It’s 9:30 A.M. on a Friday in Cambridge, England, and ARM CEO Warren East looks annoyed. Bloomberg has just reported that UBS has downgraded his company’s stock to a sell. The news is plausible, since ARM (which used to stand for Advanced RISC Machine) depends on the troubled cellphone market. But it turns out that Bloomberg simply got the wrong ARM. “They confused us with a South African mining company” — African Rainbow Materials — East explains.
This sort of thing happens all the time to ARM ARMH, a microchip designer with $546 million in revenue last year. Nestled in this leafy college town an hour from London — and 5,000 miles from Silicon Valley — it’s easy to overlook. But ARM may be the most important technology company no one has heard of. Its low-power chip designs serve as the brains of 98% of the world’s cellphones. And the company’s technology can be found in myriad other gadgets and devices, including digital TVs, iPods, videogames, video recorders, and even Pleo robotic dinosaurs.
ARM draws basic blueprints; essentially the company designs the core of the chip. Chipmakers such as Qualcomm QCOM, Texas Instruments TXN, Samsung, and many others license those designs and then customize them, sending ARM a small royalty — usually a few cents — for every gadget that rolls off the assembly line. “I don’t think they’ve ever lost sight of who their customers are, what their business is,” says Michael Rayfield, general manager of the mobile business at Nvidia, the graphics-chip maker. “ARM has done a nice job of maintaining leadership in the ultra-low-power [chip design] that is necessary for cellphones.”
As cellphone sales, which represent more than half of ARM’s revenues, have plateaued in recent years, the company’s financial growth has slowed. Revenue grew 78% from 2004 to 2006, for example, but only 6% last year. Profits, at $120 million in 2008, have followed a similar trajectory. (ARM’s profit margins, 21%, are slightly higher than those of its peers.) The flattening growth helps explain why ARM’s shares are trading at only half of what they were two years ago.
But ARM is at the center of a crucial technological shift. The lines between phones and computers are blurring as devices like iPhones and BlackBerrys, which use ARM designs, offer everything from e-mail and web surfing to movies and spreadsheets. These days the most popular services — Facebook, YouTube, iTunes, Google — no longer require being tethered to a PC. With traditional computers beginning to lose their primacy, competition in chips for smartphones and other non-PC devices is intensifying.
For ARM it means facing off against a ferocious new competitor: Intel INTC. The chip titan has declared its ambition to conquer cellphones and other gadgets the way it did personal computers, and steamroll ARM in the process. For its part, ARM is trying to expand by entering a nascent market in which Intel already has an advantage: netbooks, the mini-computers designed mostly for e-mail and web browsing. East thinks his corporate David can prevail over the chip Goliath. As he puts it, “ARM is 1,750 people, but we punch way above our weight.”
An organist in his local church, East, 47, is mild-mannered but capable of the occasional brash proclamation. He acknowledges that Intel is a brilliant competitor and a matchless manufacturer. But, as he puts it, “Intel’s got an old-fashioned business model. ARM’s got a 21st-century business model.”
What he means is that Intel designs and builds all its own chips. It limits its product line to maximize efficiency and protect its profits. And along with Microsoft MSFT, Intel is famous (or infamous) for using its power to essentially dictate how PCs function.
By contrast, ARM licenses its blueprints to all comers and leaves the manufacturing and sales to its partners. Since ARM doesn’t manufacture, it can encourage customers to build whatever they like. That flexibility is especially valuable in the cellphone market, where handset makers often insist on custom chips and software to give them an edge in speed or battery life. “What ARM has done that I think is really good is similar to the strategy we have: It’s to find a lot of partners,” says Qualcomm CEO Paul Jacobs, a customer and an admirer. “It’s not just you as a company doing it all yourself. You leverage what you do by enabling a lot of other people.”
As it happens, Intel and ARM were partners before they were rivals. Intel built an ARM-based chip architecture, XScale, that served as the brain of early BlackBerry and Palm devices. But Intel sold the mobile business in 2006 after it decided that its own X86 architecture could compete in the world of tiny low-power circuits. Last year it introduced XScale’s successor, the Atom. So far the chip has been used in netbooks. A version for phones is at least a year away.
Intel has stumbled in previous attempts to conquer consumer electronics, so it’s far from automatic that the giant will crush ARM. Still, Intel executives argue that the Atom will bring the strength of the PC platform onto phones. “Performance, power, compatibility — you have to deliver all three at the same time,” says Pankaj Kedia, a director in Intel’s mobile Internet devices group. “We are going to do that. We don’t think ARM can.”
ARM traces its beginnings to 1982, the dawn of the PC revolution. While Bill Gates and Steve Jobs were building Microsoft and Apple AAPL in the U.S., European technologists wondered how to get in on the action. A catalyst came in the form of a British Broadcasting Co. documentary, The Mighty Micro, which introduced TV viewers to the age of computing. The BBC later decided to sell an affordable homegrown machine to the enthusiasts it inspired. Acorn Computer in Cambridge won the contract to produce the BBC Micro, Britain’s answer to the Apple II. The Micro was an unexpected hit: Acorn anticipated selling 12,000 but instead moved 1.5 million.
When it came time to plot an encore, a group of young engineers decided to design a processor to serve as the next-generation machine’s digital brain. The chip would be based on “reduced instruction set computing” or RISC, an approach that stressed simplicity and efficiency. They named it the Acorn RISC Machine (later changed to Advanced RISC Machine): ARM.
