Ali,
Some Intel background from todays mercury news.
Pretty interesting reading:
Posted at 4:31 p.m. PST Saturday, February 20, 1999
Intel sets standards in PC market
BY TOM QUINLAN
Mercury News Staff Writer
After years on the margins of the computer industry, Packard Bell Electronics Inc. made a bold bid in 1993 to join the top ranks. It handed control of its technology to Intel Corp. By buying into the chip company's Pentium initiative, Packard Bell took from Intel everything it needed to deliver state-of-the-art personal computers -- not just the chips Intel typically furnished, but critical surrounding components as well. Within two years, the company nearly doubled its market share, trailing only PC pioneer Compaq Computer Corp. in sales. The real achievement, however, was Intel's: The Santa Clara company had rewritten the rules for success in the personal computer industry. Suddenly, a component supplier had more power to move the market than the PC companies themselves. It's as if the firm that supplied the engine block to General Motors suddenly came to dominate the auto industry -- controlling the kinds of spark plugs, oil filters, carburetors and batteries to be used in the finished car, and reaping the bulk of the profit from each auto sale. Intel was not only Inside, it was In Charge. But today, Intel's empire is under siege. The Federal Trade Commission is seeking to curtail the company's power, even as the Justice Department is targeting the other half of the Wintel duo, Microsoft Corp. And the emergence of popular new digital devices threatens to undermine Intel's command of the microprocessor market. Still, few outside the PC business understand the scope of Intel's power -- or how the company uses it. Under six years of Intel's guidance, the personal computer has undergone a dramatic transformation from a balky, esoteric device to a mass-market product that many people find indispensable. And the computer industry has become an economic dynamo, generating hundreds of billions in sales each year -- and, not incidentally, much of Silicon Valley's affluence. No hardware company is more responsible for these developments than Intel. None has benefited nearly as much as this $26 billion corporation, which has become one of the world's most impressive profit machines. None has Intel's influence. ''If not us, who?'' asked Intel Chairman Andy Grove, referring to Intel's leading place in its industry. ''We are a major beneficiary and participant in the PC industry. If we don't do our fair share of the heavy lifting and moving the industry forward, who's going to do it?'' Intel is more than simply the leading PC microprocessor maker, with nearly 79 percent of the market in 1998, according to the IDC research firm. The company has a nearly unbreakable hold on almost every aspect of PC performance, from memory to graphics to sound. So potent is Intel that it can determine whether technical innovations succeed or fail, whether certain companies in its industry prosper or struggle, and -- on some occasions -- whether products make it to market or die on the drawing boards. Consumers can thank Intel for the performance and increased simplicity of today's desktop computers, which have improved without an increase in price. And thousands of companies with millions of workers can thank Intel for the dollars PC sales now produce -- although most of the money has flowed to companies that support Intel's new industry order. Intel has relied on three weapons to exercise this dominance: its money, its technology and its ability to set PC design standards other companies must meet. Yet throughout the computer industry, critics are charging Intel with increasingly autocratic behavior to sustain and extend its domain. Many of these critics are reluctant to voice their fears publicly -- citing respect, fear or a host of other reasons. But the company's recent actions have led others to question openly whether Intel's use -- and potential abuse -- of its power now outweighs the benefits the company has brought to the industry. ''There's an argument that a benign dictatorship is the most efficient form of government there is, but Intel's recent actions bring into question how benign they really are,'' said Linley Gwennap, publisher and editorial director of the San Jose-based newsletter Microprocessor Report and a longtime observer of the chip industry. A federal judge in Alabama has ruled that the company broadly abused its microprocessor monopoly, and an FTC complaint into similar but narrower charges is headed to hearings early next month. In addition, the FTC is continuing a deeper probe into Intel's business practices -- a level of scrutiny that could ultimately rival the Microsoft antitrust investigation led by the Department of Justice. A hard look at the industry shows that Intel's impact is beyond dispute. Its decision in the early 1990s to enter the market for chipsets -- an electronic traffic cop that directs how data is moved between the processor and the rest of the computer -- drove many smaller companies out of that business. More recently, such larger companies as Acer Inc., Compaq, Digital Equipment Corp. and S3 Inc. have abandoned or surrendered the rights to promising technology rather than risk prolonged conflict with Intel. But as aggressive as Intel is, the company never has focused on destroying its rivals. And Intel has a compelling retort to the tales of companies that have suffered because of its actions: Look at how many have benefited. ''We've created wealth for more than 300 companies in the industry,'' said