The phone dies the month the new one ships. Not all at once, usually; it starts holding less charge, then it lags, then an update it did not ask for lands and the lag gets worse, and by the time the replacement is on the shelves the old one feels like a chore you are being punished for keeping. Somewhere in there the suspicion arrives, and it is hard to shake: that the thing was built to fail on schedule, timed to the release, so that you would have no choice but to buy the next one. It is a suspicion about motive, and it feels obvious. This chapter asks whether it is true, and it turns out to have two answers, because the phrase “planned obsolescence” covers two very different things. One of them is a documented conspiracy with minutes and fines. The other is a trade you made at the shelf without noticing. The work of the chapter is to separate them, and then to hand you a way to tell, for your own dead phone, which one you are looking at.
The case that is fully documented
Start with the hard proof, because it exists and it is worth being exact about. On 23 December 1924, in Geneva, the largest light-bulb manufacturers in the world, among them Osram, Philips, and the international arm of General Electric, signed an agreement and set up a body to run it. The cartel took the name Phoebus. One of the first things it did was to shorten the life of the bulb. By early 1925 the members had agreed a standard life for an ordinary household bulb of 1,000 hours, down from the 1,500 to 2,000 hours that had been common before, and they built a central testing laboratory in Switzerland to enforce it. The enforcement is the part to read closely. Factories sent their bulbs to be checked, and, in the words of the engineer who read the surviving records, if a factory’s bulbs lasted longer or shorter than the regulated life, the factory was obliged to pay a fine. A bulb that lasted too long was a violation. A Tokyo Electric memo from 1927 complains that the penalties for making bulbs too good were discouraging the company from making them at all.
It worked, and the cartel measured that it worked. The average life of a member’s bulb dropped by about a third between 1926 and the 1933–34 fiscal year, from around 1,800 hours to about 1,205, moving down toward the 1,000-hour rule; by the end of that stretch no member factory was making a bulb rated above 1,500 hours. Sales, meanwhile, rose: the cartel sold about 335.7 million bulbs in 1926–27 and about 420.8 million four years later. A shorter-lived product sold in greater numbers, which was the point. This is planned obsolescence in the literal sense the phrase promises: a deliberate agreement to make a thing worse so that people would have to replace it more often, documented in the makers’ own papers. It is not a legend. The legend is what grew up around it, the idea that because this happened to light bulbs in the 1920s, everything you own is rigged the same way. Between the fact and the legend is the whole of this chapter.
Even the courts have been more careful than the legend. When the United States government sued General Electric over its lamp monopoly, the case that produced the judgment reported as 82 F. Supp. 753 in 1949, the Phoebus agreement was in evidence. Yet on ordinary household bulbs the court declined to find that GE had deliberately shortened their life to sell more; the shortening it did find proven concerned particular products, flashlight and Christmas-tree lamps, not the bulb in the ceiling. A court holding the cartel’s own contract still would not extend the charge to every bulb in the house. That restraint is a useful model for reading the modern cases, which arrive dressed as proof and mostly turn out to be something narrower.
What the modern cases actually say
The cases that revived the phrase for a new generation are the smartphone ones, and they are worth reading for what each document actually holds, because the legend treats them as the present’s Phoebus minutes, and the documents say something narrower. The trigger was a real thing Apple did: an update quietly slowed older iPhones. Apple’s account was that ageing batteries could no longer deliver peak current, causing sudden shutdowns, and that the software throttled the processor to prevent them, extending the phone’s usable life rather than ending it. The objection was that this was done without telling anyone, so that a slow phone read as a worn-out phone and pushed the owner toward a new one. Three authorities acted, and the three documents say three different things.
Italy’s competition authority fined Apple 10 million euros and Samsung 5 million in 2018, the statutory maximum in each case, for unfair commercial practice: pushing updates that degraded performance without adequate warning. That is a regulator’s finding of fault, the only finding of fault among the three. France’s consumer-fraud directorate settled with Apple for 25 million euros in 2020, and the detail that matters is which charge survived. The complaint that set the case going, brought by a group whose name translates as Stop Planned Obsolescence, alleged exactly the thing the legend would predict: deliberate obsolescence, which is a crime in France. The regulator did not uphold that charge. What it retained, and what Apple settled, was a misleading commercial practice by omission, the failure to tell buyers the update could slow their phones. The distance between the charge brought and the charge settled is the distance between the legend and the record, drawn in a single case. The United States class action ended in a settlement of between 310 and 500 million dollars, roughly 25 dollars a device, and the approval order states in its own words that Apple denies any wrongdoing. That settlement was itself thrown out by an appeals court in 2022 over the legal standard the lower court had used, then approved again, with payouts beginning in 2024. None of the three is a finding that a phone was designed to fail. Two are settlements, which buy peace rather than admit a plan, and the one regulator that investigated the plan itself declined to find it.