Meanwhile, the rest of Acorn was going to seed. Operations were a mess and parts shortages plagued the company. Customers began canceling orders. By the end of the decade Acorn had nearly tumbled into bankruptcy more than once.
A knight in shining armor arrived, however. Apple, Acorn’s competitor across the pond, was searching for a chip to power the Newton handheld computer, a forerunner of the PalmPilot that it was then planning. Acorn needed cash, and Apple needed the chip. So in 1990, Acorn spun off ARM as a standalone company, and Apple paid $3 million for a 30% stake (which it gradually unloaded over the years).
The new venture was led by a swashbuckling former Motorola executive named Robin Saxby. Equal parts dreamer and realist, Saxby envisioned ARM’s then-radical business model, figuring ARM could gain market share by liberally licensing its ideas to chip companies and collecting royalties when they succeeded. He knew the strategy could work only if he ran a tight ship. So he moved ARM into a converted 18th-century barn and devised clever ways to save money, such as having employees wire and install the telephone system themselves.
During the early ’90s, plenty of tech luminaries dismissed ARM’s approach of focusing on tiny low-power chips instead of the more muscular designs favored by larger competitors. Jim Clark, then chairman of rival Silicon Graphics, called ARM’s technology a “toy.” But Saxby’s penny-pinching ethic and gambler’s nerve set ARM up for its big break in the mid-’90s. After years of disappointment when promising licensee products like the Newton failed in the marketplace, ARM won a spot in a breakthrough new digital cellphone from Nokia NOK. The Finnish company became the dominant manufacturer of handsets, and phonemakers beat a path to ARM’s door. The year before the Nokia phone arrived, ARM had $45 million in revenues; the year after, it was $99 million.
Today cellphones aren’t the growth business they once were, which means ARM faces challenges even apart from Intel.
ARM’s occasional stumbles have added to its headaches. Most recently the company took a timid approach to mobile graphics chips. Rather than going it alone, it partnered with an outfit called Imagination Technologies on the iPhone 3G, only to see Imagination aggressively improve its own offering and waltz off with the sole contract for the new iPhone 3GS. (ARM’s technology is still used in chips that control other functions on the iPhone.)
Smartphones are lucrative for ARM. The company earns more than six times the royalty from a smartphone vs. a plain-vanilla cellphone, because the tiny computers use more of its technology. However, smartphone sales haven’t amounted to enough so far to really drive ARM’s revenues.
And outside the smartphone category, there simply aren’t many billion-unit chip opportunities around. East has talked about a new breed of “smart” home appliances as a potential market — think networked thermostats and washer-dryers that turn on when electricity costs are lowest — but that vision is a long way from materializing. “There’s no Nokia,” East acknowledges, no single company that can propel ARM in this new market in the way Nokia did with cellphones 15 years ago.
In the meantime, Intel seems to be girding for battle. At a conference last year, Kedia, one of the company’s mobile Internet device directors, took a swipe at ARM. “The shortcomings of the iPhone,” he charged, “have come from ARM.” He claimed that ARM’s architecture made the iPhone slow and implied that Intel’s processors would do better. After ZDNET reported the comments, Intel’s mobile czar issued an apology, admitting that Intel’s Atom chips don’t yet operate on low enough power to work in a phone.
The kerfuffle exposed a hole in Intel’s argument that its dominance of PCs will carry over to mobile. Apple already uses Intel chips in all its computers and should be a prime customer for Atom. But instead it seems to be pouring resources into an ARM-based strategy. Apple CEO Steve Jobs has said the company is designing its own system-on-a-chip — a single integrated circuit that contains all the intelligence of a computer — specifically for iPods and iPhones.
There may be even more to Apple’s chip design ambitions. Richard Doherty, principal at the Envisioneering Group consultancy, believes Apple will use ARM-based chips in a new class of computer, perhaps a tablet, within a year. That would give the whole ARM community a boost. “The ecosystem is what counts here,” Doherty says. “If Apple comes out with something amazing, I’m sure that other ARM licensees are going to be able to draft along with them.”
Others are already flirting with the idea of ARM-based PCs. Asus, Elitegroup, and Pegatron have showed off laptops with ARM-based chips. Phil McKinney, chief technology officer for Hewlett-Packard’s PC group, is mum on rumors that HP HPQ is eyeing ARM for laptops. But he tells Fortune that such chips could make sense for buyers who don’t want a traditional PC: “I think you’re going to see some interesting devices coming out from a variety of players that are going to be ARM-based.”
For its part, Intel announced a strategic alliance with — of all companies — Nokia in late June. At first blush, it appeared to be a victory for Intel. The two companies will collaborate on a mobile computing device. However, not only did Nokia not commit to using Intel’s Atom chip, but a Nokia executive went so far as to note, on the conference call announcing the pact, that the new deal won’t affect
Nokia’s “great” relationships with ARM licensees.
ARM, indeed, has great relationships with customers. The company may even benefit from the fact that it isn’t Intel. More than one tech exec would prefer not to see Intel extend a crushing near-monopoly into a new arena. ARM’s CEO remains sanguine. “We’ve got the winning technology because we came from low power in the first place, and that’s what the market wants right now,” says East. “I’m not going to sit here and tell you that Intel is going to win, am I?” Certainly not. But if you have to race against the likes of Intel, it’s nice to know you have a headstart.