In 1924, eight of the world's largest light bulb manufacturers met in Geneva and agreed, in writing, to make their product worse.
Bulbs at the time lasted up to 2,500 hours. The cartel standardized them to 1,000. Companies whose bulbs lasted too long were fined. The paperwork still exists. The Phoebus Cartel is not disputed. It is historical record.
This was the first documented case of an industry deliberately degrading its product, coordinating that degradation across competitors, and building enforcement mechanisms to sustain it.
It was also a template.
The Pattern Repeats
The cartel collapsed in 1939 when World War II disrupted European commerce. Bulb lifespans did not recover. The 1,000-hour standard had been normalized into engineering practice. The knowledge of how to build longer-lasting bulbs atrophied because no one was working on the problem.
Alfred Sloan at General Motors understood something similar in the 1920s. He called it "dynamic obsolescence." The idea: redesign cars annually, not to make them better, but to make last year's model feel wrong. Ford's strategy had been durability. The Model T was designed to last. Sloan's strategy was aspiration. Not a better car, but a newer one. By 1931, GM had surpassed Ford in sales. The aspiration model won.
In 2017, Apple pushed software updates to older iPhones that throttled CPU performance. Users noticed their phones felt slow. Many bought new ones. Apple eventually paid $500 million to US class action plaintiffs and $113 million to more than 30 state attorneys general. The company said it would never deliberately shorten the life of its products. The settlements came anyway.
John Deere went further. Modern tractors run on diagnostic software. Farmers who need to repair their equipment cannot read their own error codes without dealer authorization. A $500,000 tractor with a correctable fault sits idle until a dealer arrives. Deere settled a class action for $99 million in 2026. It agreed to give farmers access to diagnostic tools for at least ten years on large agricultural equipment. The agreement was described as a win for farmers. The baseline it was measured against is worth noting.
What Buildings Do
The same logic applies to the built environment, more quietly and over longer time horizons.
The average concrete and steel building in the United States lasts 67 years. Brick buildings average 110 years. Stone, maintained properly, lasts indefinitely. A Roman concrete structure has been standing for 2,000 years. A medieval abbey, 800. A postwar housing block in Britain was often designed to a 60-year standard and required major structural intervention at 25 to 30 years.
This is not a failure of engineering. It is a consequence of how buildings are financed.
A developer who builds a structure, sells it, and exits has no stake in what happens in year 40. The cost of cheap materials, of collapsing maintenance layers into sealed systems, of integrating services that cannot be replaced without replacing the building — those costs land on whoever holds the property decades later. The developer captures the upside. The future absorbs the downside. This is not incompetence. It is rational behavior within the incentive structure.
Stewart Brand described what a healthy building should look like: site (permanent), structure (decades), skin (years), services (years), space plan (months), stuff (daily). Each layer changes at its own rate. Traditional construction respected this. Stone walls outlasted the wooden floors inside them. The floors outlasted the plaster. The plaster outlasted the furnishings. Each element could be replaced without disturbing the others.
Modern curtain wall construction collapses these layers. The structural system, the weatherproofing, and the mechanical services are integrated into a single assembly. Glazing seals fail. They do it reliably, typically within 20 to 30 years. When they go, repair requires replacing the entire facade. When embedded services degrade, they cannot be reached without removing what surrounds them. The building is not designed to be maintained. It is designed to be replaced.
The Supplement Cycle
The same pattern appears in food.
Industrial processing removes nutrients: fiber matrices, fat-soluble vitamins, living enzymes, fermentable substrates for gut bacteria. These are then repackaged and sold as supplements. The US supplement industry generates approximately $50 billion per year. Most of what it sells was once present in whole food before processing extracted it.
Weston A. Price documented this in 1939. He studied traditional diets in isolated populations and found near-zero dental decay, no obesity, and strong skeletal development. The same populations, after adopting industrial food, showed decay and declining health within a single generation. His photographs remain the most direct comparison available.
Remove the completeness. Sell the cure for the resulting deficiency. It is the same business model applied to a different product.
The Counter
Japanese potters have been repairing broken ceramics with gold lacquer for more than 500 years. The practice is kintsugi. The crack becomes visible. The repair becomes the history. The object is worth more for having been broken and healed.
Kintsugi is not merely an aesthetic practice. It encodes a philosophy: that things are worth maintaining, that age and repair are evidence of value rather than failure, that an object worth repairing is an object worth having.
This requires a precondition. The object must be built well enough, and loved enough, to be worth the effort.
Planned obsolescence destroys that precondition. If nothing is designed to last, nothing accumulates the kind of meaning that makes repair worthwhile. The throwaway economy is not just economically wasteful. It destroys the conditions for human attachment to made things.
The EU Right to Repair Directive (Directive 2024/1799, adopted June 13, 2024) establishes that manufacturers must offer repair services beyond the warranty period, make parts and documentation available for up to ten years, and disclose repairability scores. Current scope covers washing machines, dishwashers, phones, tablets, and a handful of other categories. Buildings remain outside scope.
The directive is meaningful progress. The baseline it is measured against is worth naming: this is what it took to make things repairable by law, when repairable-by-design had been the norm for most of human history.
The Deeper Problem
The Phoebus Cartel did not invent the principle. It documented it at industrial scale. The principle is older: design the system so that value flows toward those who control it, regardless of what happens to the product or the person who uses it.
Applied to buildings: shorten lifespan, externalize maintenance costs, extract value at construction, leave the problem for the next owner. Applied to food: remove the nutrition, sell the supplement. Applied to software: throttle the device, drive the upgrade. Applied to agriculture: require seed purchase every year.
The timescale and domain change. The mechanism is the same. Remove the completeness of the thing. Sell access to what was taken out.
The alternative is not a return to some imagined past. It is a different feedback loop. Traditional builders stayed. Craftspeople signed their work. The person who designed a building often answered to someone who would live in it. When the feedback loop is closed, things are built to last. When it is open, they are built to be replaced.
The Phoebus Cartel opened the loop. Everything since has been downstream of that decision.