Int. J. Services and Operations Management, Vol. 5, No.

6, 2009 799

The effect of technological platforms on the global

supply chain: a case study on Intel’s platform
business in the PC industry

Hirofumi Tatsumoto*
School of Business Administration
University of Hyogo
8–2–1, Gakuen-nishimachi, Nishi-ku
Kobe-shi, 651–2197, Japan
E-mail: tatsu@biz.u-hyogo.ac.jp
*Corresponding author

Koichi Ogawa
Intellectual Asset-based Management Endorsed Chair
University of Tokyo, Japan
E-mail: NAE01471@nifty.ne.jp

Abstract: The purpose of this paper is to investigate the effect of the

technological platform on the global supply chain by examining its
characteristics and diffusion mechanism. Using architectural analysis, we
developed a case study on Intel’s platform business in Taiwan in 1990s. It
defines the characteristic of platforms as a separator of the technologies used in
finished products into completely modular and integral technologies. It explains
the difference in technology diffusion speed at a finished product level and a
platform level, and shows the shift in production of finished products from
developed countries to developing countries, while the platform remains in the
developed country. The growth of platforms has destroyed the advantages of
traditional firms based in developed countries and encouraged market entry by
new firms based in developing countries.

Keywords: product architecture; technology diffusion speed; separation

mechanism of platform.

Reference to this paper should be made as follows: Tatsumoto, H. and

Ogawa, K. (2009) ‘The effect of technological platforms on the global supply
chain: a case study on Intel’s platform business in the PC industry’, Int. J.
Services and Operations Management, Vol. 5, No. 6, pp.799–816.

Biographical notes: Hirofumi Tatsumoto is an Associate Professor of

International Business Management at the University of Hyogo in Japan. He
holds an MA in Economics from the University of Tokyo. His main area
of research is strategic management in particularly the electronics, software
and semiconductor industries. He is part of the research project on
‘International Standardisation’ of the New Energy Development Organization
(NEDO), Japan. His research was published as chapters in Monodukuri
Management, Economic Analysis of the Game Industry and Strategic Use of
Consensus-based Standards and he has published articles in the SEC Journal

Copyright © 2009 Inderscience Enterprises Ltd.

800 H. Tatsumoto and K. Ogawa

(accepted by the Information-Technology Promotion Agency Japan or IPA),

Organizational Science, the Journal of Intellectual Property Association of
Japan and Akamon Management Review (all in Japanese).

Dr. Koichi Ogawa is a Professor of the Intellectual Asset-based Management

Endorsed Chair, University of Tokyo, Japan. Professor Ogawa holds a Doctoral
degree in Electronics Engineering. His paper ‘Architecture-based approaches
to international standardisation and as evolution of business model’ with two
co-authors was awarded second place in the International Electrotechnical
Commission (IEC) Century Challenge in 2006. He has been in charge of R&D
and business development for 30 years in Fujitsu Ltd., Japan. Previously, his
main field of research was computer storage technology; since 2004, it has
been on the management of technology and product-architecture theory. His
articles have been published in journals including the Journal of Applied
Physics, the Journal of IEEE Transactions, Japan Journal of Applied Physics
and Akamon Management Review. He was one of the Japanese delegates to
ISO/JTC1/SC23 as the chairman of a local working group, and contributed to
the international standardisation of the 5.25-inch WORM and 3.5-inch MO
optical disk cartridge.

1 Introduction

The effect of product architecture on innovation is one of the central notions in

the existing research on technological innovation (Abernathy and Utterback, 1978;
Henderson and Clark, 1990; Christensen, 1992; Ulrich, 1995; Baldwin and Clark, 2000).
Product architecture is the basic concept used to allocate the functions of a product to its
components, and is classified into modular and integral architecture (Ulrich, 1995).
Modular architecture includes a one-to-one mapping from functions to components,
and has decoupled and ruled interfaces between components (Ulrich, 1995; Baldwin
and Clark, 2000; Fujimoto, 2007). In contrast, integral architecture includes a complex
mapping from functions to components, and has coupled and rough interfaces between
components (Ulrich, 1995; Fujimoto, 2007).
The growing use of modular architecture increases the importance of innovations in
modular cluster industries (Langlois and Robertson, 1992; Sanchez and Mahoney, 1996;
Baldwin and Clark, 2000). The rise of the computer industry in Silicon Valley is the most
typical and successful case of modular cluster innovations (Baldwin and Clark, 2000).
The platform has played an important role in modular cluster innovations (Gawer
and Cusumano, 2002; Iansiti and Levien, 2004). In modular architecture, the platform is
the special module that provides a way to mix and match other modules and construct
numerous combinations to seek the most valuable design. The platform promotes the
building of the ecosystem of complementors and leads to modular cluster innovations
(Baldwin and Clark, 2000; Gawer and Cusumano, 2002; Iansiti and Levien, 2004).
However, little is known about the diffusion mechanism of platforms before
establishing the ecosystem and modular cluster innovations. Early studies on platforms
have focused on the management of the self-sustaining process in modular cluster
innovations after the adoption has reached a critical mass (Gawer and Cusumano, 2002;
Iansiti and Levien, 2004), in which network externality fuels further growth (Shapiro and
Varian, 1998; Rogers, 2003).
 The effect of technological platforms on the global supply chain 801

