The Chip in Your Phone Was a Cost-Cutting Accident
The Compass Team
October 11, 2026
On April 26, 1985, a small British company switched on the first microprocessor it had ever designed and measured the power draw: one tenth of a watt.
They had been aiming for under one watt, and only because a plastic package was cheaper than a ceramic one. Sophie Wilson later called the result "a complete accident".
The chip was the ARM1. About 25,000 transistors, a tenth the complexity of the chips it competed with. The team behind it: two designers, Sophie Wilson and Steve Furber, plus a handful of engineers at a company called Acorn, in Cambridge.
Acorn was chasing cheap, not speed. To fit the chip in a plastic package instead of a ceramic one, it had to run under one watt. So the team cut every corner it could, choosing simplicity because it couldn't afford the alternative.
Acorn's co-founder Hermann Hauser later said the team had two advantages over the big chipmakers: no money, and no people. Both pushed every decision toward simplicity.
Steve Furber called the approach "Victorian engineering margins". They were so careful to land under a watt that they landed under a tenth of one.
No one had asked for a low-power chip. The whole industry was racing toward faster clocks. Acorn built one that was slower and simpler, because that's what a broke company does.
Then Acorn nearly died, and the chip survived only because Apple needed a processor for a handheld called the Newton. Apple refused to build on a rival's intellectual property. So they cut a deal: Apple put in $3 million, Acorn sent 12 engineers, and a chipmaker called VLSI donated the tools.
In November 1990 the team moved into a converted turkey barn outside Cambridge and named the company Advanced RISC Machines. Apple's one condition: keep a rival's name off the door.
Then the new CEO made the weirder bet. Robin Saxby decided ARM would build no chips at all. It would design the blueprint, license it, and collect a royalty on every chip anyone else made.
That looked like giving away the game. It was the moat.
Any phone maker could license ARM, so they all did. Nobody had to bet the company on a factory. A neutral designer whose customers were rivals was the one shape that could survive any player's failure.
The Newton flopped. The licensing model outlived it.
Nokia put ARM in the 6110, then every phone followed. Today ARM runs 99 percent of the world's smartphones, inside more than 350 billion chips shipped. SoftBank bought the company for $32 billion in 2016, a $40 billion Nvidia takeover collapsed in 2022, and the 2023 IPO valued it at over $54 billion.
The company that built nothing became the most pervasive thing in computing. The future belonged to whoever optimized for the battery instead of the benchmark.
Every phone is a receipt.
Sources
Share this article