ARM: Britain’s most successful tech company you’ve never heard of
In a loose collection of offices on an underwhelming business park outside of Cambridge sits Britain’s most successful technology company, ARM. You’ve probably never heard of it, but ARM’s designs are at the heart of the iPhone and nearly every other modern smartphone. It has fingers in almost every other area of technology, from fitness trackers to server farms. It records profit margins that analysts have described as “impossible” (in a good way), and goes a long way to helping justify the “Silicon Fen” label sometimes applied to Cambridge’s tech scene. So how did one company get so successful without anyone really noticing? And, more importantly, what does ARM actually do?
To explain that, it’s easiest to start at the beginning, long before the smartphone was even imagined.
Unlike ARM, Acorn Computers was a household name. Founded in Cambridge in 1978, the company was one of a number of small start-ups latching on to the hot new craze for what were then known as microcomputers, and its first product, the Acorn System 1, was a fairly standard example of the time. Selling for £80, it was aimed at the university market, outfitted with just a small LED display, a keypad and a cassette interface.
The Acorn System 1 was followed by the System 2, 3 and 4, as well as the consumer-focused Acorn Atom, but the company’s real breakthrough came in 1981, when it was chosen by the BBC to produce a machine to accompany a TV series aimed at increasing computer literacy across the UK. Acorn offered up its Proton computer as a candidate, which was rebranded as the BBC Micro and released in December that year. By 1984, around 80% of British schools had one of the computers, with the government subsidising half the cost of its purchase. Acorn looked like it would be a big player in the future of the information age.
So the company set about laying the groundwork for its future. With graphical user interfaces on the horizon, it knew that it would need to boost the speed of its hardware considerably. But for a company the size of Acorn, with just over 400 employees, it would find it hard to design a chip from scratch. Buying one from an external company, as it had with the Micro, wasn’t an option as those processors just weren’t fast enough.
The answer, according Acorn’s then-design manager Stephen Furber, “fell like a bolt from the blue” in a stack of papers dumped on his desk by Acorn co-founder Hermann Hauser. Furber and his co-designer, Sophie Wilson, had found research from the Berkeley campus of the University of California into a new type of processor: one that simplified the set of instructions it would follow, in order to enable a sleeker, more efficient design. This style of processing was called “reduced instruction set computing”, or Risc, and the Berkeley Risc designs had been put together by just two people, David Patterson and Carlo Sequin. Both academics are still at Berkeley, and in February 2015, their contribution to the history of computing was commemorated by the Institute of Electrical and Electronic Engineers with the unveiling of a plaque at the university.
Patterson and Sequin’s invention was crucial to the Acorn researchers. As Furber later wrote: “Here was a processor which had been designed by a few postgraduate students in under a year, yet was competitive with the leading commercial offerings. It was inherently simple, so there were no complex instructions to ruin the interrupt latency.
“It also came with supporting arguments that suggested it could point the way to the future, though technical merit, however well supported by academic argument, is no guarantee of commercial success.” The computer powered by the chip that came out of the design was called the Acorn Risc Machine – or ARM.
For a long time, it looked like Furber’s concerns about commercial success would be borne out. The Risc chips bolstered Acorn’s business, and soon brought it what should have been its biggest coup to date: a partnership with Apple Computers in 1990 to design a new processor for its Newton handheld computer. The new company formed by Acorn and Apple to carry out the design was also called ARM, which now, “following a judicious modification of the acronym expansion”, stood for Advanced Risc Machine.
The Risc designs were perfect for the Newton. The simplified instruction set allowed low-power computing, which was crucial for building what was essentially a smartphone 20 years before its time. But the unusual architecture hurt Acorn itself, which was forced away from the PC market. The “Wintel” monopoly (Microsoft’s Windows operating system and Intel’s processors) was gradually gaining steam, and you couldn’t run Windows on Acorn’s ARM chipped computers. Acorn itself lacked the scale to encourage developers to work directly on its platform. Excelling in technical merit wasn’t translating to a healthy bottom line.
