(A3.1) If you were an accountant for Intel, what specifically would be the relevant accounting research question with respect to the Intel Pentium chip flaw?
(A3.2) What constrains Intel's decisions about how to account for the Pentium chip flaw?
(A3.3) What do you need to know, estimate, and assume to answer the research question?
(A3.4) What would you recommend to Intel management with respect to accounting for the Intel Pentium chip flaw? Why?
Accounting for the Intel Pentium Chip Flaw 9-101-072 R E V : O C T O B E R 4 , 2 0 0 2 ________________________________________________________________________________________________________________ Research Associate Lisa Brem prepared this case under the supervision of Professors Gregory S. Miller and V.G. Narayanan. This case draws on material from “The Intel Pentium Chip Controversy (A) and (B), “HBS Nos. 196-091 and 196-092, prepared by Research Associate James Evans under the supervision of Professor V.G. Narayanan. HBS cases are developed solely as the basis for class discussion. Cases are not intended to serve as endorsements, sources of primary data, or illustrations of effective or ineffective management. Copyright © 2001 President and Fellows of Harvard College. To order copies or request permission to reproduce materials, call 1-800-545-7685, write Harvard Business School Publishing, Boston, MA 02163, or go to http://www.hbsp.harvard.edu. No part of this publication may be reproduced, stored in a retrieval system, used in a spreadsheet, or transmitted in any form or by any means—electronic, mechanical, photocopying, recording, or otherwise—without the permission of Harvard Business School. G R E G O R Y S . M I L L E R V . G . N A R A Y A N A N Accounting for the Intel Pentium Chip Flaw What we view as an extremely minor technical problem has taken on a life of its own. —Andrew S. Grove, Chief Executive Officer, Intel Corp.1 In December 1994 IBM stopped shipping personal computers containing the Intel Pentium microprocessor. It supported this action with analysis provided by the IBM Research Division, which indicated that “common spreadsheet programs, recalculating for 15 minutes a day, could produce Pentium-related errors as often as once every 24 days. For a customer with 500 Pentium-based PCs, this could result in as many as 20 mistakes a day....”2 Andrew Grove, Intel’s president and CEO retorted quickly: “[b]ased on the work of our scientists analyzing real world applications, and the experience of millions of users of Pentium processor-based systems, we have no evidence of increased probability of encountering the flaw...You can always contrive situations that force this error. In other words, if you know where a meteor will land, you can go there and get hit.”3 Although Mr. Grove's opinion of the Pentium “bug” was clear, IBM's action forced Intel's top management to address a tough question: Should Intel undertake a broad recall of Pentium Chips? History Microprocessing chips are the “brains” of computers sold for both home and business use. Intel was the first, the biggest, and regarded by many as the best microprocessor-maker in the computer industry. Intel's 1993 gross margin of 63% and net margin of 26% were clear indications of the strength of the company (see Exhibit 1 for Intel's recent financial statements). Intel's success was based on constant innovation, huge spending in marketing, production and cycle-time improvement, 1Aaron Zitner, “Consumers get break on computer chip,” The Boston Globe, December 21, 1994, p. 1. 2Bart Ziegler, “IBM, Intel Continue Their Wrangling Over Liability for Pentium Repair Bill,” The Wall Street Journal, December 14, 1994, p .B6. 3[[Intel]]
[email protected], “Tell us what you think!,” Copyright (c) 1994, Intel Corporation. All rights reserved. This document is authorized for use only by Jennifer Robinson in ACCT-6140-1,Current Trends Acct Standards.2020 Summer Sem 05/04-08/23-PT2 at Laureate Education - Walden University, 2020. 101-072 Accounting for the Intel Pentium Chip Flaw 2 and relentless challenging of copycat competitors by “pummeling rivals in court.”4 Intel’s first microprocessors for computers, the 8086 and 8088 (sold in the first IBM personal computer), were followed by its 80186 (used primarily in controllers), 80286, 80386, and 80486 editions, each with markedly more memory and processing speed than the previous generation (see Exhibit 2). By 1993, when the first Pentium was introduced, a number of competitors—including AMD, Cyrix, and NexGen—were already producing clones of Intel’s 486. These competitors were getting faster at cloning Intel's designs, and the 486 clones appeared on the market much faster than the clones of 386 and 286 chips. The Pentium had been designed with more testing than any previously produced manufactured product.5 “They ran the hell out of all the software they could find,” said Joseph Costello, CEO of a chip-design software maker.6 A full year of testing involving trillions of random mathematical processes was performed to detect any problems with the chip before and soon after it went on the market; a quadrillion more were performed in the year after that. However, the rollout of any new microprocessor was unavoidably risky since chip-testing technology was a generation