Intel's Revenue Barely Moved. Everything That Mattered Collapsed Anyway.

Read Intel's 2024 top line and you'd never know anything was wrong. Revenue came in at $53.1 billion, down a rounding error — about 2% — from the year before.³ For a company being eulogized in every business section on earth, that number is almost insolent in its calm. And it is the most misleading figure in the whole autopsy. Because in the same year that revenue barely flinched, Intel's contract-manufacturing arm lost $13 billion, the company suspended a dividend it had paid for decades, more than 15% of the workforce was marked for the door, and the board pushed out the CEO it had hired three years earlier to save the place.²⁴⁶ The patient's pulse looked fine. Every organ underneath had failed.

The official story is that Intel had a rough patch, a bold turnaround, and a CEO who retired. Almost every load-bearing word of that is doing PR work. Intel did not have a patch — it had a decade-long structural unwind. The turnaround was the right diagnosis at a dose that nearly killed the patient. And the CEO did not retire.

The press release said "retired."⁵ Then the reporting arrived. Bloomberg, citing people familiar with the matter, found the board had lost confidence in Gelsinger's turnaround plan and gave him a choice: retire, or be removed. He chose the gentler word.⁶ That gap — between the verb Intel published and the verb that actually happened — is the whole story in miniature. Intel's collapse is a long string of euphemisms finally catching up with the arithmetic.

The collapse hid in the segments, not the total

Here is the trick a sticky revenue line plays on you. A company can hold its top line flat while the economics underneath it invert completely — if it is propping up a money-losing new bet with the cash from an old, declining one. That is precisely what Intel was doing. In April 2024 it broke its foundry business out as its own reporting segment for the first time, and the curtain came up on a $6.96 billion operating loss for 2023, on $18.9 billion of foundry revenue.¹ That was the warning. Then 2024's number landed: a $13 billion foundry operating loss — nearly double the prior year — even as group revenue held at $53 billion.²³ The business wasn't shrinking. It was bleeding internally while standing perfectly still.

The mobile miss was never one bad meeting

Every Intel autopsy eventually reaches for the same tidy origin story: the company turned down the chip for the first iPhone over a single pricing call, and the smartphone era left it behind. It is a wonderful story and it is mythology. The damage was not one rejected deal — it was a multi-decade pattern of decisions that each made local sense and together cost Intel an entire computing platform. The most telling of them: Intel had its own competitive ARM-based mobile processor line and sold it off in the mid-2000s, walking away from the exact architecture that would go on to power nearly every phone on the planet.⁸ You cannot miss a market in one meeting. You miss it the way Intel did — by repeatedly choosing the margin you could see over the platform you couldn't yet.

This matters for the autopsy because it relocates the cause of death. The romantic version puts the fatal error in one room on one afternoon, which is comforting — it implies a single villain and an easy fix. The structural version is harsher: Intel's organism was built to defend the x86 franchise, and an organism optimized for one thing will keep declining the next thing, again and again, each time with a defensible spreadsheet in hand. The mobile miss wasn't a stumble. It was the system working exactly as designed, against a future the design couldn't price.

The right diagnosis at a lethal dose

Gelsinger's bet — to turn Intel into a contract chip manufacturer competing for the world's most advanced silicon — was not the dumb part of this story. It may have been the correct read of where the industry's power had migrated. The problem was the dose. Standing up a leading-edge foundry is one of the most capital-hungry projects in industry, and Intel tried to fund it while its legacy business was softening, not booming. By mid-2024 the strain forced a $10 billion cost-reduction plan, headcount cuts north of 15%, and — most telling of all — the suspension of the dividend, the single clearest signal a mature company can send that it has run out of room.⁴ Even the rescue from Washington shrank in transit: a CHIPS Act package floated in March as up to $8.5 billion was finalized in November at $7.86 billion, materially less, against a hole measured in tens of billions.⁷ Right medicine, wrong patient, far too late in the disease.

Isn't this just a turnaround that needs more time?

The fair objection is that none of this is fatal yet. Revenue held. The foundry has a stated path to break-even by the end of 2027.² Federal money, even at $7.86 billion, is real, and the fabs being funded — in Arizona, New Mexico, Ohio, and Oregon — are physical assets a competitor cannot conjure overnight.⁷ On this reading, 2024 was the trough of a deliberate, painful, multi-year investment, and calling it a collapse is just impatience. That argument is honest and not absurd. But it misses what the board's own action proved. The people with the most information, the most patience, and the most to lose by acting early did not extend the runway — they removed the pilot, while publishing the word "retired."⁵⁶ A board that believed in the timeline does not force out the author of the timeline two months after finalizing its government funding. The clearest evidence the turnaround thesis had failed is that the people running it stopped believing in the person running it.

Intel did not collapse the way companies are supposed to — sales falling off a cliff, customers fleeing, the number everyone watches turning red. Its number stayed almost serene: $53 billion in, $53 billion out, two years running.³ The collapse happened everywhere the headline couldn't reach — in a foundry loss that doubled, in a dividend that vanished, in a workforce cut by a sixth, in a CEO eased out under a word the board didn't believe. The lesson of Intel isn't that a great company can fail. It's that a great company can fail for a decade while its revenue line lies to you about it — and that the most expensive misreadings are the ones the top line lets you keep believing.