Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Abstract
Innovation has frequently been categorized as either radical, incremental, architectural, modular or niche, based
on the effects which it has on the competence, other products, and investment decisions of the innovating entity.
Often, however, an innovation which is, say, architectural at the innovator/manufacturer level, may turn out to be
radical to customers, incremental to suppliers of components and equipment, and something else to suppliers of
critical complementary innovations. These various faces of one innovation at different stages of the innovation
value-added chain are what we call the hypercube of innovation. For many high-technology products, a technology
strategy that neglects these various faces of an innovation and dwells only on the effects of the innovation at the
innovator/manufacturer level can have disastrous effects. This is especially so for innovations whose success
depends on complementary innovations, whose use involves learning and where positive network externalities exist
at the customer level. We describe the hypercube of innovation model and use it to examine RISC (Reduced
Instruction Set Computers) and CISC (Complex Instruction Set Computers) semiconductor chips, and supercomput-
ers, and suggest how firms can better manage the relationships along the innovation value-added chain using the
model. The model forces innovation managers to think in terms of their customers, suppliers and complementary
innovators.
something else to suppliers of critical comple- 'competence enhancing' depending on what the
mentary innovations; all of which have implica- innovation did to the knowledge base of the inno-
tions for the success of the innovation. These vating entity.
various faces of one innovation at different stages Roberts and Berry (1985) prescribed how an
of the innovation value-added chain are what we innovating entity can enter a new business de-
call the hypercube of innovation. pending on its familiarity with the technology and
Schumpeter himself described innovation as "a market, and the newness of the market and tech-
historic and irreversible change in the way of nology to the innovating entity. These business
doing things" and "creative destruction" (Schum- entry options range from acquiring other firms
peter, 1947). Abernathy and Utterback (1978) with the technology and market competence or
found that as a technology evolves, product inno- performing the R & D internally for the familiar,
vation gives way to process innovation, making it to venture capital investments for the unfamiliar
difficult for the innovating entity to revert to new and new. The familiarity of the technology to
product innovations; that is, the competence of members of the innovation value-added chain
the innovating entity is effectively destroyed. Us- was not investigated.
ing the automobile industry, Abernathy and Clark Given the nature of some of the industries
(1985) grouped innovations into four categories studied by these authors, the conclusions they
depending on the impact of the innovation on the arrived at vis-?t-vis what the innovating entity
innovating firm's capabilities and knowledge of should do should not differ much from the con-
its technology or market. They did not address clusions that would be arrived at from an analysis
the impact of each of the innovations on the that uses the hypercube model. However, an ex-
capabilities and assets of the suppliers of compo- amination of the effects of an innovation cannot
nents, customers, and suppliers of complemen- be limited to the impact on the capabilities, com-
tary products. Using extensive data from the pho- petence and assets of the innovating entity for
tolithography industry, Henderson and Clark industries where at least one of the following is
(1990) classified innovations according to whether true: complementary innovations are critical to
the innovation overturned the existing knowledge the diffusion and success of products; learning by
of core concepts and components, and the link- customers is critical, expensive and often results
ages between them. They classified an innovation in lock-in; positive network externalities at cus-
as radical if the core concepts of the innovation tomers are common and equipment and critical
as well as the linkages between them overturned components (that go into the innovation) from
existing ones; architectural if the core concepts suppliers can be innovations in their own right.
were being reinforced while the linkages between An analysis must also look at the impact of the
these core concepts and components of the prod- innovation on the capabilities of suppliers of
uct were changed; incremental if the core con- components, customers, and complementary in-
cepts were reinforced while the linkages between novators.
them were unchanged; modular if the core con- Stated differently, studies that have catego-
cepts were overturned while the linkages between rized innovation have the innovating entity asking
the concepts were unchanged; radical if the core the question: 'What is the impact of this innova-
concepts and linkages between them are over- tion on my organizational capabilities, compe-
turned. As was the case with the Abernathy and tence, existing products, knowledge of compo-
Clark analysis of the automobile industry, the nents, key concepts and linkages between them'.
