Chips with everything
By Kate Bulkley
Apr 18, 2001
Arm Holdings is in the chip business. They're mad about them. They design them. They license them. But they do not, repeat do not, make them. Which is why, unlike some of their rivals, they're not having to cash them in
As Robin Saxby enters his office at ARM Holdings, he has the look of a man on a mission. You might think his only preoccupation would be to shore up the short-term share price of his microprocessor design company during the downturn in the technology industry. But this is not so, and there are two main reasons. First, ARM Holdings has a stock price of which most tech companies can only dream. Second, Mr Saxby's pre-interview mission was about showing off yet another ARM success story, the Nintendo GameBoy Advance, a handheld computer game-player, for which ARM supplied the microprocessor.
At a time when most technology companies are cutting their earnings forecasts faster than you can say "economic slowdown", it is refreshing to meet a CEO who can consider a future beyond the next quarter's results. ARM has what equity analysts call "momentum". The semi-conductor market is performing as badly as most of the rest of the tech industry, but ARM has a business model that sets it apart. Unlike giants such as Intel, Texas Instruments and Motorola, it does not build chips, but simply innovates, designs and licenses "the chip's engine". The Intels of the tech world are having a hard time selling semiconductor chips because of slowing demand for products ranging from mobile phones to PCs, but the need for improved microprocessing is still high. Only 25 per cent of ARM's income stream comes from royalties on products that include its microprocessors. Most of the rest comes from licensing and developing its designs, and the time lag between design and sale of a product gives the company a smoother earnings stream.
When Mr Saxby, 54, decided to join ARM in 1990 as chief executive, he had lived through six downturns in the highly cyclical chip business. "I said I was too old for more of that," he recalls. "The ARM business model was an evolution and I certainly don't claim to have invented it, but I did say something like, 'We will build chips over my dead body'. Howard Brooks, an equities analyst at brokerage Beeson Gregory, says: "Semiconductor companies don't cut back on R&D in hard times, because with the length of design-to-production cycles they would kill their future business, but [in a down cycle] they do ship less product. So while the semiconductor companies are hurting, ARM's business model powers on." ARM is among the few tech companies which have issued no earnings warnings. In the midst of the tech fallout, it posted first-quarter 2001 revenues up 58 per cent, on target for the twelfth consecutive quarter. And Mr Saxby expects results to be in line with analysts' estimates of 40 to 42 per cent overall growth in revenues for the year. "Our strength is our technology," he says. "We were the first people to really worry about cost and power efficiency [of microprocessors on semiconductor chips]. We thought we had a good strategy and we implemented it.
Even in a down-cycle world you need new technology and the big companies in particular are trying to keep their development budgets together. We describe ourselves as architects not builders. The building industry is struggling, but there is plenty of room for new architecture and new designs." The semiconductor business is highly cyclical because it is based on product sales. A downturn in sales means there are too many chips and not enough products to put them in. Probably the biggest contributor to this is the slump in mobile phone sales to the major system operators, including Vodafone and Orange. The complications of the new-generation phone systems have made forecasting the growth of the mobile phone business very tricky. Factor in the enormous cost of third-generation phone licences, and glitches in next-generation phone technology, and it is little wonder the mobile market is drawing breath.
But even as ARM's near-term royalty payments from mobile sales might suffer a substantial hit, handset manufacturers such as Nokia, spurred on by the debt-burdened mobile operators, want their new phones to be much more powerful. To help the operators meet their revenue goals, the next generation of phones need to provide more services and therefore more revenue to operators. Ramping up the capabilities of these new phones' capabilities is a design issue, which ARM is in a unique position to help solve. "It will be no surprise to us if our royalties from [mobile] handsets will be less [in the coming quarters]," says Mr Saxby. But he believes that although margins will likely be tighter, this can be counter-balanced by making new licensing deals and keeping a close eye on variable costs. "If our partners come out with world-beating phones, then we are in good shape. What I do know is all the best phones are likely to have our technology in them." As it is, ARM has its microprocessor designs in 75 per cent of the world's mobile phones. But only 12.5 per cent of its revenues come from this area, limiting its exposure to the market. On the plus side, the pressure on the mobile phone operators to squeeze out more revenue on new phones will likely mean ARM's expertise in design will be increasingly called on.
