The occasion for this text is a small conference and a large finding. At the end of April 2026, some fifty states met in Santa Marta, Colombia, as a “coalition of the willing” — a mini-summit intended to breathe new life into the deadlocked UN climate negotiations. The powerful were not there. Not the USA, not China, not India, not Russia, not the Gulf states. Ten countries that together cause more than sixty per cent of global emissions stayed away. Germany was invited — the Federal Environment Minister stayed at home and sent a state secretary. The declaration was signed. At home, Germany is pursuing a course that undermines the declaration.
The gesture is familiar. It is not new. It is the gesture of the last thirty years.
I. Thirty Years of Bookkeeping
In 1995 the first Conference of the Parties to the Framework Convention on Climate Change opened in Berlin. At that time global CO₂ emissions stood at around twenty-three gigatonnes a year, and the atmosphere contained about 360 ppm. Today, thirty conferences later, emissions stand at around thirty-seven gigatonnes a year, and the atmosphere at 425 ppm. The global mean temperature has risen by nine tenths of a degree. The curve shows not a single turning point that could be attributed to the conferences — no dent, no plateau, no kink that could not equally be explained by other causes (economic crises, falling technology costs, the COVID lockdown).
That is not to say the conferences have achieved nothing. The Paris Agreement of 2015 established the 1.5- and 2-degree targets, which have since served as the reference point in almost every climate-policy dispute. The Loss and Damage Fund has been set up, if inadequately filled. The Glasgow methane pledge has led to concrete programmes in some countries. The cost curve of renewable energy was accelerated by international coordination. All of these are real results.
But measured against the declared mission — stabilising the climate at a non-dangerous level — the record is devastating. In thirty years not one single gigatonne of reduction has been pushed through against the trend that would remain once other explanations are taken away. The curve has gone on rising, unimpressed by the conferences. Anyone who compares this with the claims of the process will find that the gap can no longer be explained as “incremental progress”. It is structural failure.
II. The Format as a Structural Trap
Why does it not work? The first and most obvious reason lies in the procedural rule: consensus. Every country has a veto. In Belém at the end of 2025 the word “fossil” could not be included in the final document because Saudi Arabia and Russia objected. A conference that is not allowed to name the central problem of its mission has not given up its mission — it has placed its negotiating rule above it.
The consensus rule is no accident. It was designed in 1992 to bring every country to the table. The assumption was that climate protection would not seriously threaten the strategic interests of any country. That assumption was wrong. For oil-exporting states, serious climate protection means the end of their business model. For industrialised states, it means the costly rebuilding of infrastructures that were built for a fossil logic. For catching-up states, it means giving up development paths that were open to the industrialised countries.
If climate protection harms interests, and the procedure gives a veto to those whose interests are harmed, climate protection will not be decided. That is not a question of will. It is a structural property of the configuration. It works reliably, like a well-designed mechanism.
The answer of the “coalition of the willing” in Santa Marta is the form in which it is admitted that this mechanism has failed. But the alternative — a smaller, voluntary coalition — immediately runs into the next problem: it has no mandate, no binding force, no power to impose sanctions. What it decides applies to those who were willing anyway. The powerful, who cause sixty per cent of emissions, remain outside. The small coalition can assure one another that it is serious. On the atmosphere this has no effect.
III. The Mechanism of Normalisation
If the format does not solve its declared task — why is it continued? This question leads to the actual function of the conferences, which lies beneath their declared mission.
The format has become decoupled from its function. It has become self-perpetuating, with its own logic, its own profession, its own careers, its own media rhythms. A COP brings together fifty thousand participants. Every delegation has its tasks, its criteria of success, its internal yardsticks. NGO staff follow their issues through the negotiating texts. Scientists present their findings. Diplomats negotiate over wording deep into the night. Opening ceremonies, press conferences, bilateral meetings, debates over the final document — all of them formal elements of a contest that takes place as such, whether the atmosphere keeps heating up or not.
This is not hypocrisy. Nor is it the failure of individuals. Most of those involved work conscientiously, honestly, often with great personal commitment. They do what they can within the framework they have been given. The problem does not lie with those involved. It lies in the framework.
The framework perpetuates itself because too many livelihoods depend on its continued existence. Climate diplomacy has become a profession of its own. Climate NGOs are legitimised by their participation in the process. The media have an annual high point they prepare for. Universities have set up degree courses that prepare students for this world. Foundations finance observer status. A whole industry of climate debate lives off the conferences carrying on. If they stopped, this industry would stop. Nobody wants to let it stop.
That is the mechanism of normalisation. As long as the format is running, the message is: “We are working on it.” As long as we are working on it, the catastrophe is something that is being dealt with, something in progress, something that appears socially manageable. If the format stopped, it would become visible that it never worked. As long as it carries on, the impression that it could work stays alive — and today that impression is worth more than any actual effect.
IV. Germany as a Case Study
In Santa Marta, the Federal Republic provides the textbook example. Federal Environment Minister Carsten Schneider does not travel there in person. He sends State Secretary Jochen Flasbarth, who signs the declaration. At home, a course is being pursued that systematically undermines the declaration. New LNG terminals have gone into operation over the last two years. The coal phase-out, once brought forward to 2030, has de facto been pushed back to 2038 again. The heating law was watered down before it had even taken effect. Industrial policy continues to prioritise fossil-dependent sectors. Germany's climate target pathways are revised in every quarterly report, almost always in the same direction.
