Inside Intel

COPYRIGHT

Fourth Estate

An imprint of HarperCollinsPublishers Ltd.

1 London Bridge Street

London SE1 9GF

www.harpercollins.co.uk

First published in Great Britain by HarperCollinsPublishers 1997

Copyright © Tim Jackson 1997

Tim Jackson asserts the moral right to be identified as the author of this work

All rights reserved under International and Pan-American Copyright Conventions. By payment of the required fees, you have been granted the nonexclusive, nontransferable right to access and read the text of this ebook on-screen. No part of this text may be reproduced, transmitted, downloaded, decompiled, reverse-engineered, or stored in or introduced into any information storage and retrieval system, in any form or by any means, whether electronic or mechanical, now known or hereafter invented, without the express written permission of HarperCollins ebooks

HarperCollinsPublishers has made every reasonable effort to ensure that any picture content and written content in this ebook has been included or removed in accordance with the contractual and technological constraints in operation at the time of publication

Source ISBN: 9780006387977

Ebook Edition © DECEMBER 2016 ISBN: 9780008240615

Version: 2017-01-05

DEDICATION

To Emily Marbach

CONTENTS

COVER

TITLE PAGE

COPYRIGHT

DEDICATION

PROLOGUE: A Half-Billion-Dollar Mistake

PART ONE · INNOVATION

1 The Odds-On Favourite

2 2,000 Starts a Week

3 The Third and Fourth Men

4 Into the Potato Patch

5 A Saviour from Bucks County

6 The Rebels

7 Yellow Snow

8 The Microprocessor

9 Public Company

10 Second Source

11 Turning Failure to Profit

12 A New Standard

PART TWO · DOMINATION

13 Borovoy Wins a Patent Battle

14 A Competitor on the Horizon

15 Penang Burning

16 Gopen Beats the Union

17 The Stop-Gap

18 Marriages and Divorces

19 Organization and Alpha Particles

20 The Microma Mistake

21 Crush!

22 Whetstone’s Design Win

PART THREE · EXCLUSION

23 Seeq and Destroy

24 Checkmate Powell

25 Microcode

26 Gold

27 The Vancouver Complaint

28 A Scandal in Malaysia

29 Davidian’s Bonus

30 The New CEO

31 An Anonymous Caller

32 Lagging the Koreans

33 Raising the Tax

34 The Two Webbs

35 Departures

36 A Question of Drafting

37 The Traitorous Two

38 A Hacker Inside

39 Tech Support Screws Up

40 The 10x Force

41 Winning the Platform Wars

KEEP READING

NOTES

INDEX

ACKNOWLEDGEMENTS

ABOUT THE AUTHOR

OTHER BOOKS BY

ABOUT THE PUBLISHER

PROLOGUE

A Half-Billion-Dollar Mistake

ARTHUR ROCK WAS PROBABLY standing staring out of the floor-to-ceiling windows, gazing at the evening lights of the December city skyline, when the call came in.

His office suite, on the twelfth floor of a tall tower in San Francisco’s financial district, had a group of comfortable sofas at one end dominated by a collection of bold modern pictures. But Rock, who had suffered a bout of polio as a child, liked to take calls standing up. Next to his desk stood a light-coloured wooden reading lectern, which allowed him to work on his feet – like a clerk in a nineteenth-century government office.

It was hard, though, to imagine anyone further removed from the nineteenth century. Rock was a billionaire venture capitalist, who had made his money by investing in some of the most successful technology companies in history. He saw his job as doing more than simply picking winners and then sitting back to watch them get on with it. When Arthur Rock invested money in a company, he sat on its board of directors and helped to guide its strategy. In times of crisis, when his investments might be at risk, he’d be ready if needed to step in with firm advice on how the company should recover its correct course.

The incoming phone call gave no clue that today might be one of those occasions. As Rock’s secretary put the call through to his speakerphone, he heard the soft-spoken voice of Gordon Moore, chairman of Intel Corporation, opening the proceedings with his customary calm. Ever methodical, Moore would often start Intel’s board meetings by checking that all the participants had received the appropriate papers from Jean Jones, his trusted secretary for nearly thirty years.

Rock didn’t need to be told why the Intel directors were holding today’s board meeting by conference call instead of at the company’s headquarters in Santa Clara, just over an hour’s drive south of the city. Like the other Intel directors, he had read all about the crisis in the New York Times and the Wall Street Journal.