One example is the development of Giselle, an ARM microprocessing architecture based on Java computer language. This allows web content to be translated into mobile phones and personal digital assistants (PDAs), which mobile operators such as NTT DoCoMo of Japan believe is key to increasing the revenue potential of third-generation (3G) phones. ARM (Advanced RISC Machines), was founded in November 1990 in Cambridge as a joint venture between the now-defunct Acorn Computer Group, Apple Computers and VLSI Technology.
After an IPO in April 1998 at a price of £5.75 per share, and two stock splits (four-for-one in April 1999 and five-for-one in April 2000), it is now capitalised at £2.6bn. This week, the company's share price is around £3.11. ARM has grown to become the world's leading company supplying intellectual property (IP) to the semiconductor market. Dataquest Research says it had more than 21 per cent of this market as measured by total revenue in 1999, ahead of US competitors such as MIPS Technologies and RAMBUS. The UK company has become the global, de facto standard for the 32-bit embedded chip market with revenues for the year end 2000 up 62 per cent to £100.7m, providing a pre-tax profit of £35.4m that was a 97 per cent increase on 1999. The threat to ARM is, perversely, that it has been so successful, even described as "the Microsoft of the mobile world" by Beeson Gregory's Howard Brooks. Companies such as MIPS are chasing hard and believe the next significant battles will be fought in the 64-bit and 128-bit markets. But Mr Saxby believes ARM's head start will help it head off any threats.
ARM originally created chip architecture for desktop computers. Its genius was to see opportunities in the portable markets that, in the early Nineties, seemed like the distant future. A standard plug-in computer had little need for energy-saving chips because of a constant flow of electricity, but Mr Saxby and his founding team of 12 engineers saw the trend towards portability. ARM realised high-power consumption would handicap powerful portable devices unless the microprocessor could be hugely energy efficient.
One of the other big problems in the early days was to convince large chip-makers to work with a little start-up in Cambridge. But, one by one, convinced by Mr Saxby's renowned sales skills and ARM's power-efficient designs, they signed up, and today ARM's licence agreements with 56 companies read like a Who's Who of the chip business, including Intel, Analog Devices, Samsung, ST Microelectronics, Infineon (formerly part of Siemens), Texas Instruments, Philips, Sharp and Sony among others. Twenty new licences for ARM designs were signed in the last quarter alone. Particularly significant was the signing of Motorola as a licensee in October 2000. Before Mr Saxby joined ARM, he spent 11 years in sales, marketing and engineering at Motorola, but the US chip giant was slow to see the value of the designs from the company of its former employee. "Traditionally, semiconductor companies have designed their own microprocessors, but the cost of doing this is so high now that unless you have established a global standard, it is difficult," says Mr Saxby. "I would loved to have licensed Motorola with ARM [technology] sooner, but they said, 'Why do we need your technology?' Finally, because the end-customers like ARM architecture, they said [to Motorola], 'Please supply us with ARM on your chips'."
The only major name still holding out against the ARM onslaught is Hitachi. In fact, the Japanese electronics conglomerate established a collaboration with ST Microtechnology to spin out a free-standing company called Super SH which has a business model remarkably similar to ARM's. "This should help, not hinder us," says the unflappable Mr Saxby. "ST is already our partner." He sees this new joint venture as another operation that might license ARM's technology, but this may be an overly rosy view. ARM's position must be continually renewed as technologies progress. ARM is inside 75 per cent of mobile phones, but new mobile technologies including GPRS, Bluetooth and DWCDMA mean new designs. So while the pressure in the industry is to create "the next big thing", Super SH, with its backing of Hitachi, an ARM-style business model and ST Microtechnology's expertise, could be the next big threat. Jim Tully, chief analyst of the semiconductor group at Dataquest, says: "Every time there is a discontinuity in a product line because of new technologies like the transition from [today's] GSM phones to GPRS and 3G, there is risk and a potential threat to ARM."
