Report on Business Magazine
FUTURETECH
CHIPS and GRAVY ATI Technologies is at the top of the graphics-chip business. The trick is to stay there.
Clive Thompson
05/28/1999
The Globe and Mail
Metro
Page 96
All material copyright Thomson Canada Limited or its licensors. All rights reserved.
The kid pauses beside the bridge over a slimy canal, looking briefly up a massive, ruined, postapocalyptic castle. He reloads an enormous rocket-launcher gun. "See, here," he says, pointing at the computer screen. "That castle, see how realistic it looks? And look at the water in the moat, look at the movement. It's just eerie how real that is."
Standing in a Toronto computer store, the 14-year-old boy--Joseph, though he won't give his last name, since he's supposed to be in school--is contemplating the peculiar importance of 3-D computer graphics in today's simulated life. He's playing Quake II, the infamously gory first-person computer game, which the store has loaded onto a demo IBM Aptiva. The graphics on the computer are particularly good because, as the kid explains, the computer comes equipped with a hot graphics chip--the "Rage Pro Turbo," made by ATI Technologies Inc., an equally hot Canadian high-tech company. Thus, when four guards--the alien enemies in the game--come around the corner, the resulting bloodshed is rendered in exquisitely fine detail.
"If you don't have a good [graphics] chip," Joseph says, as he calmly strafes the room with rockets, "you don't get this smooth movement. My old computer's got no chip, so everything's got these rough edges, it's crap. Now look at that," he says triumphantly, as a rocket hits a guard and explodes. An orange cloud of unsettlingly fractal beauty erupts, as chunks of flesh fly in graceful, smooth arcs. "That's what you're lookin' for!"
Indeed, there's quite a bit of emotion wrapped up in the graphic-chip industry. By making it possible for computers to generate colourful pictures, 3-D animations and video, graphics chips in the past 10 years have transformed computers from plain-text, egghead tools to full-fledged entertainment and creative devices. And as computer games, the multimedia Internet and digital TV become ever more mainstream, graphics chips have turned into a very big business indeed: about $1.8 billion in the past year.
It is a business that has recently become dominated, unexpectedly, by ATI. "Unexpectedly" because nobody is accustomed to thinking of Canada when they hear the word "microchips"; that's usually the terrain of Silicon Valley and intergalactic firms such as Intel. Yet last year, the strategy of the Markham, Ont.-based ATI suddenly clicked. Years of pouring money into research and development, aggressive marketing and following the zigzagging graphics market paid off. Its share of the world graphics-chip market jumped abruptly to 32%, more than doubling that of its nearest rival, Santa Clara, Calif.'s S3 Inc.
Breaking the psychologically suggestive $1-billion mark last year on its sales of graphics boards (known to consumers as graphics cards) and chips, ATI, with 1,700 employees, is now Canada's third-largest high-tech company and has seen its stock price rise as high as $27, from its 1993 issuing price of $4. With $74 million in the bank at the start of this year, it had no debt. And all the while, industry analysts have gone faintly berserk with positive reviews. "They are just on top of the pile," raves Jon Peddie, president of Jon Peddie Associates, a California graphics-chip industry research group. "They have a strong brand, the broadest product line, they're accepted by both [computer makers] and consumers. They haven't missed a step. It's remarkable."
High praise--but inside the ATI bunker, they're not sitting on their laurels. They know they're riding a crest; they also know the profoundly fickle nature of the graphics-chip industry. With devastatingly slim margins, demanding computer-manufacturer customers, and products that are nearing obsolescence after nine months, the terrain is littered with failed companies that, too, once held the brass ring.
More problematically, future trends remain obscure. There's a plethora of areas the graphics industry is advancing into--but which one will prove profitable? Digital TV? Interactive cable services? Laptops? Quasi-computer "appliances?" Take your pick. Last year, the company's 48-year-old president and co-founder, Kwok Yuen ("K.Y.") Ho, won Ernst & Young's Canadian Entrepreneur of the Year Award, and in January he was named one of North America's top 25 executives by BusinessWeek. But despite his endlessly sunny demeanour, he preserves a black sense of humour about the cutthroat dynamics of the business. "It's like driving a sports car," he jokes. "You can either arrive early, or it can kill you in an accident. Either way, at least you have the excitement."