impact of the innovations on the capabilities and In the hypercube of innovation approach we are
assets of suppliers, customers and suppliers of suggesting that in addition to probing what the
complementary products was not considered. In innovation will do to its competence and assets,
their study of the US cement, and minicomputer the innovating entity must-also ask the question:
industries, Tushman and Anderson (1986) classi- 'What will my innovation do to the competence
fied innovations as 'competence destroying' or and products of my suppliers, OEM (original
A.N. Afuah, N. Bahram / Research Policy 24 (1995) 51-76 53
I
Unchanged I o ,ion : ovatiou I / I /
Linkages between core
concept~& components
. . . . . . . :. . . . . . . q / /I
IArchitectural ~ Radical / / ~ /
Changed IInnovation ' Innovation/ 1,~ Innovator f
[. . . . . . . . . . . : | J Innovation Value-Added Chain
/ i ~ Supplier
YT - x Reinforced Overturned
C o r e Concepts
Fig. 1. The hypercube of innovation. The X and Y axes are the innovation-classifying factors. The Z-axis is the innovation
value-adding chain of supplier of key components, innovator, customer and supplier of complementary innovators.
54 A.N. Afuah, N. Bahram / Research Policy 24 (1995) 51-76
complementary innovators. It depicts relation- sional space being determined by the 'intensity'
ships that are multidimensional in nature. In par- of the innovation along each of these dimensions,
ticular, the hypercube is a four-dimensional cube where intensity is a measure of how radical the
with each of the stages of the innovation value- innovation is, using an ordinal scale, say, of incre-
added chain representing a dimension, and the mental = 1, modular = 2, architectural = 3, and
location of any innovation in this four-dimen- radical = 4, with intensity increasing from incre-
I o-.°,al ', I I
Supplier
Unchanged
Llnkagee betweenc o r e
coaeepb & component3
Changed
r_°:7__ :7::11
i
I
Incremental t Modular
Unchanged Innovation , Innovation
I Reinforced Ovemuned
Llnlmg~betweencore ....... I ......
t'ott~pk & ~ponents Core Com~pts
Architectural, Radical
Changed Innovation I! Innovation
I
Reinforced Ove~arned
Core Concepts
Innovator
Supplier
I
Incremental , Modular
Unchanged Innovation , Innovation
Linkages between c o r e I
....... I ......
concepts & components Architectural I Radical
Changed Innovation ' Innovation
I
t !!
Reinforced Overturned
Core Concepts
Customer /
Incremental i Modular ]
Unchanged Innovation j Innovation [
Llokqm tmtwt~ core ...... II......
concept8& eomponentJ
Changed
•c•mental
Innovator /
Reinforced Ovemmaed
Core Concepts
Fig. 2. The hypercube of innovation exploded to show the various faces that an innovation can assume along the innovation
value-added chain.
A.N. Afuah, N. Bahram / Research Policy 24 (1995) 51-76 55
Table 1
The effects of the innovation on the innovating entity
The innovating entity
Asset or activity Possible impact
Core concepts Enhances or makes obsolete core concepts from previous innovations
Linkages between core concepts and components Enhances or destroys previous knowledge of linkages
Product components Remain the same or change
Competence Enhances or destroys other competencies (skills and knowledge
from previous innovation)
Existing products Enhances use of previous products or cannibalizes them
Complementary innovations from previous products Can use or not use complementary innovations
Institutional support Can receive or not receive any government or other institutional
research subsidies
Table 2
The effects on customers
Customer
Asset or activity Possible impact
Learning Enhances or destroys skills and knowledge acquired from previous product
Built-up assets Enhances or destroys use of assets built around previous innovations
Network externalities Enhances or destroys positive network externalities
Complementary innovations from previous products Can use or not use complementary innovations from older products
Product design Enhances or makes obsolete previous design
Design knowledge Enhances or destroys previous design knowledge
Product components Remain the same or change
Table 3
The effects of the innovation on complementary innovators
Complementary innovators
Asset or activity Possible impact (range - - best to worst)
Inertia of old complementary products Keeps up with inertia of old complementary innovations
Momentum of new products Keeps up with the momentum of new complementary innovations
Product design Enhances or makes obsolete previous design (of complementary product)
Design knowledge Enhances or destroys previous design and manufacturing knowledge
of complementary product
Product components Remain the same or change
Competence Enhances or destroys skills and knowledge from previous products
Existing products Enhances use of previous products
Positive network externalities Enhances or destroys positive network externalities for complementary
products
Complementary innovations from previous products Can use or not use complementary innovations from older products
56 A.N. Afuah, N. Bahram / Research Policy 24 (1995) 51-76
chain, viz. suppliers of critical components that tal, if the core concepts are reinforced while the
go into the innovation, customers (OEM, and linkages between them are unchanged; modular,
end-user), and suppliers of complementary inno- if the core concepts are overturned while the
vations. Fig. 2 further explodes these stages of linkages between the concepts are unchanged.