Intel’s heavily advertised flagship product, its new Pentium microprocessor, was flawed. The company had known about the flaw for some months and kept quiet about it, believing that only a tiny fraction of the millions of users of the chip across the world would ever be inconvenienced by the problem. But the flaw had been discovered by a mathematics professor, whose data had been posted on the Internet. Intel’s attempt to play down the issue had angered customers, worried investors, and ultimately provoked a storm of criticism in the media, culminating in a damaging report on CNN.

As the problem escalated, Intel’s response had been doggedly consistent. The company had kept repeating that the flaw was unimportant, and kept insisting that it required no corrective action from most owners of computers equipped with a Pentium chip. Intel had admitted that a small number of specialized users, mostly scientists or mathematicians, might need their Pentiums replaced. But Andy Grove, the company’s combative chief executive officer, insisted that the people best equipped to decide who was in this category were Intel’s own engineers. As a result, customers who wanted to return a flawed Pentium would have to call Intel and convince the company that they really needed a replacement.

Intel’s approach to the issue had the support of the computer industry. The last thing that makers and retailers of PCs wanted was to be put in a position where they had to open up hundreds of thousands or even millions of PCs already in the distribution channel on their way to end-users, and go through the time-consuming and expensive business of replacing their existing processors with new ones. So the industry, which understood from experience that no microprocessor was ever perfect on first launch, stood staunchly behind its leading chip manufacturer. Company after company issued press releases telling customers that the Pentium flaw was nothing to worry about, and Intel’s approach to the problem was just fine.

That wasn’t how the general public saw the issue. Inflamed by critical comments on the Internet, newspaper readers and TV viewers couldn’t understand why a maker of microprocessors should be any different from the rest of manufacturing industry. If you were in any other business and discovered a major problem with one of your products, you offered customers a free exchange, no questions asked. It was as simple as that. You didn’t screw them around by telling them the problem probably wouldn’t affect them and force them to call you and jump through hoops before they could get a replacement. Yet that was what Intel seemed to be doing – and its attitude seemed to be brazenly arrogant, since the problem with its new chip, advertised as the latest and greatest in computer technology, was a simple one: it couldn’t do long division properly.

For a few uneasy weeks the alliance of expert insiders had succeeded in defying the instincts of the less well-informed outsiders. It had looked as though Andy Grove’s boldness and willingness to take criticism and unpopularity would eventually triumph. The chorus of complaints against Intel had begun to die down, and consumers had seemed to be getting the message that the Pentium flaw wasn’t such a big deal after all.

But the situation was precarious. It depended on the giants of the computer business maintaining complete unanimity. If just one major vendor changed sides, then it would no longer be possible for Intel to convince its customers that all the experts agreed there was no problem.

That change had now happened. International Business Machines, the computer giant that had unwittingly helped Intel rise to dominance of the industry, had broken ranks. For reasons that Intel had not been able to fathom, the company had put out an announcement saying that it was putting shipments of new machines equipped with Pentium chips on hold until the issue was resolved.

The result was pandemonium as Intel’s customer-support lines were flooded with anxious callers. A new round of criticism was unleashed in the media. Finally, the company was forced to face one of the toughest decisions of its history. Should it guarantee to exchange all the flawed chips – even for customers who used their computers for nothing but playing games – and face up to wasting half a billion dollars on an unnecessary returns programme? Or should Intel stick to its guns? Should it save some money in the short term, but run the risk of throwing the computer industry into turmoil and causing long-term damage to the value of its brand?

This, Gordon Moore explained, was the question that Intel’s board now had to decide.

Maybe Arthur Rock should sit down after all.

The decision the Intel board took that day was undoubtedly the right one. Reversing completely its former position, the company announced that it would now offer a no-questions-asked exchange to anyone who owned a flawed Pentium. Andy Grove issued a grovelling public apology for the irritation that his previous stance had caused, and the company set up an emergency call centre to handle the expected rush of customers dialling in to take up the offer. To cover the costs of the incident, the company set aside $475m – more than half its earnings during the last quarter of 1994, when the board meeting took its decision – and turned one million of the world’s most advanced microprocessors into nothing more than ornamental tie-pins and key chains.

Once Intel had backed down, the bad publicity went away almost instantly. Within weeks, public confidence in the corrected Pentium chip began to return. Within months, the two class-action lawsuits that had been filed against the company as a result of the Pentium flaw began to move to settlement. And Intel, scarred and chastened by the experience, began to return to the business it knew best: building microprocessors for personal computers.