Computer critic Sherry Turkle once famously argued that, as computers have become more powerful, they've moved from calculation to simulation. Calculation doesn't take much computer speed. Even a relatively slow computer--say, an old 386 from 1991--can calculate spreadsheets or crunch databases well enough for the average businessperson's needs. The push for faster and faster chips has come largely from the desire for simulation, such as pictures and animation.
Displaying even a small bit of animation or a complex Web site will eat up a disproportionately large chunk of your computer's main processing chip (such as Pentium II), slowing its performance--an annoying phenomenon that every computer user has experienced. Hence the market for the graphics chip. Working alongside the computer's main processor, it concentrates solely on rendering graphics, neatly removing this load from the main processor. Some times the chip comes as part of a computer, pre-installed by the manufacturer; some times customers buy an ATI graphics card off the shelf and install it on their computer themselves. Either way, the end result is smoother images, with richer 3-D effects, moving at a much faster pace. And, as Ho notes, this isn't restricted to video gamers: "Everything is graphical now--even the spreadsheets, the commercial applications."
The need for a graphics chip seems obvious now, but it was a harder sell back in 1985, when Ho and two partners founded ATI in east Toronto. Ho was a recent immigrant from Hong Kong; trained as an electrical engineer, he'd worked at home for Control Data Systems Inc., Philips Electronics and National Semiconductor. In Canada, though, he couldn't find work; as an immigrant, he supposedly didn't have enough "relevant experience." Nonetheless, with the recent release of the graphically driven Macintosh, Ho could see which way the computer industry was headed. Risking their savings, he and two friends, Lee Lau and Benny Lau (not related), put together $300,000 to found ATI. With a staff of six, it was considerably leaner than most people would imagine a chip company. "The secretary, the receptionist and the shipping department were all one person," Ho laughs. Since they didn't have enough money to build a plant to assemble the chips--a multimillion-dollar proposition--they outsourced assembly to a Californian chip factory.
The initial months were a fight against obscurity. They designed a graphics chip, but as an unproven, small Canadian company, Ho recalls, "computer manufacturers would say, 'Why should we deal with you?'" Within four months, they spent their initial stake. The Overseas Union Bank of Singapore came to the rescue with a $300,000 loan, which was soon increased to $1.5 million.
ATI did have one thing going for it, though: the high-volume nature of the chip business. Once a computer maker decides it likes a graphics chip, orders of hundreds of thousands of chips are common. In ATI's case, sales picked up dramatically in the second half of its first year, when Commodore placed an order for 7,000 chips a week. By the end of its first year, ATI had done $10 million in business. "That was the first real lifesaver," Ho says.
Much of its sales success was based on technical innovation, paired with low-priced products. Faced with a deeply fragmented computer market in the '80s, for example, ATI released The Graphics Solution--a popular board that combined several different graphics functions, making it a one-size-fits-all product. On top of these technical smarts, ATI quickly developed an ability to woo and maintain large client companies (now including Hewlett-Packard, Compaq and IBM), which were scrambling to make their computers highly graphical. Ho's salesmanship is renowned; even today, he estimates he spends more than 80% of his time massaging clients.
Peddie remembers meeting Ho at a computer conference in California in the late '80s, when ATI was still a relative unknown. With his ever-present toothy grin and deeply inflected English, Ho didn't necessarily cut a typical figure for an executive. Once Peddie talked to him, though, he saw Ho's grasp of the coming importance of graphics chips. "You realized right away that he was killer smart," Peddie says. "He's like Bill Gates. If you look at the guys who are successful in this business, they're not the typical MBA types. They're the nerds. They understand the techie stuff. But they also know how to sell."
By the early 1990s, ATI was well established, with a solid stable of clients and a handful of products. But as the company was discovering, getting onto the playing field and staying there are two very different things. Though high-tech industries are known for quick turnaround, the graphics-chip industry has raised it to a painful art form. "The prudent investor," says one analyst, "could reasonably conclude that this is a brutal industry."
