The Simple Version

Tech companies build products they know will become obsolete because staying relevant in the market is more valuable than any single product, and obsolescence is how they force that relevance cycle to keep turning in their favor.

Why This Seems Strange but Isn’t

In most industries, building something that lasts is the goal. A bridge engineer doesn’t design for planned failure. A pharmaceutical company doesn’t want its drug to stop working. But technology operates under a different physics, one where the half-life of any given product is essentially predictable from the day it ships.

Apple knows that the iPhone it releases this fall will be outclassed by the one it releases next fall. Microsoft knew that Windows XP would eventually become a liability. Amazon knew that the original Kindle would look embarrassingly primitive within five years. None of this is accidental. None of it is failure.

The confusion comes from applying an industrial-age lens to a post-industrial product. In tech, the product is rarely the point. The relationship is the point.

The Upgrade Cycle as Business Model

When Apple introduced the iPhone in 2007, it wasn’t selling a phone. It was inaugurating a subscription to Apple’s vision of the future, priced as a hardware purchase. Every subsequent iPhone is a renewal notice.

This is why the company invests so heavily in making switching costs real. iMessage lock-in, iCloud photo libraries, the Apple Watch that only works properly with an iPhone, the AirPods that pair seamlessly to Apple devices and awkwardly to everything else. The obsolete product creates the conditions for its own replacement, but only within the same family.

The strategy is self-reinforcing. The more products you own from one company, the more painful it becomes to leave when any single product ages out. Obsolescence, in this reading, is a retention mechanism.

This is related to how per-user pricing grows revenue without asking permission. Both are structures that expand a company’s claim on a customer over time, without requiring any single dramatic ask.

The Internal Logic of Deliberate Shortening

There’s a second, less obvious mechanism: companies that obsolete their own products control the terms of succession. Companies that don’t, get obsoleted by someone else on someone else’s terms.

Intel spent years optimizing for a world that ran on x86 chips in desktop and laptop computers. When mobile computing arrived, the company was structurally and culturally unprepared to compete on power efficiency. Apple, by contrast, made the deliberate decision to move away from Intel chips toward its own ARM-based silicon, accepting the massive cost and disruption of that transition rather than waiting until external pressure made the choice for them. The M1 chip, released in 2020, validated the bet comprehensively.

Amazon did something similar with AWS. The company built internal infrastructure for its own e-commerce operations, then recognized that the infrastructure itself was a product worth selling. It cannibalized its own cost center before anyone else could build a competing one. That willingness to make existing investments obsolete is what gave AWS its years-long head start in cloud computing.

The pattern holds across company size. Tech giants don’t expand randomly; the order they enter markets is the strategy. Deliberate obsolescence is part of that sequencing. You build something good enough to win a market, then replace it before a competitor can use your stagnation as an opening.

The R&D Budget Is a Bet on Future Irrelevance

Here is a useful way to think about research and development spending: it is, functionally, a company paying to make its current products obsolete on its own schedule rather than a competitor’s.

This is why companies like Google fund projects that seem to have no commercial relationship to their core business. Google X (now simply X) has produced self-driving car technology, internet-beaming balloons, and glucose-monitoring contact lenses. Most of these will fail or be spun off or quietly shelved. Some will create new product categories. The point is not the success rate of any individual project. The point is that Google’s core business, search advertising, generates enough cash to fund a portfolio of bets on futures where search advertising may not be the dominant revenue model.

If search itself becomes obsolete, Google wants to have been the company that made it obsolete. The R&D budget is insurance against being the victim rather than the agent of change.

What This Means for the Customers Left Behind

None of this is entirely benign from a consumer’s perspective. Planned obsolescence in software often means that perfectly functional devices get cut off from security updates, not because the hardware can’t handle them, but because supporting old hardware complicates the product roadmap. Microsoft ended support for Windows 10 in October 2025, pushing millions of users toward hardware upgrades on a timeline set by the company’s development preferences, not by any failure of the machines themselves.

The e-waste implications are substantial and mostly unaccounted for in the profit-and-loss statements of the companies driving the upgrade cycles. The social cost is real, and it is being externalized.

But understanding the mechanism doesn’t require endorsing the outcome. The reason tech companies build products they know will become obsolete is that obsolescence, managed carefully, is more profitable than longevity. It creates recurring revenue, reinforces switching costs, lets companies control the terms of their own succession, and funds the next generation of bets.

The product dies. The relationship doesn’t. That asymmetry is the whole business.