Executive Vice President Paul Otellini. ''Ask them if they're glad we took the lead in creating standards.'' Yet the biggest share of the wealth has gone to Intel. In 1997, Intel made $6.9 billion in profits on $25 billion in sales; on about the same sales totals, Compaq netted $1.85 billion. (Microsoft made $3.4 billion on sales of $11.4 billion.) And in 1998, when most computer companies saw their fortunes fall with the Asian financial crisis, Intel boosted annual sales yet again. In essence, Intel has usurped the traditional place of the PC manufacturers -- who once made their money through the challenging task of integrating components from dozens of different companies into a cohesive whole. Rather than technology providers, the PC companies now are more akin to distributors, filling the same role for Intel that car dealers do for General Motors or Ford. ''You no longer buy a Compaq computer, you buy an Intel computer from Compaq,'' said Tom Yuen, co-founder of PC maker AST Research. Some companies embraced this trade-off. ''It's not that we don't need or have technical expertise,'' said Michael Dell, the founder of Dell Computer Corp. ''But we focused on providing our 'value add' through other means, like distribution.'' But others chafed at Intel's command. ''Intel ended up being far from a partner,'' said Gary Stimac, once the senior vice president of product development for Compaq, and later a board member and consultant for Cyrix Corp., a competing maker of microprocessors. ''They were a dictator. I'm not saying that (Intel's) enhancements are bad, but they're very, very one-sided. . . To a certain degree, the innovation of the PC is no longer there.'' Even as Intel executives disagree with Stimac's characterization, they broadly concede part of his point. ''In a sense, PCs are now commodities, because most of the intellectual property is now tied up in the microprocessor,'' said Otellini. As evidence of Otellini's argument, consider how the industry's research and development spending is allocated: In 1997, Intel spent $2.34 billion on R&D; Compaq, Apple Computer Inc. and Dell Computer -- the first two of which once had extensive R&D operations -- spent $1.42 billion combined in the same year. Some PC makers tried to resist the revolution, but almost without exception they fell in line -- or fell back. One of the latter group was Irvine-based AST Research Inc., which around 1993 was No. 4 in worldwide sales. AST executives believed their company could continue to trade on its experience and expertise in building reliable PCs, rather than accepting Intel's all-in-one technology package. They continued to purchase some processors from Intel -- but also cut a deal to use some Advanced Micro Devices processors in their own PC design. The strategy was a bust. With clone makers offering eye-catching Intel technology at lower costs, AST had neither the performance advantage nor the broad product line to distinguish itself. It steadily lost market share, money, and ultimately its independence, becoming a wholly owned subsidiary of Samsung Electronics Group in 1997.To onlookers, AST's fate proved that Intel's crusade for standards was too powerful for any single company to defy. But some saw another message in the company's downfall: You cross Intel at your own risk. When the company decided to deal with AMD, Intel no longer treated AST as a most-favored customer, some former AST executives say. There were no public announcements, but insiders noticed that AST's Intel-based offerings -- such as a server designed to compete with Compaq's products -- started showing up months later than its rivals' products. AST, in this telling, was the victim not just of market forces but of retaliation. But Yuen, unlike some of his former colleagues, says that's not so. ''I'm sure there were probably suggestions at a lower level that we were damaging our relationship with Intel, but (Andy) Grove never called me up and threatened to cut off our supply,'' Yuen said. ''Was Intel's response something we thought very seriously about before we (made the deal with AMD)? Oh, yes. But most of (AST's) problems were caused by (the company's) own actions.'' Intel has consistently dismissed the allegation that it would penalize a customer simply for doing business with a competitor. But for many, the notion of a vengeful Intel remains gospel to this day -- and AST continues to be a prime example. ''We get responses from companies all the time that they like our product, but are afraid that they might upset Intel,'' lamented one competing chip executive. ''No one wants to be another AST.'' Whatever the truth of the AST story, there is no question that Intel emerged from the PC revolution with enormous power over its industry. Not only had its strategy been compelling, its execution had been nearly flawless -- a tribute to the company's top managers, from Grove on down. Intel's vision and competitiveness date back to co-founders Gordon Moore and Robert Noyce, who started the company in 1968. It has been sustained by a cadre of top executives who share engineering backgrounds (even the company's chief lawyer and its marketing guru are engineers) and long years of experience at Intel. But it is also true that the company's triad of weapons -- money, technology and influence over standards -- has been virtually irresistible. By any standard, Intel makes money at an astonishing pace. No other company logs annual sales from microprocessors that are even one-tenth of Intel's total. No other computer company can touch Intel's profits of $13 billion in the past two