At present, ARM still has a lead and has been clever enough to attract the best engineers, hiring a record 200 people last year, and a further 40 people in the most recent quarter. Today ARM employs 659 staff. "Our challenge is to continue to pick the right partners and the right products," says Mr Saxby. "I don't worry about the competitors. It's not me being arrogant, it's just that we have been doing it for so much longer. We have to make sure all the designs we come out with are the world's best. When people say who are your biggest competitors, I say it's ourselves. It's if we do a bad implementation job, it's lethargy, it's arrogance and it's if we lost the plot." One analyst sees the future a bit differently. "It's been a rapidly growing company, but it is also a rapidly maturing company," says Mr Brooks from Beeson Gregory. "If you become the absolute top player, what is the growth potential?" The growth could be in the faster processors. Mr Saxby says the 64-bit and 128-bit markets are some way down the development track and the biggest growth area is still in 32-bit. ARM wants to make sure power consumption is controlled even when the bit speed is doubled or quadrupled. ARM's chief financial officer, Jonathan Brooks, says: "We are interested in elegant design rather than simply brute-force 64-bit. Mr Saxby says: "The reason we can grow is that in the embedded chip market, which ranges from washing machines to TV remote controls and even chips in your car brakes and air bags, the majority of these are eight-bit. But at the end of last year, the 32-bit market got slightly bigger than the eight-bit market. In older phones the processors were all eight-bit, so now we are in the high-growth sector [32-bit] of a pretty high-growth market. The embedded [microprocessor] market size was about 537 million units last year and we had about 400 million of that." The embedded chip market is expected to reach five billion units in four to five years, a 10-fold growth rate. The trend towards "system on a chip" or system-level integration, means the IP market is growing at a rate well above the overall semiconductor market and is attracting many new vendors.
The present looks bright too, because Mr Saxby expects the billionth embedded ARM processor to be shipped in the second half of this year. Of ARM's 56 licensees, only 23 are shipping products "in significant volume", he says, but "there are a lot of new products coming", all with the prospect of royalty payments to ARM. Mr Saxby enjoys earning from both ends of his deals, the fees for the design and development of a microprocessor and the royalties on the end-products that include these designs. But he admits that growing the first revenue stream from design and design tools and development is the bedrock of ARM's successful business model. "Trying to forecast when [specific] products will take off is stupid," he says. "What we can do is work on leading designs and control the licensing phase and control the design phase. Royalties will come when they are ready to come. Let's look at them as the icing on the cake."
As other tech companies have seen their share prices tumble, ARM's has held up extremely well, perhaps too well. Although its share price is trading below its highs of last year, the shares are on a forward price/earnings multiple in the mid-nineties, which is much higher than its closest competitors. A big fan of ARM, analyst Howard Brooks, has a "sell" on the shares. "I like the company," he says. "I love the business model, but the company is overvalued right now." ARM has a "smoother earnings" path than a classic chip manufacturer, but how can a 90-times-earnings multiple be supported when the company plans to grow at around 40 per cent this year? Bruce Huber, managing director at Broadview International in London, predicts the company must make a major acquisition to keep pace with its stock market rating. The risk of large acquisitions is a potential "watering-down" of the predictable business model. Plus, integrating other companies takes time and costs money.
Mr Saxby says ARM typically does one acquisition per quarter. The CEOsays the lack of hierarchy at ARM helps its employees come up with good ideas. He talks about the convergence of microelectronics and biotechnology, an area he thinks is rich for growth. From more intelligent heart pace-makers to smarter hearing aids, he believes in biotechnology. "This isn't going to have a big impact on short-term revenues, but if you look at DNA mathematically and start to compare it with computers, you can see lots of parallels," he says. "You can predict that as computer power improves, many things could be potentially possible."
This kind of forward thinking won ARM the design and development business behind Nintendo's newest portable games player, the GameBoy Advance. Design work started about five years ago, during which ARM earned substantial fees. Mr Saxby regrets, as he crashes his simulated car while playing with the new GameBoy in his office, that his teenage son is much better at these games than he is. But certainly playing this game has been worth the wait. Nintendo is already shipping 100,000 units a day of the new player in Japan alone, putting it among the world's fastest-selling consumer electronic products, and adding a little more royalty revenue icing to ARM's cake.