The explanation of the diffusion process in terms of the ecosystem has produced
important insights, but it is fundamentally incomplete. The ecosystem model has revealed
the diffusion process after the critical mass, but the diffusion mechanism before the
critical mass is still not clear, and there remains an unanswered question about how the
platform obtains the shipment volume to create the ecosystem before the critical mass.
To understand the diffusion mechanism of platforms, the global supply chain
needs to be considered because the platforms are often used in products produced
by firms based in developing countries. Recent research on global supply chains has
suggested that platforms as turnkey solutions have established new global production
networks (Sturgeon, 2002; Berger and the MIT Industrial Performance Center, 2005;
Shintaku et al., 2006). They emphasised the platform-based production network, such as
contract manufacturing in value chain modularity.
The existing diffusion models of platforms, which are based on both the ecosystem
and the global production network, provide little insight into the reasons for the rapid
and wide diffusion of the platform before the critical mass and for the resulting change
in the global supply chain. In this paper, we develop and apply a new model derived
from a case study on Intel’s platform business in Taiwan in the1990s that helps in
understanding the characteristics of the platform, the diffusion mechanism into firms
based in developing countries and the effect on the global supply chain in detail.

2 Literature review and conceptual framework

2.1 Two types of product architecture
Systems are classified into two types of product architectures: modular architectures and
integral architectures (Figure 1).
This typology reflects the dependence of the elements that constitute the system. The
system consists of many components. The system with simple and one-to-one mapping
from functions to components is called a modular system, while the system with complex
mapping and that has interdependence is called an integral system.

Figure 1 Definition of product architecture (see online version for colours)

Modular architecture Integral architecture

Safety &
Computation MPU Body
Stability
frame

Data Key Board Fuel Efficiency Engine

Input
Data Hard Drive Comfort Suspension
Store
Function Component Function Component

Ex) Personal Computer Ex) Passenger car

802 H. Tatsumoto and K. Ogawa

A typical example of an integral architecture is a passenger car. When you seek to

improve the fuel efficiency of a car, you need to lighten the body frame and to redesign
the suspension in order to meet your objective without sacrificing the functions of
comfort, safety and stability. However, in the case of a product built using modular
architecture, such as today’s PC, it is not necessary to redesign the keyboard or the
microprocessor to upgrade the hard drive from 10 GB to 100 GB.

2.2 Two types of technologies

In modular architecture, it is not necessary to finalise the design details of the overall
system in the early stages of the design process. Designers can mix and match modules to
vary the results while seeking the optimum outcome. Technologies in which designers
can individually vary designs and evaluate them with mix-and-match flexibility can be
defined as modular technologies.
In contrast, an integral architecture requires designers to plan the complete
and detailed design at an early stage. Technologies in which designers mutually organise
and adjust designs with coordinate-and-fit optimisation can be defined as integral
technologies.
In terms of technology diffusion speed, there is a considerable difference between
the two technologies because they require different background knowledge: codified
knowledge and tacit knowledge (Teece, 2006).
Modular technology uses codified interfaces specified by the modular architecture.
Designers can make choices within the limits set by the design rules without consulting
other designers. Knowledge is easily transferred among designers because it is clearly
codified and has rules in open environments (Baldwin and Clark, 2000; Teece, 2006). On
the other hand, integral technology is based on the premise of sharing tacit knowledge
among the designers. Tacit knowledge is hard to transfer to others because such
knowledge is difficult to articulate and demonstrate (Teece, 1986; Nonaka and Toyama,
2004; Teece, 2006).
Consequently, modular technology diffuses at very high speeds because it is easily
adopted by new organisations, while integral technology often remains in particular
organisations and diffuses at very low speeds.

2.3 The hierarchy of modular architecture

Ideally, modular architecture is supposed to be completely modular. However, actual
modular architecture products are not so. Strictly speaking, modular architecture products
should be a completely decomposable system. But they are actually often a nearly
decomposable system (Simon, 1996). In this case, the coordination of knowledge and
organisation cannot be achieved only by mix-and-match mechanisms developed using
design rules. Rather, interactive management of the participants and their activities is
required (Brusoni and Prencipe, 2001). This suggests that actual modular architecture
products consist of both modular and integral technology.
The concept of the hierarchy of architecture helps us to understand this complex
phenomenon (Alexander, 1964; Clark, 1985; Eppinger et al., 1990; Ulrich, 1995; Simon,
1996). Product architecture has a hierarchical structure and different technologies can
 The effect of technological platforms on the global supply chain 803

be applied on different levels. Completely modular architecture should use modular

technology on all levels, but actual modular architecture often consists of both modular
technology layers and integral technology layers.
For example, a DVD player shows that a modular architecture product consists of
layers with different technologies (Figure 2). Manufacturers mix and match modules such
as the optical pickup, spindle motor, and controller Large Scale Integration (LSI) circuit
to source better designs and to enable new DVD players to come to market promptly.
Thus, the system-level layer of a DVD player is a modular technology layer. But the
core components of a DVD player consist of integral technology layers. For example,
the optical pickup is made of lenses, beam splitters, detectors, lasers, magnets and
mechanical actuators, which require mutual adjustment in almost all design processes and
in the manufacturing process.