behind chip- making technology. In addition to detailed testing, Intel had strategically prepared this new chip for market with the increased trademark protection available to a name, the “Pentium,” rather than using the 80586 moniker. Intel began to steer the market toward this newer, faster chip—and away from competitors’ clones of prior models—by attempting to convince personal and corporate computer users that the 486 was outmoded. With an advertising blitz estimated at $150 million, the company advertised the logo Intel Inside® and the name Pentium in every available type of media. By the end of September 1994, Intel had shipped nearly two million Pentium chips to computer producers (see Exhibit 3). The Flaw Discovered Early in the summer of 1994, mathematics professor Dr. Thomas Nicely of Lynchburg College in Virginia discovered inconsistencies in his calculations performed on his Pentium-driven PC. Nicely was trying to prove that PCs could do mathematical work heretofore only performed on larger systems and thus was involved in intense and continuous number crunching far beyond that of a typical user. Nicely discovered the division flaw occurred only with rare combinations of numbers, and was not in his software, but in the processor of his Pentium.7 The problem arose in the floating-point processing unit of the chip, which handled numbers expressed in scientific notation. At the end of October, Nicely published a note on the Internet querying other users about the Pentium flaw. A discussion soon emerged at the Internet news group “comp.sys.intel.” The tone quickly changed from a calm discussion of arcane technical tests to flaming accusations and threats aimed at Intel. 4Don Clark, “Intel Balks at Replacing Pentium Chip Without Asking Owners Any Questions,” The Wall Street Journal, December 14, 1994, p. A3. 5Jim Carlton and Stephen Kreider Yoder, “Computers: Humble Pie: Intel to Replace Its Pentium Chips,” The Wall Street Journal, December 21, 1994, p. B1. 6Robert D. Hof, “The ‘Lurking Time Bomb’ of Silicon Valley,” Business Week, December 19, 1994, p. 118. 7As one example of the flaw, the solution to the following calculation, 4,195,835—( (4,195,835 / 3,145,727)) x 3,145,727, should be zero, but a computer with the flawed Pentium chip provides an answer of 256: Walter S. Mossberg, “Intel Isn’t Serving Millions Who Bought Its Pentium Campaign,” The Wall Street Journal, December 15, 1994, p. B1. This document is authorized for use only by Jennifer Robinson in ACCT-6140-1,Current Trends Acct Standards.2020 Summer Sem 05/04-08/23-PT2 at Laureate Education - Walden University, 2020. Accounting for the Intel Pentium Chip Flaw 101-072 3 On November 7, 1994, an Electrical Engineering Times article by Alexander Wolfe, based on the Internet discussions, prompted Intel’s response that it had uncovered the flaw during tests the previous June. However, Intel had run a series of tests and concluded that an error would occur only once every nine billion random calculations, or every 27,000 years for most users. Further, Intel had remedied the problem for the next planned version of the Pentium, but had not informed customers who had purchased a flawed processor. Following the article, Intel offered to replace the flawed chips, but only on a limited basis: users first had to demonstrate that the flaw was likely to occur in the work they performed on their computer. The Internet discussion group continued to “flame” the company on-line and eventually attracted the attention of reporters. On November 22, 1994, CNN broadcasted a story that revealed the chip flaw for the first time to the general public, Intel’s own computer-making customers, and the rest of the media. On November 24, the beginning of the Thanksgiving holiday weekend, the front-page of the New York Times Business Section headlined “Flaw undermines the accuracy of Intel’s Pentium.”8 The Boston Globe carried the same story on its front page, and the news continued to unfold over the next month. By November 25 Intel’s stock had dropped two percentage points from its high in late September. The mounting consumer pressure to fix the Pentium flaw was hitting PC manufacturers as well as Intel. But who would take responsibility for fixing the flaw? Dell, a computer manufacturer, began advertising its Pentium with a built-in computer fix to remedy the latent flaw. On November 28, Sequent, a mainframe manufacturer, stopped shipping Pentium machines until a software solution could be installed.9 On November 30, IBM—a major Intel customer—announced that it would replace the Pentium processor in any of its machines at the customer’s request. IBM, however, did not have Intel’s support to fulfill such a promise and ran the risk of having to purchase replacements on its own account. As the second week of December passed, the media coverage began to abate and Pentium flaw stories lost their front-page status. December Pentium sales continued to increase as planned, but several thousand Pentium owners were calling Intel daily. Intel rallied over a thousand of its employees to respond to these calls and carefully assess whether the users were performing functions that would be at risk of engaging the flaw in the floating point unit.10 On December 12, 1994, IBM, without