years; only far-flung IBM is close. In fact, Intel in 1997 was the sixth most profitable company in the world, surpassing General Motors and Philip Morris. And Intel's net profit margin -- profits as a percent of sales after expenses -- was 28 percent, more than 2 1/2 times any other company in the top 10. But what has ensured Intel's dominance isn't how much it makes. It's how much it spends -- and where it puts the money. Intel executives reinvested $6.5 billion in their company last year, some of it in surprising places. This hardware behemoth employs more than 1,000 software developers at its Intel Architecture Laboratories in Oregon. It funds hundreds of independent software and hardware developers to create products that use Intel technology. And it operates a separate $1 billion venture capital fund with investments in more than 200 companies. Were that fund spun off separately, it would be among the nation's ten largest, according to Venture Economics, a market researcher. ''It's enlightened self-interest,'' explained Grove. ''Say you are dealing with a games developer, a modest-sized company with a much shorter planning horizon and financial horizon. It's not obvious to them that their future depends on the next-generation Intel processor, so we have to give them a measure of encouragement,'' including money and technical assistance. ''If they don't bet on our success, we won't succeed,'' he continued. ''Sometimes you have to take a holistic view of business.'' Yet what Grove sees as encouragement provided by Intel's money, others see as manipulation. The prime example: the co-op portion of the ''Intel Inside'' campaign, which created a powerful financial incentive for computer makers to use and promote Intel processors. Intel unveiled this potent program -- similar to the co-op programs that retailers use to subsidize newspaper ads -- in 1991. While its importance has grown since, its structure has remained largely the same. If a manufacturer included the phrase ''Intel Inside'' in its advertising and on its PC packaging, Intel rebated up to 6 percent of the value of the manufacturer's orders. It was an attractive program, recalled Dan Hutcheson, president of VLSI Research, a market research firm in San Jose. But it had such an obvious catch that initially it seemed inconceivable manufacturers would agree. ''They had spent all this money and effort on creating a brand image for themselves,'' Hutcheson said. ''Why would they give that away to Intel for a few extra dollars?'' But in a hotly competitive market, those extra dollars were important -- particularly in years when PC profit margins fell. ''1992 was a bad year for a lot of companies,'' said Yuen, the former AST head. ''For some of them the question wasn't whether they would sell their brand equity to Intel or not, it was whether they would stay in business.'' Between 1991 and 1996, Intel returned $1.4 billion to hundreds of manufacturers through the ''Intel Inside'' program. But the payoff was enormous. ''We created brand name awareness for ourselves,'' said Dennis Carter, Intel's top marketing official and a corporate vice president. ''And we didn't even have a retail product.'' Competitors charge that Intel uses the co-op funds for other purposes as well. They say the company often targets specific markets where AMD or Cyrix has a strong product, crafting special promotions for manufacturers who choose Intel instead. In his April ruling in a suit filed by Intergraph Computer Corp. against Intel, federal Judge Edwin Nelson also took a dim view. The ''Intel Inside'' program ''reinforces and magnifies Intel's monopoly power,'' he wrote. Intel won't comment on the specifics of Nelson's ruling or competitors' charges, but has defended ''Intel Inside'' as no different in concept or execution from co-op funds run by Coca-Cola or General Motors. For all the infighting that characterizes the PC industry, it would be difficult to find anyone who wants to abandon standards. Standard designs helped touch off the PC explosion. They keep it going by accelerating the acceptance of hot technologies like DVD playback and high-speed modems -- which have a ready-made market of hundreds of millions of PC users. But what irks many in the industry is how standards are controlled. As a practical matter, the power is held by Microsoft and Intel. Intel officials have their own view. ''I don't think we dictate things,'' Grove said. ''I think we work with other people to establish innovations and improvements. Intel is often a prominent member of such groups, but they are set up such that one company doesn't dictate.'' That may be technically correct, given that multi-company ''standards bodies'' exist for almost every aspect of the PC. And it is also true that Intel has proposed its share of standards that never took off, such as its plans for ''I2O'' technology, intended to balance the internal operations of the PC. But Intel has won far more than it has lost, and industry-standards bodies have mattered little to that outcome. What matters instead is market power. ''Standards are really created by one company that's powerful enough throwing its weight behind something,'' said Michael Feibus, founder of Mercury Research, which tracks the chip industry. ''Standards committees are pretty ineffectual. That means Intel.'' Take servers, the powerful computers that dole out data over a network. Several years ago, Intel created a standard for these machines and made its design -- a sort of paint-by-numbers guide to building sophisticated computers -- widely available. That allowed dozens