Figure 2 Hierarchy of architecture with two types of technologies

Product Levels of hierarchy Technology

DVD Player

System level:
Modular technology
Modular architecture

Controller Optical Spindle

LSI pickup motor

Component level:
Integral technology
Integral architecture

Lens Actuator Laser

2.4 The separation mechanism of platforms

As a consequence of hierarchical architecture, actual modular architecture has two
coordination mechanisms reflecting the two technologies. The first is the automatic
coordination mechanism with mix-and-match flexibility realised by sharing codified
interface information. The second is the interactive coordination mechanism that allows
for a sharing of tacit knowledge among designers. The second mechanism is usually
coordinated by special companies whose capabilities span a range of technological
fields that are wider than the range of activities that their clients actually perform
in-house. They are referred to as systems integrators (Brusoni and Prencipe, 2001) or
platform leaders (Gawer and Cusumano, 2002). Both are essentially the same in respect
of their coordination mechanism. The former utilises an activity-based interactive
804 H. Tatsumoto and K. Ogawa

coordination mechanism, whilst the latter achieves coordination inside of the module and
via the special module on the outside. In this paper, we will focus our attention on
platform leaders.
As we explained above, platforms are special components provided by companies
that have wider capabilities or knowledge than their clients (Gawer and Cusumano,
2002; Iansiti and Levien, 2004). In our definition, the platform is not equivalent to
other modules in a modular architecture product, because it includes the coordination
mechanism within it and converts the relevant levels of the product architecture from a
nearly decomposable system into a completely decomposable system. The conversion
clearly separates the technologies used in the product into modular and integral. This
conversion is defined as the separation mechanism.
The separation mechanism changes the actual modular architecture into a completely
modular architecture that allows modules to evolve freely within the limits set by design
rules and designers. This change radically increases the importance of the design rules,
and, in turn, leads to a powerful ecosystem of complementary products with increasing
mix-and-match flexibility. As a result, the separation mechanism sharply and steadily
accelerates modular cluster innovations based on the platform.

2.5 The effect of platforms on global supply chains

This acceleration of modular cluster innovations deeply affects the global supply chain of
modular architecture products.
The technology diffusion speed of actual modular architecture products is moderate
because the two technologies need a mixed and inseparable coordination. However, since
the platform splits the two technologies and converts an actual modular product into
a completely modular product at the system level with integral-technology layers at
the platform level, the technology diffusion speeds of the finished products and of the
platform become very different.
The technologies of the finished products rapidly diffuse into firms based in
developing countries because they use completely modular technology with a high
diffusion speed. On the other hand, the platform technology remains longer in the
original firm, which tends to be based in developed countries, because the platform
uses integral technology and has a low diffusion speed.
The separation mechanism therefore destroys the advantages of the original
manufacturers of the finished products; and new firms, which may be situated in
developing countries, may adapt the modular technology faster than the original firm.
Consequently, the separation mechanism may cause a production shift of the finished
products from developed countries to developing countries while keeping the platform in
a developed country.
We explore the validity of our framework through a brief summary of the competitive
and technical history of the PC industry. By drawing on a detailed case study of Intel’s
platform business in Taiwan, we try to explain the characteristics and influences of the
platform. The Taiwanese electronics industry had a big influence on the growth of the
Taiwanese economy after the 1990s. The cooperation between a US CPU vendor and
Taiwanese motherboard or notebook PC manufacturers is a typical platform success
story. This paper focuses on the collaborative industry-development process between
Intel and Taiwanese firms.
 The effect of technological platforms on the global supply chain 805

3 Method and research design

3.1 Method
Data on product architecture were gathered from various technical documents and
many discussions with PC developers. These data provided the necessary information,
including the interdependence of components and the mapping information from
functions to components. This type of data is useful in understanding the product
architecture of a PC.
In addition to these data, we gathered interview data and data on the time trends of
the price and shipment volume of PCs. The data included many interviews with US,
Taiwanese and Japanese PC industry players. The interviewees included the managers of
Intel’s design centre in Taiwan and in Japan, who used to be in charge of the platform
enabler, and Taiwanese motherboard and notebook PC manufacturers. Today, 90% of
motherboards and notebook PCs are produced in Taiwan, and Intel supplies both the
CPUs and the chipsets for these Taiwanese companies.