The other kind of short life
If most short-lived products are not Phoebus, what are they? The answer is the second meaning of the phrase, and it is not a conspiracy at all. It is a trade, made at the shelf, usually by the buyer, between how long a thing lasts and what it costs. A cheaper material, a sealed case that makes a phone thin and waterproof but glues the battery in, a part sized to the price point rather than to a fifty-year life: each buys a lower price or a wanted feature with some durability, and each shows up later as a thing that wears out sooner than you wish it had. The maker did not have to plot your return trip to the store. It only had to offer you the version that sold, and the version that sells is usually the cheaper one.
One way to see the trade is to price it, and the light bulb, of all things, is where the modern data is unusually complete, because within living memory the same fitting has taken four technologies. The federal energy agency priced all four as they stood in American homes in 2020, the last year the ordinary incandescent could be counted alongside the rest before efficiency rules ended its sale. Set the up-front price of each against how long it is rated to last and the naive story, that the durable thing costs more, holds only until you look at what an hour of light actually costs.
The inversion in the right-hand panel is the shelf trade, priced. As recently as 2015 the agency’s average price for the LED was $13.53, and a shopper who reached past it for the thirty-cent bulb was making the trade the legend describes, durability passed over on price. By 2020 the LED’s price had collapsed to a few dollars while it lasted twenty times as long and drew a fraction of the current, so the durable option had become the cheap one per hour of light; since then, efficiency rules have retired the ordinary incandescent from American shelves altogether, ending its manufacture in January 2023 and its retail sale that July. While it sold, it sold because it was thirty cents and the buyer in a hurry reached for thirty cents, not because anyone conspired to keep it there. That is the ordinary machinery of obsolescence: a market that keeps offering the cheaper, more fragile version because that is the one enough people buy. No plot is needed to explain it. The demand side has a home of its own, in the way marketing and status turn a working product into an old one in the owner’s eyes, which is taken up where the volume treats how attention and desire are themselves sold; the point here is only that the supply side rarely needs a conspiracy to explain a short life.
The test for your dead phone
Which brings the chapter back to the phone on the desk. The two meanings of planned obsolescence call for two different verdicts, and the reason the folk version feels both true and slippery is that it runs them together. So here is the test, in the form of three questions you can actually answer about a particular thing that died.
First, does the maker profit from the early death in a way that works against you, and did it hide the mechanism? The Phoebus bulb is the model: the death was designed, it served the seller, and the plan was concealed. Undisclosed software that slows a phone leans this way, which is precisely why the regulators acted; a battery that simply wears with use does not, because wear is not a plan. Second, was there a durable version you could have bought and passed over, usually on price? If the shelf held a sturdier model and the cheaper one won, the short life was a trade you made, not a trap that was set. Third, could the thing be repaired, and was that path left open or closed off? A sealed case and a glued battery can be a genuine engineering choice for thinness or waterproofing, and they can also be a way to convert a repairable fault into a reason to rebuy; the tell is whether a repair was possible and priced to lose. Run those three and most dead phones resolve into the ordinary trade, a cheaper build wearing out on schedule with the warranty because that is how it was priced, with the documented exceptions, the concealed and self-serving slowdowns, sitting closer to the cartel. The suspicion that opened the chapter is not paranoid. It is just usually aimed at the wrong one of two very different things.
They design everything to break so you have to buy a new one.