of companies to enter a market previously dominated by heavyweights like Compaq, IBM and Hewlett-Packard Co. Servers dropped in price, the leading companies lost market share -- and Intel sold more of its most expensive processors. That is only part of the power that Intel gains from setting standards. The chip giant also has gained enormous influence over the course of computer industry innovation, because no PC innovation would prove popular if it failed to work well with other components. In its June complaint against Intel, the FTC argues that Intel uses its monopoly power -- derived in large part from its control of microprocessor design -- to block other companies' efforts to push new technology. Witness its complicated relationship with the Santa Clara graphics chip company S3 Inc. In the mid-1990s, S3 was the dominant supplier of standard graphic chips, which accelerate a PC's processing of complicated visual information. As software increasingly included 3D graphics -- in everything from games to engineering models -- S3 set out to create its own de facto 3D graphics standard. The company's bright idea was to establish a direct pipeline between a PC's graphics processor and system memory, to speed the processing of memory-eating 3D files. The move required the integration of the graphics processor and chipset, a project S3 code-named Plato. After spending an unspecified amount of money on Plato -- company insiders have placed the total anywhere from hundreds of thousands to more than $1 million -- S3 brought the project to Intel. To succeed, the company needed key technical information to ensure Plato would be compatible with upcoming Intel microprocessors. But Intel said no -- in part, Intel executives say, because the Plato technology used so much memory that other PC functions could suffer, and in part because Intel at the time wasn't giving details of the Pentium II to anyone. ''That was the end of it,'' said Ron Yara, senior vice president and co-founder of S3. ''Without the information for building the chip around Intel's next-generation processor there was no point in coming out with the chipset. So we killed the project.'' But sources close to the companies indicate there was more to the story. S3's plan to bite off a piece of the chipset business made it a potential competitor of the leading chipset manufacturer -- Intel. And Intel was already building its own private highway in the PC -- the Accelerated Graphics Port -- that would be reserved for the passage of graphics data. Soon every graphics chip maker -- including S3 -- needed advance information on AGP if it was to match Intel's pace. Simply put, S3 found itself in an untenable position: trying to compete and cooperate with Intel at the same time. And when its focus on Plato left it behind rivals in the advancement of stand-alone graphics chips, it lost one of its biggest customers -- Intel again, which used S3's graphics chips on many of its motherboards. S3 has not regained Intel's motherboard business. But it did assure its access to Intel's technical information, without which it might not have survived. More recently, S3 signed a technology-sharing agreement with Intel, bringing the story full circle. PC memory has become so cheap that a Plato-like combination of chipset and graphics processor has more promise. But if such technology catches on it will probably be Intel -- not S3 -- that defines it. As far as Intel is concerned, there is nothing particularly remarkable about its interaction with S3. The graphics company tried to gain an upper hand in the market place, but failed because its plan did not offer sufficient benefits to PC users. That is the sort of mistake Intel strives not to make. ''One of the primary reasons we are able to create standards is that we focus on supporting technology that the entire industry could use,'' Otellini said. ''Too many other 'standards' are really proprietary technology.'' Each time Intel introduces a significant new microprocessor, there is a staged event, at which a top Intel executive approaches the cameras to proclaim the excellence of the product. An assortment of ''development partners'' -- 30 or so -- stands alongside, promising to build the chip into enticing new devices. More than media opportunities, these events offer clear illustration of the power of Intel's technology. Companies compete for access to Intel's wares, because building the best systems first is a vital advantage in the marketplace. Ask almost anyone at Intel, and you'll hear that this is as it should be. From graphics features to networking capabilities to the tenfold increase in processor speed since 1990, innovation has consistently distinguished Intel. ''If you want to know why we hold the position within the industry that we do, the number one reason is we deliver technologically superior products,'' Otellini said. ''Our technology is second to none.'' AMD and Cyrix occasionally may offer chips with slightly faster clock speeds at somewhat lower prices than Intel. Or they may show off integrated designs with less raw processing power but significant price savings. But even these chips, Intel executives note, usually mimic Intel designs that hit the market months or years before. Rival chip makers concede Intel's technological capabilities -- but some say that Intel isn't satisfied to compete on performance and price. Instead, they say, Intel uses technology as a weapon -- transforming the industry into an array of allies and enemies. ''Everywhere you turn, you run into a company that doesn't want to work with you, or supply