generating smoke complaints and angry op-eds in the present one. Worse: they occasionally go wrong in a way that is photogenic, and the official who authorised the burn owns that outcome personally in a way no official owns the aggregate fuel load.

The Commission's own strategy leads with prevention. Read it — the sequencing is correct, the science is correct, the four pillars are in the right order. Then watch which pillar shows up in the press releases, and which one has €600 million and an airframe order attached.

Nobody is corrupt here. The gradient does the work. You don't need a conspiracy when the incentive structure reliably converts good analysis into visible hardware.

Which retires the comfortable half of my own argument. The clock is a constraint — real, physical, five years of airframe. The gradient is a choice made inside the constraint, by people, plural, each of them reasoning correctly on the incentives in front of them. So the lag has no villain and it has authors. And the scandal was never that aeroplanes take five years to build. It is that the one intervention which structurally cannot arrive in time is also the only one that photographs — so it wins the budget, and the delay it guarantees gets reported as a law of nature.

v · the cost of having to ask

One more piece of architecture, quieter than the rest.

Civil protection is a member-state competence. The EU supports and coordinates; it does not command. The Civil Protection Mechanism activates on request — a national government has to formally ask.

Asking is not an administrative act. It is a political statement that your own capability has been exceeded, made by an interior minister on a day when the opposition is already asking why the fleet wasn't pre-positioned. There is a delay between the moment the request would help most and the moment it becomes politically survivable, and that delay is not measured in paperwork. It's measured in hectares.

Subsidiarity was designed to keep decisions close to the affected. Against a phenomenon whose smoke crosses three borders in a day and whose fronts merge across a peninsula, the same design distributes the decision to whoever is most reluctant to make it. The fire is continental, the sovereignty is national, and the seam between them is where the delay lives.

Which is the previous section again, wearing different clothes. Electoral-term returns, request-based activation, seven-year budget blocks, national competence — every one of these mechanisms weights the present, the national and the enfranchised, and structurally under-weights the future, the downwind and the unrepresented. A minister answers to voters this term. Smoke crosses three borders in a day. The fuel load matures in a decade. The people in the valley below the front have no seat in the decision about whether to ask.

Subsidiarity keeps decisions close to the affected only if the affected are governments. Define them instead as the people the fire actually reaches, and the arrangement inverts: the decision sits furthest from those who bear it. That is not a broken coherence. For the parties in the room it works exactly as specified. Which is the harder thing to say, and the truer one.

vi · the clock you can't shorten

Three clocks so far: the fire's, the build cycle's, the electoral term's. There is a fourth underneath them, and it is the one that matters, and it has been hiding inside a phrase I used earlier without inspecting it — procured against a threat model written when the 2025 numbers were still being tallied.

A democracy cannot legitimately act on a fact until the fact has been counted, published, contested and accepted. EFFIS closes a season's figures well after the season ends; the model revision follows the figures; the budget follows the model; the airframe follows the budget. Ratification is itself a clock, it runs in years, and every schedule downstream inherits its lag.

Unlike procurement, you cannot simply speed it up. Procurement is slow by accident — supply chains, tooling, cold production lines — and accidents can be fixed with money and standing orders. Ratification is slow on purpose. The counting and the contesting are what make the resulting action bind; an institution that acts on unratified reality has just described every panicked state decision in the historical record. Nobody wants to redesign this clock because the fast version already has a name, and the name is worse than the lag.

That's the real bind, and it isn't cynicism — it's a design problem with an actual shape. The legitimacy that permits an institution to act is manufactured by a process slower than the thing it must act on. Which leaves the unglamorous moves: authorisations that trigger on live data and expire on their own, capacity procured against ranges instead of point estimates, orders hedged before the final number lands. None of these abolish the lag. They stop it from being inherited by everything downstream.

vii · what 2032 looks like

Predictions, offered at the usual odds:

The twelve aircraft will arrive from 2028. They will be excellent. They will be specified against conditions modelled around 2025, and by 2028 those conditions will read as a mild year. The fleet will be described on delivery as long-overdue, which will be accurate, and as transformative, which will not.

After this summer, a European reform package will be announced in September or October, with prevention foregrounded in the text and response funded in the annex. It will be slotted into the next budget cycle. Its physical output will land in the early 2030s.

And around 2032 there will be another piece, in another publication, with the same headline: politics must catch up with the new reality.

The reality won't be new. It will be the fourth consecutive new reality. But the phrase is load-bearing — it locates the failure in insufficient urgency rather than in a set of clocks nobody has proposed rebuilding. And rebuilding them means funding outcomes that never photograph, ordering hardware against a number that isn't final yet, and moving a competence away from the governments that currently hold it.

None of which photographs well either.