This discrepancy is not hypocrisy in the simple sense. It is the structural double track of the Megamachine. On the climate-policy track, the right declarations are produced; on the industrial-policy track, the wrong physical results are produced. The two tracks run in parallel because they have different addressees, follow different institutional logics, apply different criteria of success. The minister can sincerely sign in Santa Marta and sincerely approve the LNG terminals in Berlin. The contradiction is not in his consciousness. It is in the system.
That is the hallmark of a mature Megamachine: it has developed mechanisms by which contradictory statements can be true at the same time, because they are made at different functional levels. Whoever names the contradiction is not refuted — he is referred to the complexity that the contradiction supposedly requires.
V. The Megamachine as Climate Actor
If the conferences do not influence the curve, who does? The question leads to the real observation of this text.
There is no single actor who wants the climate catastrophe. The oil companies do not want it — they want returns, and wanting returns is not the same thing. Politicians do not want it — they want re-election, and re-election means not having to inflict short-term pain on their own population. Consumers do not want it — they want to be able to lead a life in which what everyone else does is available to them too. The climate diplomats least of all want it — they have been fighting against it for decades. Nobody wants it. And yet it comes about.
Lewis Mumford called this phenomenon the Megamachine: the configuration of bureaucracy, technology, economy and military that interacts in such a way that no single actor has an overview of the whole, and all actors together produce an outcome that none of them wanted. In the climate dimension this configuration is especially visible today. Capital is allocated where returns are highest — and in many sectors that is still fossil. Energy infrastructures have lifetimes of thirty to fifty years — what is built today shapes the year 2070. Transport networks are designed for fossil logistics. Agriculture is integrated into fossil-powered supply chains. Building stocks are heated with gas. Militaries run on oil. Financial systems finance extraction.
Each component is optimised for its own logic. The sum produces the climate trajectory we are on. No single will, no single decision, no single culprit. The emergent will of the system.
And now a new layer is added: AI. It optimises oil exploration. It optimises shipping routes. It optimises the financial flows that finance extraction. What we call the Metamachine in another paper extends into the climate dimension: the system gains additional capacity for self-optimisation. The trajectory does not become more flexible because there is more intelligence in the system — it becomes more entrenched, because optimisation now also takes over the points of friction that used to be human.
VI. What the Form Cannot Do
What is the COP format inherently unable to do? It cannot take hard decisions that painfully harm particular interests. Its design — consensus, voluntariness, no sanctions — rules that out. What can stand at its end is always the common ground of what all participants can just about bear — and that common ground, measured against what would be needed, is far too small.
If the format cannot solve the task, the task would have to be shifted to other formats. Possibilities exist, but all of them are politically or structurally difficult. Hard bilateral agreements between major emitters — the USA and China, for instance — could have a substantial effect, but are tied to the respective domestic political constellations, which today hardly permit such agreements. A carbon border adjustment that forces trading partners to comply — the EU has begun with this, but the effect is limited and the format fraught with conflict. Making international finance conditional on climate — the IMF and the World Bank are taking first steps, but against considerable resistance.
Or, harder to admit: shifting the emphasis from mitigation to adaptation. Accepting that the trajectory is what it is, and preparing for what is coming. This is currently avoided because it is regarded as resignation. But adaptation is not resignation — it is the sober answer to a situation in which mitigation, measured against the necessary pace, is no longer taking place.
VII. The Honest Answer Nobody Gives
The honest answer is that nobody knows what would work. The format we have does not work. The alternatives are politically unlikely or not yet tested in practice. Meanwhile the system keeps running. Every year a conference. Every year a new rise in emissions. Every year a widening of the gap between declared mission and measured result.
The role of the conferences is no longer to solve the problem. Their actual role — regardless of their declared role — is to maintain the appearance that the problem is being dealt with, while the system carries on along its course. This is not because anyone wanted it. It is because the form has developed its own logic, and this logic has become decoupled from the function.
The Megamachine perspective does not say what is to be done. It makes visible what is happening. The conferences are not the cause of the catastrophe. They are the ritual through which society processes the catastrophe — makes it manageable, makes it discussable, makes it bearable — while the physical trajectory continues, unimpressed.
That is what the Megamachine does best: producing ritual where action would be needed. Ritual is, paradoxically, more powerful than action. It can be continued indefinitely without producing measurable failure. As long as the conferences go on, the message is: “We are working on it.” On the day they stop, the truth would become visible: we never did.
It is not the task of this text to propose a programme. There are programmes enough. What is perhaps missing is a sober acknowledgement of the situation: that the format in which the international public strives for the climate no longer carries the climate. Whoever acknowledges this does not stop acting. He stops confining action to the format. He begins to look in other places — smaller, more concrete, closer to what he himself can influence. That is not resignation. It is shifting attention to where it has an effect.
The climate conferences will go on. They will go on for many years yet. During that time, whatever real change is possible will not take place in Santa Marta or Belém or Dubai. It will take place in Pakistan, where millions of private PV systems are being installed. In Spain, where wind and sun are changing their share of the electricity mix. In Uruguay, where a small economy has converted its energy supply. In the cooperative energy projects springing up locally in Germany and elsewhere. In every single Tetra-Lodge that uses bamboo instead of concrete. In every honey cooperative that tends the diversity of pollinators instead of industrial agriculture.
Real climate policy does not take place in the conference hall. It takes place everywhere else — and it is slower, smaller, less visible than the annual stage. But it is what remains measurable in the end.