the chain for better visualization. We emphasize As detailed in Table 1, the innovating entity
the fact that the innovating entity can use any has to recognize and take the necessary correc-
criteria for categorizing innovations in the X and tive action depending on whether the innovation
Y axes. makes obsolete or enhances previous designs, de-
Tables 1-4 list the range of possible impact of stroys or enhances knowledge gained in previous
an innovation at the innovating entity, suppliers, designs, cannibalizes older products, or can be
customers, and complementary innovators, and used with previous complementary products.
we briefly describe what is possible at each stage.
2.2. Customers
2.1. The innouating entity
The impact of an innovation on the capabili-
The focus of most innovation literature has ties and assets of the innovator's customers has
been on the impact of an innovation on the very important implications for the market suc-
capabilities and assets of its innovator. For the cess of the innovator. Unfortunately, most inno-
innovator, the primary concern has been the im- vation studies have focused on the impact of the
pact of the innovation on its organizational com- technology on the innovator's knowledge of tech-
petence - whether it enhances or destroys it nology and market, while ignoring the impact on
(Abernathy and Utterback, 1978; Tushman and customers' capabilities and assets. There are at
Anderson, 1986); on the core concepts and link- least four areas where the impact of an innova-
ages between those core concepts of the product tion on a customer can have serious effects:
(Henderson and Clark, 1990); on existing innova- learning, positive network externalities, compati-
tions; and on the willingness of management to bility with complementary or old products and
invest in the innovation (Henderson, 1993; Rein- continued use of old products.
ganum, 1983, 1984). For the hypercube model,
any categorization framework can be used. For 2.2.1. Learning
example, we use the Henderson and Clark (1990) Many complex high-technology products re-
model and classify an innovation as radical, if the quire that users invest time and money in learn-
core concepts of the innovation as well as the ing how to operate and maintain the products.
linkages between them have overturned existing An innovation that destroys the knowledge that
ones; architectural, if the core concepts are being the customer has acquired has a smaller chance
reinforced while the linkages between these core of being adopted than one that enhances this
concepts of the product are changed; incremen- knowledge and skills. Thus we expect a person
Table 4
The effects of the innovation on suppliers of key components or equipment
Suppliers of components and equipment
Asset or activity Possible impact (range - - best to worst)
Component a n d / o r equipment design Enhances or makes obsolete previous design of component or
equipment supplied for previous innovation
Design knowledge of components a n d / o r equipment Enhances or destroys previous design and manufacturing knowledge
of components or equipment supplier for previous innovation
Competence Enhances or destroys skills and knowledge used to supply components
or equipment for previous innovation
Old products Enhances or destroys use of previous components or equipment
A.N. A fuah, N. Bahram / Research Policy 24 (1995) 51-76 57
who buys a computer and learns the computer's products can still be used with the new innova-
operating system to be less willing to buy another tions, and (4) built-up assets will not have to be
computer with a different operating system than destroyed.
one with the same operating system; unless there
is another program that can make the new oper- 2.3. Complementary innovators
ating system transparent to the customer.
The huge success of personal computers since
2.2.2. Positive network externalities their introduction in the late 1970s would not be
A product or skill is said to possess positive as phenomenal were it not for complementary
network externalities if the value of the product innovations like spreadsheet and word-processing
to an owner increases as more people own it. software. Innovators not only have to watch out
Positive network externality has its origins from for the inertia of older complementary innova-
the telephone network where one's telephone is tions and the momentum of newer ones, but may
more valuable the more people are connected to also have to cooperate (via, e.g. strategic al-
one's network. The more friends you have that liances) with the complementary innovators to
own a computer that is compatible with yours, produce complementary innovations (see the case
the more valuable your computer is to you be- of IBM and Intel's microprocessor later).
cause you can Share software and innovative ways
of using the computer. An innovation that de- 2.4. Suppliers of components and equipment
stroys this positive network externality does not
stand a good chance of being adopted by cus- Some high technology product innovations (e.g.
tomers. aircraft and supercomputers) depend heavily on
component and equipment innovations from their
2.2.3. Compatibility with complementary products suppliers. The aircraft cannot move into super-
Using the computer example again, a personal sonic flight without the right innovations in en-
computer user who invested in a Lotus 123 gine technology. In supercomputers, most of the
spreadsheet would prefer not to switch to a new gains that we have seen in computer performance
computer that requires him to buy a new spread- have come from innovations in the semiconductor
sheet. chips that go into them.