And so Acorn and ARM’s fortunes diverged. Acorn pivoted multiple times over the next decade, building a touchscreen device aimed at replacing newspapers, partnering with Apple to provide computers to UK schools, and designing a set-top-box “thin client”, which would only work paired with a server at the back-end. But by 1998, it was losing almost £10m a year, and eventually the pivots couldn’t save it: in 1999, after a name change to Element 14, the company was bought out entirely by a private equity firm. In the same year, the entirety of its profits had come from its remaining 24% holding in ARM.
Over the same period, ARM grew from strength to strength, even as it faded from the public eye. The company’s expertise in designing low-power processors left it well placed to capitalise on the mobile revolution, firstly through what are now known as “feature” phones and then, as it renewed its partnership with Apple in 2007, with smartphones. The first iPhone contained a Samsung-manufactured ARM-designed chip, as has every one since. And in chip design, as so much else, where Apple led, the industry followed: almost every smartphone on the market today uses an ARM chip.
Where once Acorn stumbled against the Wintel monopoly, now the opposite happens. Windows and Intel are struggling to pivot their technology to the low-power needs of mobile phones while ARM dominates. By 2007, Intel had abandoned the excesses of the company’s Pentium line, which chewed up energy while offering declining marginal returns, and moved to its Core chips, which were low-powered enough to use in laptops, but it still took the company another year to launch its Atom line of mobile processors.
ARM is well aware that it has the technical lead and shows off that fact to whoever it can. Laurence Bryant, the company’s head of strategic marketing, told me that “if you look at the history of ARM, it’s been focused on low power and mobility – that’s our heritage and in our DNA”.
“We are increasing the amount of computing capability and the user experience that goes with it, without sacrificing benefits of power consumption. Low power leads to smaller and lighter batteries, lets you save space used for thermal management, and drop fans. I’m confident that our success will continue in that vein.”
But the company pins its success elsewhere. Rather than focusing on the technical benefits of its chips, it instead highlights the way it works with manufacturers to build them. Ed Gemmell, the company’s head of brand marketing, argues that the company has two competitive advantages: low power processing, and the eco-system around its chips. But both of those are served by a higher way of doing things: “We don’t tell our partners what to make.”
Gemmell says: “We provide the best mobile processors and processing units. How partners want to put those components together is up to them. We want a broad and differentiated partner base. What that does mean is some companies don’t do it as well.
“Intel had too much control over innovation on their PCs. They left very little for the PC manufacturers to do other than slot their chips in. That’s not a healthy environment, and the road to innovation will slow.”
Instead of designing chips outright, ARM takes one step back: it does the highly specialised job of laying out the central processing unit, and then hands those designs on to third parties. Most of those take the design as it is, and fill in the gaps around the edges to make what’s called a “system on a chip”, but some, such as Apple, pay extra for the right to tinker with the design itself.
That way of working has a number of pay-offs that serve to make ARM unique. It’s one of the few companies that arch-competitors trust with their trade secrets. Both Nvidia and Intel use ARM technology, even while competing with each other, and with ARM, in other areas, because they know the company won’t ever fight them in the eyes of end users. You will never buy a chip made by ARM, and that means Intel is happy to work with them. “We couldn’t do our jobs if our partners didn’t trust us,” says Monika Biddulph, who heads up the part of ARM that trains its customers in how to do the hard work of applying the designs.
But the key to its success, it seems, is that it has cracked the goal of every business in the world: it gets paid multiple times for the same work. The chip layouts that are its bread and butter are general-purpose enough that they can be reused again and again in different situations. What is in a smartphone one day ends up controlling a router for high-speed broadband a decade later, and ARM carries on getting paid either way. As a result, in 2014, half the company’s income was coming from designs that were over five years old.
As the company looks to the future, it has big plans. It already makes the majority of its processors for things that aren’t mobile phones – up from just 40% a few years ago – and its chips are now being scaled up to take on Intel at its own game, powering the servers at companies like PayPal. Even more ambitiously, its internet of things department aims to dominate in designing the chips that run the homes of the future.
But of course don’t expect ARM-branded smartfridges or thermostats: in keeping with its history, and its unassuming profile on the outskirts of Cambridge, ARM’s aim is to continue flying below the radar for as long as possible while quietly helping run the electronics we use every day.