Today, three years later, the incident has been forgotten by most people other than specialists in the microprocessor industry. Intel’s brand is stronger than ever, and its share price has nearly quadrupled since that telephone board meeting. In retrospect, the Pentium incident now looks like an aberration, a buying opportunity, a hiccup in the otherwise irresistible rise of one of America’s most powerful and successful companies.

But was it an aberration? Or did the Pentium scandal reveal something important about Andy Grove and the company that he ran?

Intel Corporation, founded in 1968, can claim credit for inventing some of the most important technologies of the modern electronics industry and bringing them successfully to the mass market. Towering among its achievements is the invention of the microprocessor, which has brought the power of computing to the desks of hundreds of millions of people all over the world, and has changed the world around us by putting intelligence into appliances ranging from vacuum cleaners to cell phones and from toys to cars.

The conditions that gave rise to this extraordinary spike of scientific innovation were put in place by two men, Robert Noyce and Gordon Moore. Both of them were leading figures in California’s electronics industry, long before the San Francisco Bay Area even became known as Silicon Valley. It was their charisma, their leadership, their contacts and their reputation that brought together a group of the most talented engineers in the world and established the framework that allowed scientific creativity to flow.

But it was only for a short time that innovation was the most significant thing that Intel did. When it released the world’s first microprocessor to the market, Intel was just two years old, and had been a public company for only one month. It did not yet stand out far from the crowd of other startup companies trying to make money from the uncertain new technology of integrated electronics.

So Intel changed. From being just an innovator, it became a company whose objective was to deliver – to make sure its good ideas were turned into practical products that customers could use, products that arrived on schedule and at prices that fell consistently year by year. This transformation was no mean feat. It forced Intel to become rigorously organized and focused, and to find a balance that allowed it to keep firm control over its operations without jeopardizing the creativity of the scientists who were its greatest commercial asset.

The result of this transformation was that Intel rose to domination of its industry. Its memory chips, the products that generated most of its sales, swept the mainframe computer industry by storm, and its microprocessors became the standard on which the entirely new industry of the personal computer was built. It wasn’t the intrinsic merits of Intel’s products that brought about this domination. Instead, it was more apparently banal things like distribution, customer support, product range, documentation, and technical development tools.

The process was by no means smooth. Nearly nine years after Intel created the microprocessor, the company found itself running last in a three-way race for market share. Yet the company refused to give up. In a matter of days it created a campaign to convince its employees that regaining leadership of the microprocessor industry, and crushing its leading competitor, was a matter of survival. The campaign, called Operation Crush, worked like a dream. Intel’s microprocessor, acknowledged even by its own engineers as technically inferior to the competition, had become the industry standard. Almost by accident, a later version of it was chosen by IBM as the basis for the IBM Personal Computer.

As the PC began to change the face of the computer industry in the 1980s, Intel once again went through a transformation. Now that it was the industry leader, the company no longer needed to focus on delivering ground-breaking new products or using marketing campaigns to overthrow a more powerful rival. Instead, the key to its continued success was to keep challengers at bay and to attack any threat to its high profit margins. So Intel planned and carried out a ruthlessly brilliant programme to change the rules of the chip industry. Instead of authorizing other companies to base their products on its designs – a practice, known as ‘second sourcing’, that was accepted as the only way to give customers security of supply – Intel resolved secretly that it would become the monopoly supplier of its chip designs. To achieve this goal the company had to flout a long-term technology agreement with a key partner. But the stakes were too high for this to be an obstacle.

So lucrative was Intel’s monopoly, though, that the company began to face the problem that plagues all successful technology companies: how to stop others from setting up new companies that threatened its position. Here Intel responded in a style that would have earned the respect of any general on the battlefield. It launched a string of long-running lawsuits – civil and criminal, state and federal – against competing chip design teams, former employees, semiconductor manufacturing plants, venture capitalists and at one stage nearly even going to war with the computer companies that were its own customers. Intel’s legal department spent hundreds of millions of dollars, and the general counsel who headed it was at one point told that one of the targets he would have to meet in order to get a good performance appraisal was a fixed number of new lawsuits to start each quarter. The strategy of suing everyone in sight wasn’t likely to win many friends, and Intel lost or settled more cases than it won. But the policy of filing writs first and asking questions later helped Intel to hold on to its monopoly profits for longer.

The other focus of this ‘exclusion’ phase of Intel’s history was branding. After years of giving its products only part numbers instead of names, the company realized at the end of the 1980s that the ultimate way to keep competitors out was to make consumers associate the Intel name with high quality and reliability. So the company ran a succession of programmes encouraging PC buyers to concentrate on the processor inside, not the name on the box. First came the ‘Red X’ campaign, in which a big red X was spray-painted over the name of an outdated Intel chip that the company wanted to supersede. Then came ‘Intel Inside’, in which the company subsidized the cost of PC manufacturers’ advertisements if they included the right slogan and logo in their copy. Finally, in a move that made it dramatically harder for competitors to attract the attention of customers, the company switched from numbering its processors to giving them names.