¢= 0
2"
Supplier
-
Innovation
L/
i • i •
Value-added Chain
i
GREEN ZONE
Electric Car
the car companies, to suppliers of key compo- Fig. 5. The g r e e n - r e d zone map for the electric car.
60 A.N. Afuah, N. Bahram ~Research Policy 24 (1995) 51-76
be avoided except where there are obvious price/ these two extremes is the yellow zone. Any inno-
performance advantages for the customer, where vation whose map passes through this zone should
the innovator is entering new markets where cus- be pursued with a lot of precautions. The innova-
tomers have not yet had time to build any innova- tor should monitor the inertia of older comple-
tion-specific skills and knowledge, and compe- mentary innovations and the momentum of newer
tence destruction is not an issue, where comple- ones, to take advantage of them.
mentary innovations, that allow customers to keep Finally, the innovating entity should perform
their competence and positive network externali- the relevant cost-benefit analysis for each level
ties exist, and when institutional requirements of the innovation value-added chain, taking into
mandate the innovation. Somewhere between consideration not only the cost of learning, net-
Complementary
.DSKKeyboard I
Unchanged Ji
Lialmgesbetween eo~ I
concepts & compoaeats - "1I . . . . . .
Supplier ~n
exl~ang-- " i .OSI2
i .Electricc~
t
.DSKKeyboard i
Unchanged .ElectricalC~
Linkagm between eor~ Reinforced Overturned
cotceF4|& components CoreCoaeep~
Changed
Reinforced Overtumed
CoreCoaeepts
Cus.__tome~
Unchanged
Ltakag~betw~-acore
cm~ep~&compo~ots
77n_ i_7°::7_
Changed
.~in~o,~..,r~'..~c~ I I f'
Reinforced Overturned
Core Concepts
iCustomer
Unchanged
Unlmgell~s~a core
eoacepts & o a m ~ l ~ t i-O~S-/~
¢-~- J .DSKKeybo~d
Chaaged
Ilmovator /
Reinforced Overturned
Core Concepts
Fig. 7. Exploded version of the hypercube of innovation. OS/2, windows, DSK and the electric car at the various levels of the
innovation value-added chain.
62 A.N. Afuah, N. Bahram / Research Policy 24 (1995) 51-76
work externalities, additional capital investments, cessor, or innovate again. Intel chose the latter
and cannibalization of old products, but also of and designed the 8088 with the 16-bit internal
such additional expenses as marketing and adver- architecture of the 8086, and the external archi-
tising for the new technology. tecture of 8-bit processors so that the 8088 could
We are now ready to apply the model to RISC, use the readily available and inexpensive 8-bit
CISC and supercomputers. complementary chips. This allowed Intel's cus-
tomers, i.e. system builders like IBM, to take
advantage of the advanced third generation fea-
3. The hypercube: The cases of RISC, CISC, and tures that the internal architecture provides while
supercomputers also using the inexpensive, more readily available
second generation complementary chips.
3.1. R I S C and CISC chips When IBM decided to enter the personal com-
puter market, and had to choose a microproces-
3.1.1. CISC chips sor for its PC, it chose the Intel 8088, although
In 1970 Intel Corporation invented the first this processor may not have been superior to
microprocessor, a microchip implementation of other microprocessors, specifically the Motorola
the central processing unit (CPU) of a computer. 68000. The 8088's 8-bit interface that allowed PC
Subsequently, Intel and its competitors like Mo- manufacturers to use readily available and inex-
torola introduced successive generations of 8-bit, pensive complementary chips may have tilted the
16-bit and 32-bit microprocessors that over time balance in Intel's favor.
got faster, consumed less power, and delivered
higher functionality. The instruction sets for these
processors - the commands which programmers 3.1.2. R I S C chips
use to tell the processors what to do - grew to be By the mid-1970s industry researchers and aca-
very large, with each instruction taking too long demics had begun to question the efficiency of
to execute, and earning these processors the name the CISC approach. Alternative approaches to
complex instruction set computers (CISC). remove the performance and cost disadvantages
These relatively complicated instructions re- inherent in the CISC approach were examined.