3.2 Research design

The research questions were broken down into two operational questions:
Research Question 1 What architectural change happens within the technological
platform in terms of the indicator ‘integral-modular’?
To examine the architectural change, we used a PC bus structure map that included
the main components. Electrical signals run through the bus, and each semiconductor
chip communicates and works with each other. The bus enables these components to link
to each other logically and functionally. Integral architecture requires much coordination
in the bus. On the other hand, PC architecture is regarded as modular when each item
works individually.
Research Question 2 What effect does the platform have on the industry in
developing countries?
It is assumed that the platform affects the global supply chain by encouraging the
production of finished products in developing countries. To examine this effect of the
platform, we focus on the growth of the Taiwanese PC industry. Intel started to supply
their platform in the 1990s. Then, the Taiwanese motherboard and notebook PC industry
expanded dramatically.
We used trend data to examine the relationships between the platform and Taiwanese
production expansion. The first data is the market-share trend of Intel’s chipset, which
shows the presence of Intel’s platform. The second set of data showed the trend in
Taiwanese production of motherboards and notebook PCs. Taiwanese firms produced
them using Intel’s platform. The positive correlation between two data can explain the
effect of the platform on the global supply chain.
806 H. Tatsumoto and K. Ogawa

4 Case study and findings

4.1 Industrial environment: the background to Intel’s platform strategy
Intel’s CPU business was facing a critical threat in the early 1990s. Some CPU
companies, such as Cyrix and AMD, started to compete with Intel in supplying
compatible CPUs. Moreover, the companies developing CPUs for minicomputers and
workstations were entering the PC CPU market.
Intel’s basic strategy was to supply more of the latest CPUs faster than its competitors
to obtain economies of scale and achieve market dominance. The quantities it supplied
allowed Intel to sell the latest CPUs at a low price, even though it had invested
large amounts in development and manufacturing. Moreover, its CPU products were
prevailing over its competitors, and then other firms provided software and hardware and
the complementary goods increased the value of Intel’s CPU. As a result, it achieved
market dominance.
This strategy was simple but difficult to execute in the early 1990s because the
development lead time for a CPU was much longer than it is today. Furthermore,
the speed of market penetration of the latest CPU was slow because traditional PC
manufacturers often delayed the adoption of the latest CPU since they tended to profit
more from existing CPUs. From the perspective of Intel, a rapid and extensive spread of
the latest CPU was required. However, Intel only supplied traditional PC manufacturers
at that time and needed to find new PC manufacturers. Therefore, it had to take a
new approach.

4.2 PCs’ architectural change by the platform

Intel needed to open new markets and gain new customers. It was targeting new ventures
and companies without a recognised brand. However, such new companies were
generally young and did not have the technology to handle Intel’s novel products.
Therefore, Intel planned to provide the CPU and chipsets simultaneously as the
platform. This new supply strategy brought about the architectural change shown
in Figure 3. Before platforming, PC included a CPU and three controllers (DRAM
controller, IO Bus controller and IDE controller). The CPU and controllers were supplied
by different specialised companies that did not coordinate the controller chips or adjust
them to Intel’s CPU. It took a significant amount of time to resolve the problem and PC
manufacturers had to make great efforts to adjust both the CPU and the chipsets. In the
meantime, Intel had to wait a long time for PCs with its latest CPU to come to market.
After platforming, the three controllers were integrated into two (North Bridge
and South Bridge) and fully fitted with the CPU. In addition, the two controllers were
designed to have a good connection with each other. The CPU and controllers became
what is known as the platform. As PCs became composed of the platform and many small
components, they changed over to the complete modular architecture.
In other words, a platform-based PC does not require much coordination because the
platform has already been adjusted with the chipset. Since the platform eliminates the
interdependence problem among components, the PC was converted from actual modular
architecture to completely modular architecture.
 The effect of technological platforms on the global supply chain 807

Figure 3 The platform of Intel (see online version for colours)

Intel has changed PC architecture

・Intel supplied both CPU
and chipsets (North Bridge and South Bridge) at the same time.
・PC drastically changed to be platform-based products.

Before After
Afterplatform
platform
Beforeplatform
platform
CPU Intel’s Platform
CPU Pentium
486

DRAM DRAM North

Controller Cache DRAM
Bridge
IO Bus
Controller PCI Bus

ISA bus
PCI
Device

IDE ISA USB

South
Controller Device Bridge
HDD PCI slot
HDD ISA Bus
ISA slot
・・・These chips were supplied by Intel.
ISA 5
Device
ISA slot

Figure 4 The trend of Intel’s chipset share (see online version for colours)

100%

90%

80%

70% Others
UMC
60%
OPTI
50% AcerLab
SIS
40% Intel’s Chipset Market Share VIA
30% INTEL

20%

10%
Intel’s platform completed.
0%
1993 1994 1995 1996 1997 1998 1999

Notes: Intel started to supply the chipset in 1993, and intensively got market share.
Intel’s chipset won the top position in the market in 1995.
Source: Dataquest and MIC (Chen, 2000)
808 H. Tatsumoto and K. Ogawa

It was in 1993 when Intel began to market its latest CPU (known as the Pentium CPU)
and the chipset as the platform to the market. Figure 4 demonstrates Intel’s intensive gain
of market share. The platform was completed in 1995, when Intel had the top share with
34% of the market.