Brilliantly successful at excluding competitors though it was, this campaign had a side-effect. It infuriated many of the most powerful players in the PC industry. They saw Intel’s promotion of its brand as a direct threat to their own, and a move that would tilt the balance of power against them in favour of low-cost, no-name companies that would now have the credibility of Intel’s backing. In one startling moment, the CEO of Compaq Computer went public with a withering attack on Intel’s strategy.

Though the issue continues to rumble on, it has to some extent been overtaken by events. Intel’s domination of the microprocessor industry is now so complete that the company has little to gain from making life harder for rivals building processors that conform to the Intel instruction set, or chip makers promoting entirely different standards. In the late 1990s Intel is in the position of a gardener who has marked out a plot and removed all the weeds. Its job now is cultivation. In this fourth and latest phase of the company’s history, the objectives are to make sure that every time a consumer walks out of a store with a PC in a cardboard box, more of the total selling price of that machine falls straight to Intel’s bottom line; to persuade consumers to replace their PCs more often – or as a second best to keep their old machines, but upgrade the Intel processors inside them; and to make the PC as a whole a more attractive product, so that people choosing between buying a new TV and a new computer will make the choice that brings in profits for Intel.

This new ‘cultivation’ phase puts Intel in the position of being less a competitor to other companies than a leader for an entire industry. It has given the company an incentive to develop new technologies not because it hopes to make money from them directly, but because they can increase overall demand for computers. For instance, Intel developed a piece of software which allowed people to make telephone calls over the Internet – and gave it away, inviting people to download it for free from its web site. No matter that there were other companies trying to sell such software, or that Intel’s new package might threaten an existing line of conferencing products that the company had developed. The point was simply that the new software package gave customers without a PC a new reason to buy one.

With booming sales and profits, fewer threats on the horizon than for many years, and tens of thousands of employees whose loyalty is assured by the hundreds of thousands of dollars they stand to make from Intel stock options, the company seems almost unassailable. But it has its weaknesses – and they, like its strengths, are intimately tied to the personality of one man. To a greater degree than most outsiders realize, Intel is the personal creation of its chief executive, Andy Grove.

Andy Grove, a Hungarian refugee who anglicized his name after arriving in the United States by boat in 1956, is one of the most extraordinary figures in American business. He is brilliantly intelligent and articulate, driven, obsessive, neat and disciplined. Intel has been built in his image. The values taught in the company’s private ‘university’ – directness in confronting problems and extreme rationality in approaching management questions – are an extension of Grove’s own personality.

Andy Grove’s slogan – some have called it ‘Grove’s Law’ – is ‘only the paranoid survive’. Daily life inside Intel follows this maxim to the full. By comparison with Microsoft, the company is almost obsessively secretive. The house joke is that its photocopiers are fed with paper that is already marked ‘Intel Confidential’ at the top of every page. The company also operates a security department whose job is surveillance not only of competitors or thieves that might harm Intel’s interests, but also of the company’s own employees. This department has several times crossed the boundaries of what is considered proper behaviour in US corporate life.

In a recent book Grove claims to have elevated paranoia to a tool of management. He argues that it helps companies to watch out for dramatic changes in the business environment that faces them – changes that he calls ‘10x forces’ because they are ten times more powerful than the forces normally encountered – and to respond quickly to them. Yet the irony is that Intel’s own record here is patchy at best. The first 10x force facing Intel in its history was the replacement of old-fashioned core memory devices in the computer business with smaller, cheaper, faster integrated circuits. The company responded brilliantly: it recognized the trend, led the change, and grew to become a significant industry player as a reward for its vision.

Later 10x forces have been recognized more slowly. A case in point is the microprocessor. Although it was Intel that sold the world’s first microprocessor, the company was very slow to see the potential of its creation – only taking it seriously when two of its best engineers left to set up in competition. For years it viewed the device as a component to be used primarily in industrial controls rather than in computers, and it turned down an early suggestion by a team of its own scientists to build the world’s first ‘desktop computer’, and a gift from inventor Gary Kildall of the operating system that could have taken the place of Microsoft’s MS-DOS. But at least the outcome – Intel’s domination of the PC microprocessor business – suggests that little harm was done.