quired ever increasing chip sizes, cost more, and In 1975, IBM's Thomas J. Watson Research Cen-
placed a limit on how fast a given operation could ter's researchers began the development of the
be executed, limiting the microprocessor's speed. IBM 801 computer. Although not a microproces-
Microprocessors differ from other chips in that sor, this computer laid out the foundations of the
they have instruction sets, and development sys- RISC approach to microprocessor design. These
tems/ software have to be written to support the IBM researchers moved away from a large num-
instruction sets. They also require a host of com- ber of complicated instructions to a small number
patible complementary chips to allow them to of very simple instructions. This change signifi-
interface with devices like printers, modems or cantly enhanced the speed of the processor with-
keyboard, and chips that control disk drives. out severely impacting the ease of use and flexi-
These complementary chips and development bility that was advertised as a key advantage of
systems can also be supplied by complementary CISC processors.
innovators. The simpler and fewer instructions meant
When Intel designed its third generation mi- smaller chip sizes and faster RISC processors.
croprocessor (16-bit), the 8086, it discovered that Successive generations of RISC processors could
a lot of the less expensive complementary chips in be designed faster than same-generation CISC
existence then were for the previous generation processors providing RISC with a time-to-market
of microprocessors, i.e. 8-bit microprocessors. In- (T-I'M) advantage. These factors would lead to
tel had two choices: wait until the complementary large RISC price/performance ratio advantages
chips catch up with its third generation micropro- over same-generation CISC processors.
A.N. Afuah, N. Bahram / Research Policy 24 (1995) 51-76 63
By 1982, Patterson and co-workers at Berkeley products ranging from Desktop to Data Center
(Patterson and Ditzel, 1980; Patterson and Se- minis and mainframes.
quin, 1981), and Hennessey and co-workers at Hewlett Packard has also moved into adopting
Stanford (Gill et al., 1983) had implemented the RISC for its hardware platforms as well. The
RISC concepts in a single VLSI (very large scale PA-RISC architecture is now available across the
integrated) circuit. The RISC1 processor at range of Hewlett-Packard's hardware platforms.
Berkeley, and the MIPS processor at Stanford In the next two sections the hypercube of
laid the foundations of the two most successful innovation is used to examine innovations in RISC
RISC architectures in the industry today: SUN and CISC microprocessors. Fig. 8 provides the
Microsystem's SPARC RISC processor and MIPS context for RISC and CISC use.
Corporation's family of Rxxxx (R2000, R3000,
R4000) processors. 3.2. The hypercube model and CISC
Despite pursuing the RISC principles first,
IBM has only recently developed a relatively suc- In Fig. 9 we have classified some CISC proces-
cessful VLSI RISC processor of its own, the sor innovations according to whether they were
RS6000, and plans on co-developing the Power incremental or radical innovations, and below, we
series of RISC processors with Motorola and look at the effects of some of these innovations at
Apple Computer. the various stages of the innovation value-added
Digital Equipment Corporation at first used chain.
RISC chips from the MIPS family of processors Integrating the CPU of a computer on a single
to power its DECStation workstation products, chip, the microprocessor, in 1970 was a radical
but has recently introduced its 64-bit Alpha mi- innovation to Intel. The design methodologies
croprocessor which will be used for Digital's were not only different, semiconductor fabrica-
CI User/Software Developer I CI
CI ApplicationSoftware [ [ HardwarePackage ] CI
CI OpemtmgSy~em [ CI
CI Compilers [ [ SystemArchitecture I CI
MI MicroprocessorArchitecture II Complementary
ICs I CI
MI ChipTopLevelArchitecture I
MI LogicDesign [
SI Sermconductor FabricationProcess
I
SI SemiconductorPhysics&DeviceModels
Fig. 8. Computer system knowledge areas and their providers. Key: CI, complementary innovator; MI, microprocessor innovator
(primary innovator); SI, supplier innovator.
64 A.N. Afuah, N. Bahram / Research Policy 24 (1995) 51-76
tion capability had to be pushed to its limits to computer-aided design (CAD) tools were radical.
achieve the desired integration levels. The de- To OEM customers, this was also a radical inno-
mands put on suppliers of critical equipment like vation since they had to learn the instruction sets
Complementary
.uVAX, u370 !