4.3 The standardisation of the PC

Intel’s strategy did not stop at supplying the chipset. Intel standardised other components
in the PC, such as the internal bus, the size of components on the platform and various
interfaces of the platform connected to external peripherals.

Figure 5 The standardisation areas in the personal computer (see online version for colours)

1. The platform composed of

CPU and chipsets
CPU
PentiumⅢ
Standardized Interface
licensing
Intel’s Platform
Pin Layout
Closed Interface

AGP Chipset SDRAM

Licensing Interface

2. Making inside platform to Integral and Black box

Hub I/F
Black
AC’97 Box standardizing

Chipset PCI Bus

USB

HDD
3. Standardizing interfaces PCI slot
outside platform
Firmware
Hub

In many cases, Intel initiated standardisation alone, and in other cases they worked with
other companies. Consequently, most areas of the PC were standardised under the control
of Intel’s strategy and classified into three categories: the standardised interfaces, the
closed interfaces and the licensing interfaces (as shown in Figure 5). The standardised
interfaces are accessible and open to the public. The closed interfaces are limited for use
inside Intel. The licensing interfaces are also limited, but some specific firms licensed
by Intel can access them. Almost everything except the inside of the platform was
standardised. PC manufacturers could use various standardised devices, components and
peripherals and use only Intel’s chipset, because the interfaces of the peripherals were
standardised based on the chipset. The standardisation helped complementors prepare
their products for various customers via a single interface. It attracted a number of
complementors, and built the huge ecosystem by reducing the amount of coordination
required between finished product manufacturers and component vendors. In addition,
 The effect of technological platforms on the global supply chain 809

standardisation of the interfaces enabled the diffusion of the platform to accelerate

the openness of the PC. After its conversion into a completely modular product, any
company could produce a PC by purchasing the components and the platform. The focus
of competition in the PC industry changed totally and drastically.
Another effect of standardisation was increased consumer confidence. Consumers
assumed that standardisation guarantees function and interoperability and it encouraged
them to trust even unbranded PCs. However, product differentiation has become difficult
in terms of function and quality because of this standardisation.
A chipset with standardised interfaces rapidly promoted the conversion of the PC into
a completely modular product. Standardisation occurred in the notebook PC as well as
the desktop PC.

4.4 The expansion of Taiwanese production

The architectural change of the PC prompted many new companies to enter the market.
These new companies included many based in developing countries.
Figure 6 shows that Taiwanese production of motherboards and notebook PCs has
been increasing since 1995, the year when Intel completed the platform. This suggests
that Intel established the platform with the partnership of Taiwanese manufacturers and
that Intel’s chipsets contributed to the rise of the motherboard and notebook PC assembly
in Taiwan.

Figure 6 Intel’s platform works as a turnkey solution and has stimulated the Taiwanese PC
industry since 1995 (see online version for colours)

The production trend of Taiwanese MB/Note PC Turn-K ey-Solution

400,000,000

350,000,000 NotePC
PC) PC)

MB
300,000,000
Dollars (MB/Note

250,000,000 Intel has supplied chipset

K unit(MB/Note

and reference for motherboards

200,000,000
K New Taiwan

150,000,000

100,000,000 Intel’s platform completed.

Taiwanese production
50,000,000
growth
0
1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

year

Source: The internet information search system, Department of Statistics,

Ministry of Economic Affairs, ROC
810 H. Tatsumoto and K. Ogawa

In the early 1990s, Taiwanese PC manufacturers were already producing motherboards