Oversimplified Moderate confidence
The claim rests on something real, which is why the ruling is not that its opposite holds. Some products truly were built to die. A cartel of the world’s light-bulb makers wrote a shorter life into a rule in the 1920s, built a laboratory to test for it, and fined the factories whose bulbs lasted too long, and the record survives in the makers’ own papers. So the reverse of this claim, that nothing is ever built to fail and every short life is an innocent accident, is the less reliable reading of the two, and that is what keeps the ruling off backwards. What the claim gets wrong is the word “everything,” and the motive folded into “so you have to.” Most short lives are not a plot against the buyer. They are durability traded for price at the shelf, a choice the buyer makes, knowingly or not, when the cheaper build wins. The modern light-bulb data shows the trade has even reversed: by 2020 the durable bulb cost less per hour of light than the throwaway one, and the throwaway one has since been retired from sale by efficiency rules, so for that product “they make it flimsy to sell you more” is simply backwards. And the modern cases that look like proof mostly are not. Two of the three headline instruments are settlements, which buy peace and admit nothing, and the one regulator that investigated a planned-obsolescence complaint upheld a narrower charge, failing to disclose a slowdown, rather than a finding that the slowdown was designed to force a sale. The claim takes a real and documented practice, bounded to particular makers and products, and states it as a universal law about everything you own, with a hidden motive attached. It is oversimplified rather than either true as stated or reversed.
Sources
- The documented conspiracy: the Phoebus cartel, founded in Geneva on 23 December 1924, set a 1,000-hour standard for household bulbs in early 1925 (down from 1,500 to 2,000 hours common before), ran a central testing laboratory, and fined factories whose bulbs lasted longer than the rule; measured member-bulb life fell from about 1,800 hours in 1926 to about 1,205 by fiscal 1933–34. M. Krajewski, “The Great Lightbulb Conspiracy,” IEEE Spectrum, 2014, from the OSRAM and Landesarchiv Berlin records.
- A court’s restraint: United States v. General Electric, 82 F. Supp. 753 (1949) — with the Phoebus contract in evidence, the court declined to find deliberate shortening of household-bulb life, the proven design-life findings concerning flashlight and Christmas-tree lamps instead.
- The modern instruments, by document class: Italy’s competition authority fined Apple 10 million and Samsung 5 million euros in 2018 (regulator fine, statutory maxima); France’s fraud directorate settled with Apple for 25 million euros in 2020, retaining a misleading-omission charge and not the planned-obsolescence charge the complaint alleged (settlement); the United States class action settled for 310 to 500 million dollars, the approval order recording that Apple denies any wrongdoing, with the approval vacated on appeal in 2022 and granted again in 2023 (settlement).
- The shelf tradeoff: U.S. Energy Information Administration, Appendix C to “Updated Buildings Sector Appliance and Equipment Costs and Efficiency” (March 2023), 60-watt-equivalent lamps in 2022 dollars, averaged over the 2020 installed stock, the incandescent’s last vintage before efficiency rules ended its retail sale in 2023 — an LED at about $4.56 and 21,000 hours costs roughly 15 times the incandescent to buy and about 22 cents per 1,000 hours of light against the incandescent’s 30, before electricity, so durability is the cheaper choice per hour.
- Confidence is moderate under the rubric, which scores the weaker of evidence directness and construct match. Evidence directness is strong: the cartel record is archival, and the modern instruments are read here as exactly the kind of document each is. The binding, weaker leg is construct match. “Designed to break so you have to buy a new one” asserts a hidden intent spanning every product, while the evidence establishes specific bounded cases and, for the rest, a shelf tradeoff that is a different construct from intent; the motive the claim names is rarely the thing that can be measured. The direction is not in doubt, some things are built to die and most are not, which keeps the ruling firm and off low; the gap between the claim’s universal-intent construct and what the evidence tests is what keeps it off high.
What the test is for
Planned obsolescence is real, bounded in the record to particular makers and products, and usually smaller than the story about it. The version with minutes and fines happened, and could happen again wherever a few makers can agree and hide it. The version most people mean when they curse a dead phone is not that. It is a market doing what markets do, offering the cheaper and more fragile version because that is the one that sells, and letting the buyer trade a longer life away at the shelf for a lower price or a thinner case. The test in this chapter is worth keeping because the difference is not academic: the first case is a wrong to be regulated, and the modern regulators who fined the slowdown were acting on it; the second is a trade to be made with open eyes, and the remedy is a durable option openly priced, not an accusation. Knowing which one you are holding is the difference between a complaint that can be redressed and a purchase that can be made better next time.
The dead phone was the small suspicion, a private grievance about a single object. The next chapter takes the larger one, the money that vanished and the fury that nobody paid for it. If a product built to fail is one way the machine can turn against the person using it, a scheme built to deceive is another, and it raises a harder question than obsolescence does, because the wrongdoing is plain and the punishment somehow is not. The chapter turns from the thing that broke to the person who took your uncle’s savings and walked away.