components for your product because they don't want to offend Intel,'' noted one former Cyrix executive. ''It's like competing against a Japanese keiretsu, with all these interlocking companies.'' Intel's non-disclosure agreements with partner firms exacerbate this tension, rivals say. Companies often sign these agreements, commonly known as NDAs, when they decide to share technology but want to protect their intellectual property from wider circulation. But Intel often uses a restrictive variant known as the three-way non-disclosure agreement: It forbids signers from dealing with anyone except certified Intel customers and partners, who have their own agreements with Intel. To avoid disputes with Intel, those companies will sometimes err on the side of caution if there's the slightest doubt about whether they should reveal the information. ''We can't get the specifications of the next generation memory chip from a . . .manufacturer,'' complained one Intel competitor, describing a problem that could delay the release of a new chip by months. ''Intel's so big that they're a market all by themselves. (Suppliers) just don't want to take a chance.'' Intel says its NDAs are necessary to safeguard its technology and future product plans. So dominant is Intel's technology that few computer-industry companies, and probably no hardware companies, could survive a complete blackout. In its public disputes, Intel never has denied another company long-term access to products or technical information. But in recent years it appears to have grown bolder about threatening to turn off the switch. Such threats are at the heart of the FTC's June complaint against Intel. The FTC notes that Intel insisted that three firms -- Digital Equipment, Intergraph and Compaq -- return critical technical information shortly after each sued Intel for patent infringement. Confronted with Intel's demands, Digital swiftly settled its case, receiving money and product discounts totaling more than $1 billion from Intel in exchange for technology rights and a chip-manufacturing facility. Intergraph continued to push its complaints in court, and won a preliminary guarantee of access to Intel technology as part of Alabama Judge Nelson's ruling earlier this year. But the more telling illustration of Intel's power comes from its years-long chess game with the one PC company whose size and resources approach Intel's: Compaq. Compaq has long sought to maintain the upper hand in its relationship, though its efforts often have flopped. In 1994, it sought to continue pushing profitable systems based on the then-standard 486 microprocessor, rather than moving quickly to Intel's new Pentium. In a matter of months, the Texas computer maker found itself swamped with obsolete systems, which it had to dump while rapidly introducing Pentium-based computers. That, however, was merely the end-game of a yearlong struggle between the two companies for leadership of the industry. Earlier in 1994, Compaq had endorsed AMD's 486 processor line, and had promised to feature the AMD-K5 -- a Pentium-class chip -- in future products. But as Compaq reached that agreement with AMD, it encountered restrictions on its access to advance information about Intel's most powerful processor, the Pentium Pro. This would have been tough to withstand under any circumstances. But Compaq executives say they soon came under pressure from key customers to get back with Intel's program. ''Intel went directly to major corporate accounts to tell them what its next generation processors would be and also what the computer architectures would be around those chips,'' said Stimac, formerly the senior vice president of product development for Compaq. ''If you didn't follow (Intel's) litany of presentations with many major accounts they would come back at you with 'Intel said X, Y and Z.' '' Compaq's frustration soon boiled over. In September, the company's normally taciturn CEO, Eckhard Pfeiffer, appeared at a little-known trade show in Spain and offered a stunning denunciation of Intel's attempts to control the industry. The tirade stopped just short of accusing Intel of illegal behavior. But Pfeiffer did complain that Intel was using its control over the microprocessor to dictate terms to the computer industry as a whole. Then, in December, Compaq filed suit against Packard Bell, alleging that Packard Bell had violated Compaq's patents. It was clear to most in the industry that the suit's real impact would fall on Intel, which was giving key technologies to Packard Bell and other PC makers, and indeed Intel was ultimately added to the suit. But by then, Compaq's gambit had largely failed: AMD encountered trouble with its K5, which wouldn't be ready in time to replace the Pentium. In January 1996, Compaq signed a technology cross-licensing agreement with Intel that effectively brought Compaq into the fold. As part of the agreement, it granted Intel rights to some of the disputed technology in the Packard Bell case. Today, most Compaq executives downplay the earlier dispute. But Stimac remains bitter. ''They withheld information from us because we didn't want to sign over our intellectual property,'' he said. ''We probably could have brought it to the Justice Department, but by the time it would play itself out, it would be a lose/lose situation.'' Intel executives dismiss Stimac's complaints as typical of the posturing that goes on in negotiations. ''A lot of what Compaq wanted was better pricing,'' Intel's Otellini noted. ''They felt as our biggest customer they should get a better deal. We don't tend to do that.''
|