or notebook PCs. However, their production was limited because they could not
manufacture the latest PCs that needed the latest CPU and chipset. These Taiwanese
manufacturers had no known brand, and the major US Original Equipment Manufacturers
(OEMs) outsourced PCs to them. However, the latest premium PC was not outsourced.
The US OEMs produced them by themselves. The Taiwanese manufacturers mainly
provided economic versions of motherboards and notebook PCs that were functionally
obsolete and not so expensive.
This exclusion from the latest premium market made it attractive for Taiwanese
companies to adopt new strategies to produce the latest motherboards. Their gross margin
ratio exceeded about 60% when they produced the latest premium motherboards for
the latest PC with Intel’s platform in 1995. Figure 6 demonstrates that Intel’s chipset
enabled them to enter this market and they extended their production quickly with
Intel’s platform.
Moreover, Intel offered not only the platform but also a turnkey solution, including
the chipset, evaluation kit, list of materials, reference design, design know-how and
manufacturing know-how. It even contained the data necessary for production. This
complete solution stimulated the Taiwanese PC industry and contributed to the economic
growth of Taiwan.
Taiwanese firms had great success in the field of motherboards, and they again
achieved it in the notebook PC market.
Formerly, traditional notebook PC firms in the developed countries had competitive
advantages because they had the entire system knowledge required for development
and manufacturing, including, for example, the heat design, and the power consumption
design controlled by the chipset. However, Taiwanese companies could not buy the
latest chipset for notebook PCs because the traditional firms were developing them
in-house. Therefore, it was difficult for Taiwanese manufacturers to enter the latest
notebook market.
Actually, Taiwanese firms had started the production of notebook PCs in the early
1990s, but their production was limited. They could produce only obsolete economic
versions of the notebook PC because they did not have enough knowledge of the entire
system. Their production share was only about 20% in 1995, although it was about 80%
in 2005.
The Taiwanese notebook PC industry faced a substantial change after Intel
energetically began to provide the chipset for the latest notebook PC. The chipset market
for the notebook computer was previously regarded as a niche market. However, Intel
began to supply a large amount of chipsets to this market. Their main target consumers
were Taiwanese manufacturers.
The chipset for the notebook PC was designed carefully with respect to power
consumption and heat radiation. Moreover, this chipset solved the battery-life problem
by standardising the power-saving control function, named Advanced Configuration
and Power Interface (ACPI). These problems need a whole system knowledge of the
notebook PC. However, after the chipset for notebook PCs was introduced, Taiwanese
manufacturers could develop and produce even the latest versions of notebook PCs
without system knowledge.
As the result, Taiwanese manufacturers expanded production quickly. The
technological platform helped the Taiwanese production. And in this process, Taiwanese
manufacturers contributed to the diffusion of the technological platform. Taiwanese
 The effect of technological platforms on the global supply chain 811

manufacturers invested ardently and expanded production on a large scale. In other

words, Intel established the platform with the partnership of Taiwanese manufacturers
and as production by Taiwanese manufacturers increased, Intel’s platform spread all over
the world.

4.5 The price erosion of the standardised components and the stability of
the platform
The technological platform converted the PC to a completely open modular product with
standardised components. In addition, the turnkey solution accelerated the expansion of
Taiwanese production. This drastic change affected the PC market.
Due to standardisation in almost all the interfaces, the severe price competition has
occurred in the PC parts markets. For example, the sharp price drop happened in hard
drives and DRAMs whose interfaces were fully standardised. Figure 7 demonstrates the
trend in the average sales price of key components of a PC. The price of hard drives and
DRAMs has continued to fall. However, Intel has kept CPU prices stable because Intel’s
platform has retained an integral architecture.

Figure 7 Trend in the average sales prices of key components of a PC (see online version
for colours)
変 [ 準]

120% Intel’s platform completed at 1995.

Intel has successfully kept CPU prices.
100% ASP of Intel’s CPU (unit price)

80%

60%

40%
ASP of HDD (unit price)
20%
ASP of DRAM (unit price)
Severe price erosion
0%
1995 1996 1997 1998 1999 2000 2001 2002 2003
Notes: ASP of HDD, DRAM has declined rapidly, but ASP of CPU has been stable.
Both HDD and DRAM are outside the platform, CPU is inside the platform.
Sources: All ASPs are originally calculated by the authors based on the data on
CPUs from Microprocessor design report, HDDs from Techno
system research, DRAMs from iSuppli and various other information

This double-sided market results from the architectural position. Intel’s platform has
retained an integral architecture. Moreover, Intel has made its platform into a black box.
On the other hand, the hard drive and DRAM have become modular in architecture
because they connect via a standardised open interface on Intel’s platform. The
standardisation accelerated the openness of the market. This means that the platform has
812 H. Tatsumoto and K. Ogawa

a double-sided characteristic: it has a closed integral architecture inside the platform and
an open modular architecture on the outside of the platform. Because of this characteristic
of the platform, the more the platform has diffused, the more the price of the standardised
components has fallen.
The average sales price of hard drives fell to about 40% in 2003. DRAM suffered
from more rapid price drops. In 2000, even though the memory capacity had been
increased ten times, the average sales price of a DRAM was about 40% lower than in
1995, the date the Intel platform was completed.
Figure 8 shows the trend of the DRAM price per megabit, instead of the unit price. In
1995, a severe price erosion occurred. The DRAM price per megabit was $3 in 1995 and
falling. Finally, it was $0.08 per megabit in 2000. The price had become about 1/30 in
five years.

Figure 8 Trend in DRAM price per megabit (see online version for colours)

5.00
4.50
4.00 Intel’s platform completed in 1995.
3.50
3.00
2.50
2.00 Serious Price Erosion
$3/Mbit
1.50
1.00 $0.08/Mbit
0.50
0.00
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Notes: After the platform was completed, a drastic fall in DRAM bit pieces happened.
The price dropped about 1/30 after only five years.
Source: iSuppli

The architectural conversion of the PC had a big influence on competition in the hard
drive and DRAM market. Seagate expanded its market share rapidly in the hard drive
market after 1995, and Samsung did the same in the DRAM market. The strategic change
happened in the standardised component market such as DRAM and hard drives, where
large-scale production and cost leadership became key success factors.
As prices for main components fell, the price of the final product, PC, also declined.
Moreover, product differentiation also decreased because almost all of the areas of the PC
were standardised and commoditisation increased.
The architectural conversion of the PC affected the strategy for competitiveness. The
companies that tried product differentiation by technology faced difficult situations. The
companies that focused on circulation channels gained more competitive advantage than
the ones relying on technology development.
 The effect of technological platforms on the global supply chain 813

Cheap PCs spread all over the world because of the conversion of the architecture.
Consumers who could not previously obtain a PC could now buy one. Therefore, the
complete modularisation enabled PCs to diffuse all over the world at the consumer level.

