Look inside almost any modern connected device -- a smartphone, a smartwatch, a Wi-Fi thermostat, a battery-powered sensor node -- and you will find a processor core designed by ARM. It is one of the most successful engineering ideas in computing history. And here is the strange part: ARM has never manufactured a single one of those chips. It does not own a factory. It sells blueprints.
A three-person project in Cambridge
The story starts at Acorn Computers in Cambridge, England, in the early 1980s. Acorn had built the BBC Micro, a hugely popular educational computer in the UK, and it needed a faster processor for its next machine. The commercial chips available at the time were disappointing, so a tiny team decided to design their own.
The acronym everyone knows today originally stood for Acorn RISC Machine. Sophie Wilson designed the instruction set and wrote the simulator; Steve Furber led the physical chip design. RISC -- Reduced Instruction Set Computing -- was the key bet. Instead of piling on complex instructions, they kept the instruction set small and simple, which made the chip easier to build, cheaper, and remarkably power-efficient.
The first silicon, the ARM1, was fabricated by VLSI Technology and delivered to Acorn on 26 April 1985. When the team powered it on, it simply worked -- first try. For anyone who has designed hardware, that is almost unheard of; new processors normally need several rounds of revisions to shake out design errors. A famous piece of Acorn lore is that the early ARM chips drew so little current they could keep running on leakage alone after the power was disconnected.
From a British computer to the whole world
Acorn the computer company faded, but the processor design did not. In 1990 the ARM team was spun out into a separate joint venture, and the acronym was quietly re-expanded to Advanced RISC Machines. The new company made a decision that defined its future: it would not build chips. It would license the designs and let other companies manufacture them.
That licensing model is why ARM ended up everywhere. A chipmaker can take an ARM core, wrap it in its own custom peripherals, radios, and memory, and ship a product without designing a CPU from scratch. Hundreds of companies did exactly that, and the low-power heritage from those Cambridge origins made ARM the natural choice as computing went mobile and then went embedded.
Why it matters for IoT and embedded design
For IoT specifically, that power efficiency is not a nice-to-have -- it is the whole game. A connected sensor deployed in the field might need to run for years on a coin cell or a small battery. The ARM Cortex-M family, the microcontroller-class cores, is built precisely for that: enough compute to run firmware and a network stack, sipping current the rest of the time in low-power sleep modes.
When you pick a microcontroller for a project -- an ESP32, an STM32, a Nordic nRF, a Raspberry Pi Pico -- you are almost certainly picking an ARM core underneath the vendor's branding. The architecture decision made at the silicon layer ripples all the way up through your firmware, your toolchain, and your battery budget. Understanding what ARM is, and why it was designed the way it was, helps you reason about those trade-offs instead of treating the chip as a black box.
The takeaway
ARM proves that owning the factory is not the only path to dominating an industry. A small team in Cambridge optimized for simplicity and low power, then chose to share the design rather than hoard it. Four decades later, that decision put their descendants inside a majority of the world's connected devices.
At Fluidwire we build IoT and embedded systems from silicon to cloud, and nearly every board we prototype has an ARM core at its heart. If you are turning a hardware idea into a working device, get in touch -- we can help you choose the right microcontroller and build the firmware around it.
Top comments (0)