Apple Didn't Switch to Its Own Chips for Speed. It Switched for Control.

In June 2020, Apple announced it would stop putting Intel processors in its Macs and start designing the chips itself.¹ After fifteen years of buying the best silicon Intel made,³ it was walking away to build its own. When the first of those chips, the M1, arrived that November, the reviews fixated on one thing: it was astonishingly fast, and astonishingly efficient.² That was true, and it was the least interesting thing about the decision. Apple did not spend a decade and a fortune learning to make processors in order to win a benchmark. It did it to take back something it had given away in 2006 - the right to decide what the most important part of its computer would be.

The official story is that Apple switched to its own chips because they were better. The real story is that Apple switched because it could no longer tolerate the one thing every customer of a chip supplier has to accept: that the most consequential component in your product is designed by someone else, on someone else's schedule, for someone else's priorities. The M1's speed was the proof. Control was the point.

When you buy the brain, you buy someone else's roadmap

Buying a processor from Intel is buying more than a chip. It is accepting Intel's release cadence, Intel's design priorities, and Intel's problems - and Intel, by the late 2010s, had problems. Its manufacturing roadmap stalled for years, and the cost showed up in the numbers: by one independent analysis, over the five years before the M1, Intel's single-thread performance improved about 28% while Apple's mobile chips improved about 198%.⁵ Apple was watching the part it depended on improve at a seventh of the rate of the part it controlled. Worse, Intel's chips were designed to serve every PC maker on earth, which meant they were optimized for no one in particular - certainly not for the specific machine, operating system, and ambitions Apple had in mind. This is the build-versus-buy fork in its sharpest form. Buying is safe, proven, and cheap. Building costs years and billions and might fail outright. But buying means your ceiling is set by a supplier whose interests only partly overlap with yours - and Apple had decided its ambitions for the Mac no longer fit inside Intel's roadmap.

The bet wasn't made in 2020. It was made in 2008.

Here is what makes the decision look brave rather than reckless, and it is the part the 'Apple builds a fast chip' coverage missed entirely. Apple was not a computer assembler that suddenly decided to become a chip company. It had quietly been one for over a decade. It bought the chip designer P.A. Semi in 2008, and in 2010 shipped its first self-designed processor, the A4, in the original iPad.⁴ Every iPhone and iPad since has run on Apple's own silicon, which means that by 2020 the company had spent ten years and a billion units learning to design world-class, low-power chips - on products where a failure would have been survivable. The Mac transition was not a leap into the unknown. It was Apple pointing a decade of accumulated, battle-tested capability at a new target. The fork looked sudden from outside; inside, it was the cashing-in of the longest setup in the industry.

The bet only looks obvious because it worked

It looks obvious now, the way every successful bet does. Two honest objections cut against the easy version. First, the risk was real and large: designing a desktop-class processor is a different problem from a phone chip, the industry is littered with companies that tried to build their own silicon and failed, and a stumble would have meant shipping Macs slower than the Intel ones they replaced. Second, didn't Apple just get lucky that Intel faltered at the same moment? Partly - but it mistakes timing for cause. Apple did not bet the Mac on a chip it hoped it could build; it bet the Mac on a chip-design muscle it had already spent a decade proving on a billion iPhones. The genius was not the courage to leap. It was the patience to make the leap small - to have already done the hard, risky part on products where failure was cheap, so that by the time the Mac's turn came, the outcome was closer to inevitable than brave. Luck favored a company that had spent ten years removing the luck.

For most of its modern life Apple was the best in the world at one thing: assembling other people's finest components into something better than the sum of them. Apple silicon was the moment it stopped being an assembler of the most important part and became its designer. The M1 was fast, and the laptops were great, and the batteries lasted all day - but the durable win was quieter and larger. For the first time in fifteen years, no one outside Cupertino got a vote on what the Mac could become. Apple had bought back its own future, one chip at a time, and the price of admission was the willingness to spend a decade earning the right to pay it.