5 Discussion
5.1 Results from findings
The purpose of this study is to examine the technological characteristics of the platform
and to consider how the technological platform has affected the global supply chain.
For Research Question 1, we obtained two findings from the case study:
1 Intel started to supply CPUs and chipsets as a platform. The CPU and chipset
became a large chunk of elements that were coordinated and fully adjusted to each
other only by Intel. Intel’s platform is classified as an integral architecture.
2 However, the platform has fully standardised interfaces on the outside. Intel strongly
promoted the standardisation of interfaces of I/O components. As a result, many
complementors supply products to connect to the standardised interfaces.
The platform has a double-sided characteristic: it has a closed, integral architecture inside
the platform and a completely modular architecture on the outside. Because of the
double-sided characteristic of the platform, the more diffused the platform, the more the
PC became completely modular. Consequently, a PC became composed of one platform
and many small, standardised components. The standardised components have faced a
sharp price drop. On the other hand, Intel has successfully maintained the price of the
platform because the platform retains a black box inside with integral architecture.
For Research Question 2, we obtained two further findings from the case study:
1 The Taiwanese motherboard and notebook PC manufacturers have quickly expanded
their production following the introduction of the platform. Formerly, they could
access only limited markets, but the platform enables them to enter both the latest
premium market and the worldwide mass market.
2 Intel has provided not only the platform but also a turnkey solution. The turnkey
solution includes the technological platform, design know-how and useful
information for production. This complete solution stimulated the Taiwanese
PC industry and, subsequently, the Taiwanese economy.
The platform and complete turnkey solution affected the global supply chain for PCs. The
platform accelerated developing-country production of components, while production
of the platform itself remained in the developed country. The growth of developing
countries’ production has created a huge global market for PCs within a short time. From
Intel’s perspective, the finished products carried their integral products, such as the CPU,
to the worldwide market at a very high speed.
Figure 9 illustrates this mechanism and the effect on the global supply chain. The
platform decouples the technologies of the finished products into two technologies, one
with a low technology diffusion speed and the other with a high technology diffusion
speed. This decoupling strongly affects the global supply chain.
814 H. Tatsumoto and K. Ogawa

Figure 9 Changing technology diffusion speeds through the separation mechanism of the
platform (see online version for colours)

Changing technology diffusion speed by the platform The effect on global supply chain

Platform based finished products

Technology diffusion speed

・Developing country firms expand

d
production of finished products

ee
sp
rapidly.

n
・Many new market entrants make a

si o
Changes to the components

gy ar
ffu
of the platform diffuse at a huge global market for finished

lo ul
di
no od
slow pace but the platform products.

ch m
based finished products
te te
st le
diffuse at a fast pace.
fa mp
co

Traditional New economic

economic model collaboration model
Decoupled point by the platform Developed country firms
remain in finished products
market but supply only limited
domestic market.

The platform
complete integral
speed
slow technology diffusion ・Developed-country firms keeps
producing the platform.
Finished products ・Standardization reduces finished
without the platform products cost.
Years since delivery ・Advanced technology introduced
by the platform.

After the platform was established, firms in developing countries expanded their
production of finished products quickly and intensively. It was very attractive to use the
platform, because it enabled them to serve not only the obsolete-but-economical market
but also the advanced market. In addition, firms based in developing countries could offer
a much lower cost base than firms in developed countries.
Because of the dramatic cost reductions, PCs were reduced to affordable prices
for Brazil, Russia, India and China – the BRICs market, and a huge global PC market
has emerged, which would have been limited and small without the platform. The
platform noticeably enhances the new economic collaboration model between developed
and developing countries. Accordingly, the platform has resulted in a radical production
shift from developed countries to developing countries.