Part of the difficulty is the narrow profit margin on graphics chips. With eight million transistors per chip, today's graphics chips are nearly as complex to build as major Pentium processors. Yet they retail for less than one-tenth the price-- $30 versus the $300 to $1,000 for a Pentium chip. Partly, the prices are lower because the testing standards, which rack up development expenses, aren't as rigorous as for main processor chips. But the low prices are also a tough historical fact; graphics chips were originally considered a "peripheral," and the industry had to price itself accordingly.
To make matters worse, even a successful graphics chip has a perilously short shelf life. According to "Moore's Law," the number of transistors that can be jammed onto a main processing chip doubles every 18 to 24 months, thus doubling processing power. Graphics-chip power, in comparison, doubles at twice that breakneck rate. In the spring of 1998, ATI's main graphics chip offered 8 megs of memory; one year later, it was up to 32 megs, for the same price.
As a result, computer makers are constantly scouting for the latest cutting-edge chip, making the air of competition amongst graphics-chip companies positively neurotic. There are about five major players, and 15 minor ones. If a company misses a product cycle--such as the computer-makers' summer buy in preparation for the student market--they can be quickly shut out, and wind up with millions of instantly obsolete chips sitting in a warehouse. "It's a very, very volatile business. Success or failure is measured in quarters," says Dean McCarron of Mercury Research, a major Arizona-based graphics-chip-industry research firm.
Such was the recent fate of S3, ATI's biggest competitor. In the fourth quarter of 1997, the company enjoyed a market share of 35%, while ATI trailed a distant second, with 18%. But S3 had already made crucial mistakes, including working on an ill-advised chip architecture that ignored the trend toward high-end 3-D games. One year later, S3's market share was down to a mere 14%. "We had all of our resources on one chip, and ultimately it wasn't where the market led," says Andrew Wolfe, an S3 vice-president.
Meanwhile, ATI's market share rocketed to 32% in 1998, neatly reversing positions with S3. Revenue and profit soared by 92% and 350%, respectively. What had gone so right?
First, by working closely with Intel and computer makers, Ho had neatly managed the transition to high-end 3-D graphics. In addition, Ho's salesmanship had brought in high-end clients that gave ATI name-brand visibility. At the 1999 MacWorld expo, when Apple rode a wave of euphoria over the iMac's exploding sales, Steve Jobs announced on stage that Apple's new G3s would use top-line ATI chips exclusively. "We're going to recapture the gamers' market, and that's why we want ATI," says Craig Michaels, Apple's product marketing manager for the iMac. For ATI, the public association with Apple was a masterful bit of PR.
ATI's relationship with Apple also illustrates its savvy in managing its older products. For the iMac, ATI sold Apple on ATI's cheaper, older-generation chips. By blending medium-level performance with a bargain price, they satisfied the iMac's need for a balance between power and affordability, Michaels says. This strategy--unloading slightly older chips at discount prices, while still pushing high-end wares--is one that all chip-makers strive for, but which ATI has mastered, analysts say. It has allowed ATI to get rid of old inventory, and make important inroads into the lucrative and growing sub-$1,000 PC business.
At the same time, Ho has kept his finger on two other growing trends in computer usage--first, the increased desire to view and manipulate video, including DVD movies. While most graphics-chip companies have focused on gaming and Internet graphics, ATI has also developed chips that allow computers to manipulate video images, such as from a TV or a camcorder. Bal Sahjpaul, ATI's manager of Internet strategies, has found that customers are using ATI's All-in-Wonder card for everyday purposes, such as conducting video-phone conferences or editing videotapes on their computers: "It's become a real creativity device for people." Second, with an eye on the lucrative and exploding laptop market, Ho pushed ATI to produce cutting-edge chips tailored for portasble LCD screens, used in laptop computers. Though neither of these two areas comes close to being the size of the desktop market, both are growing quickly. A recent report by Credit Suisse First Boston found that ATI's share of the laptop market went from zero in 1997 to 8% in 1998, and is still growing.
Still, though, there is a broad streak of luck running through ATI's triumphs. ATI's executives are self-aware enough to admit it. They realize the grim rule of the chips industry: For you to win, someone else must lose. If S3 hadn't stumbled so badly, it's doubtful that ATI would be where it is today. "All this bunch of things happened right all at once and we went 'boom,'" says Phil Eisler, ATI's manager of component marketing.
Yet ATI is no stranger to bad luck itself. In 1994, it was hit by dual delays in getting out a new chip. First, testing was prolonged to eliminate bugs. Then supply problems developed in getting crucial circuits made. Third-quarter sales came in at half the company's projections; stung, investors knocked 30% off the stock price, bringing it at one point to below the $5 mark. Ho bought 19,100 shares as a vote of confidence in the company. More concretely, he began moving more of ATI's production in-house, to reduce its reliance on outsourced chip-fabrication services. "Whenever the market becomes tough, we need to get more creative," Ho says. "In this business, you cannot keep yourself lazy."
Despite ATI's successes, there are still doomsayers. Mostly, they point to the volatility of the graphics-chip market. If ATI gambles R&D money on a product that doesn't sell, or misses an important change in technology, they argue, the game will be up.
Last fall, the Boston, Mass.-based analyst firm Off Wall Street issued a scathing "sell" report on ATI. It predicted that Intel, which had entered the graphics-chip market with its own offering--the i740 chip--would eventually unseat ATI in the desktop market. Though Intel gained only a sliver of the market, Off Wall Street's Mark Roberts still feels ATI is overvalued. ATI's stream of "buy" recommendations, he argues, comes from the fact that the firm is a rarity in the Canadian investor landscape. "It's a cult stock in Canada. Investors don't have many chip stocks, so they pump ATI up." More prosaically, he points out that by mere virtue of being on top, ATI has nowhere to go but down. Weak logic, maybe, but even ATI's biggest fans can't argue with that sort of tautology. "They're positioned well, but they're the biggest, so they've got a big red target painted on their butt," Peddie admits.
Back at ATI's head office, Andrew Rodie fiddles with a black box. It's a prototype of ATI's first "set-top box"--a cable converter tricked out with the company's cutting-edge graphics and video chips. ATI is building the device, called the "Set-Top-Wonder," for General Instrument, which will retail it to cable companies looking to offer interactive, broadband services. (In Canada, Shaw Communications will offer it to its 1.5 million cable subscribers later this year.)
Once he gets the box working, Rodie, head of ATI's demo team, takes it through its paces. Standing on its own, the box can play computer games and DVD disks. Hooked up to a digital cable line, the box can receive high-resolution digital TV images. With a keyboard attached, it can surf the Web, send E-mail or conduct
E-commerce. And via the superfat data signal of cable, lightning-quick networked gaming will be a breeze. "In fact, this thing has more memory and power than many computers on the market right now, and it'll have the fastest Net connection available. It's like the Holy Grail," Rodie muses, as his seemingly teenaged assistant shoots aliens on-screen in a computer game.
For ATI, this box is the hope for the future. The fight over the desktop computer may have won it coveted market share, but as Ho argues, the trend is now toward powerful consumer electronics--"home info-appliances." Equipped with robust, high-end chips, but encased in stereo-like components with push-button ease, information appliances will demystify computing. "People are scared of computers," Ho says. "But they don't worry about using computing devices, like a microwave or a VCR. A microwave is a computer! We just don't think of it that way. That's why we have to make our products less like computers."
Come the fall, when Shaw starts offering set-top boxes to customers, Ho will have a chance to find out if he's right. As always, the margins here are slender: the Set-Top-Wonder will retail for only $300 to $400, putting it into the same league as Nintendo 64s and VCRs, and substantially undercutting the price of current DVD players. Analysts are cheering on ATI's foresight. As Mercury's McCarron notes, ATI's combination of expertise in graphics and video is unusual--its competitors are often good in one or the other, but not both. "They're positioned very well for that market," he says. "There's going to be a sea change in the products that people like ATI make, sooner or later."
As the day draws to a close, Ho could tell you about other plans he has, if you had time. He's planning to take on Intel directly in the race to make a "system on a chip," aimed at creating a new generation of super-cheap, sub-$300 computers. And there are plans for video products, for markets he can't even yet disclose. God knows he'll need them. "Our business is like rowing a boat up a stream," he laughs. "We're working, working, working, just to stay in place."
|