6 Conclusions

The platform made the architectural conversion of a PC into a completely modular

product possible by decoupling the two technologies in finished products into completely
modular and integral technologies. This architectural change, and the different diffusion
speeds of the two technologies, stimulated the PC industry in developing countries and
resulted in a new economic collaboration model.
The separation mechanism is derived from the double-sidedness of the platform.
The platform has a completely integral architecture inside and completely modular
architecture on the outside. The fast technology diffusion of the modular technology
meant that the platform accelerated the production by firms in developing countries. The
 The effect of technological platforms on the global supply chain 815

ensuing cost reductions made PCs affordable for large global markets, such as the BRICs
market. As a result, the platform strongly changed the global supply chain because it
noticeably enhanced the new economic collaboration model between firms in developed
and developing countries.
This model will become increasingly important in the future as it is extended into
other fields, such as cellular phones, DVD players, digital cameras, and even LCD panels
and photovoltaics. The growth of the industry in developing countries depends on the
use of platforms. However, at the same time, firms in developed countries that supply
the platform can access the huge global market in partnership with firms in developing
countries. This collaboration intensely affects the global supply chain. This study
examined the issue of the technological platform using the case study method. Before
further generalisations can be made, it is necessary to do more case studies in other
product segments or to use cross-sectional statistical data.

References
Abernathy, W.J. and Utterback, J.M. (1978) ‘Patterns of industrial innovation’, Technological
Review, pp.40–47.
Alexander, C. (1964) Notes of the Synthesis of Form, Cambridge, MA: Harvard University Press.
Baldwin, C.Y. and Clark, K.B. (2000) Design Rules: The Power of Modularity, Cambridge, MA:
MIT Press.
Berger, S. and the MIT Industrial Performance Center (2005) How We Compete: What Companies
Around the World Are Doing to Make It in Today’s Global Economy, New York:
Currency/Doubleday.
Brusoni, S. and Prencipe, A. (2001) ‘Unpacking the black box of modularity: technologies,
products and organizations’, Industrial and Corporate Change, Vol. 10, No. 1, pp.179–205.
Chen, W. (2000) ‘Shuanglong Qiangzhu de Jingpianzu Shichan (Competition between two dragons
in chipset market)’ [in Chinese], KGI Securities Research Report, Taiwan, KGI Securities
Co., Ltd.
Christensen, C.M. (1992) ‘Exploring the limits of the technology s-curve, Part 1: component
technologies’, Production and Operations Management Journal, Vol. 1, pp.334–357.
Clark, K.B. (1985) ‘The interaction of design hierarchies and market concepts in technological
evolution’, Research Policy, Vol. 14, No. 5, pp.235–251.
Eppinger, S.D., Whitney, D.E., Smith, R.P. and Gebala, D. (1990) ‘Organizing the tasks
in complex design projects’, ASME Conference on Design Theory and Methodology,
Chicago, Illinois, September, pp.39–46.
Fujimoto, T. (2007) ‘Architecture-based comparative advantage – a design information view of
manufacturing’, Eol. Inst. Econ. Rev, Vol. 4, No. 1, pp.55–112.
Gawer, A. and Cusumano, M.A. (2002) Platform Leadership: How Intel, Microsoft, and Cisco
Drive Industry Innovation, Boston, MA: Harvard Business School Press.
Henderson, R.M. and Clark, K.B. (1990) ‘Architectural innovation – the configuration of existing
product technologies and the failure of established firms’, Administrative Science Quarterly,
Vol. 35, No. 1, pp.9–30.
Iansiti, M. and Levien, R. (2004) The Keystone Advantage: What the New Business Ecosystems
Mean for Strategy, Innovation, and Sustainability, Boston: Harvard Business School Press.
Langlois, N.L. and Robertson, P.L. (1992) ‘Networks and innovation in a modular system: lessons
from the microcomputer and stereo component industries’, Research Policy, Vol. 21, No. 4,
pp.297–313.
816 H. Tatsumoto and K. Ogawa

Nonaka, I. and Toyama, R. (2004) ‘Knowledge creation as a synthesizing process’, in H. Takeuchi

and I. Nonaka (Eds.) Hitotsubashi on Knowledge Management, John Wiley.
Rogers, E.M. (2003) Diffusion of Innovations, 5th ed., Free Press.
Sanchez, R. and Mahoney, J. (1996) ‘Modularity, flexibility, and knowledge management in
product and organization design’, Strategic Management Journal, Vol. 17, Winter special
issue, pp.63–76.
Shapiro, C. and Varian, H.R. (1998) Information Rules: A Strategic Guide to the Network
Economy, Harvard Business School Press.
Shintaku, J., Ogawa, K. and Yoshimoto, T. (2006) ‘Architecture-based approaches to international
standardization and evolution of business models’, MMRC Discussion paper, University of
Tokyo, No. 96.
Simon, H.A. (1996) The Sciences of the Artificial, 3rd ed., MIT Press.
Sturgeon, T.J. (2002) ‘Modular production networks: a new american model of industrial
organization’, Industrial and Corporate Change, Vol. 11, No. 3, pp.451–496.
Teece, D.J. (1986) ‘Profiting from technological innovation – implications for integration,
collaboration, licensing and public policy’, Research Policy, Vol. 15, No. 6, pp.285–305.
Teece, D.J. (2006) ‘Reflections on “profiting from innovation”’, Research Policy, Vol. 35, No. 8,
pp.1131–1146.
Ulrich, K. (1995) ‘The role of product architecture in the manufacturing firm’, Research Policy,
Vol. 24, No. 3, pp.419–440.