.80xxx family 8
Unchanged . . 6 ~ family ',
L l n k a l ~ l~twt.~ t o r t
concepts & components I .I'~-PA
I ~$6000
/ Supplier Changed I .8~6
I .68000
.All mtcr~ of 2dlP$
I
~ae ~me I I ~'ARC
Reinfomed Ovemmaed
Core C ~ o n p t s
Supplier Innovator
I
Incremental ! Meddler
Unchanged Innovation ! Innovation
I
L l n l m l ~ between core I. . . . . . .
¢oneepls & components
Architectural ! Radical
Changed Innovation I Innovation
I
I
Reinforced Overturned
Core Concepts
Customer
|
/
.uVAX, u370 ! .HP-PA
.80xxx family ! .R~000
Unchanged .~xxx f~ally I
.Alpha !
Linkages between core . . . . . . I. . . . . . .
conc~Is & components 1.8086
J .68000
Innovator / Changed
! .M~S
i ~PARC
I
/
.uVAX, u370 J
.80xn family II
Unchanged .68x7~ flm0dly !
Reinforced Overturned
Llnlml~; between core Core Concepts
¢euceptl & ~m~nponents ! .I~-PA -Alpha
I .RS6~00
Changed ! .8086
I .68000
.MIPS
I .SPARC
/
Rehffe~ced Overturned
Core CoKepts
Fig. 9. Exploded version of the hypercube of innovation for CISC technologies used in m i n i c o m p u t e r s / m a i n f r a m e s vs. RISC-based
workstations.
A.N. Afuah, N. Bahrain/Research Policy 24 (1995) 51-76 65
of these processors, etc. For complementary in- To providers of compilers, the first generation of
novators, this was a radical innovation since they RISC chips was a radical innovation. To develop-
had to develop new compatible complementary ers of development systems, RISC was also a
chips for the microprocessors while also learning radical innovation since they now had to learn
the instruction sets of the processors. They had to the new instruction sets of RISC, and also had to
develop systems/software to support the micro- find ways to accommodate RISC's higher speeds
processor's instruction set. and inherent parallelism. Complementary inte-
Design and implementation of subsequent grated circuits (ICs) were hard pressed to keep
generations in the same microprocessor family, up with the RISC microprocessor speed. Chips
for example Intel's 80xxx family of CISC proces- that decouple other slower parts of the system
sors, are for the most part incremental innova- from the faster microprocessors were developed.
tions to the various members of the innovation Most of these innovations rely on rearranging the
value-added chain, partly as a result of the fact linkages between existing building blocks, and as
that microprocessors were designed to be upward such, most of the innovations to date have been
compatible with previous generations, allowing architectural in nature.
programs that were written for old processors to Once again, most of these innovations were
work on new systems. Thus programs written for adopted by CISC processors and have become
Intel's 16-bit microprocessors could run on 32-bit standard design practice for microprocessor-based
processors. Complementary innovators like Mi- system design.
crosoft who wrote the DOS operating system For the most part, the silicon fabrication tech-
used on 16-bit machines did not have to worry nology and CAD tools that suppliers provided
about writing a new operating system because CISC makers can be used by RISC makers too.
DOS could be used for the 32-bit machines. This does not mean that there have been no
radical or architectural changes in these fields.
3.3. The hypercube model and R I S C There have been many. The point is that these
innovations have been independent of RISC and
Fig. 9 also illustrates the hypercube applied to CISC architectures. The impact of these architec-
various RISC processors from key innovators. tures has been to create incremental innovation
RISC processors utilize the same building blocks in the supplier base.
as CISC processors, but the way these blocks are For RISC end-users, the picture has been very
put together is different for each processor type. different and varied. They have in general
As such, RISC compared to CISC is an architec- adopted the UNIX operating system, making
tural innovation (Henderson and Clark, 1990) for RISC a radical innovation for those who had
the innovating entity. used other operating systems like DOS. But the
For many complementary innovators, the first advantages of a non-proprietary, very low cost,
generation of complementary innovations in sup- high performance and portable operating system
port of RISC were for the most part radical since coupled with the superior price/performance of
in many cases, many changes were required in RISC microprocessors has, for most applications,
products used with CISC microprocessors. Com- made up for the loss of competence in switching
pilers used by RISC offer a good example. Fun- operating systems. To the end-user, porting appli-
damental to RISC hardware simplicity is more cations software from non-UNIX operating sys-
sophisticated compiler technology that must tems to UNIX is also costly.
translate 'friendly' programming languages such For relatively new markets like workstations or
as rORTr~N or COBOLinto a reduced and simpli- embedded control, where there were no en-
fied set of microprocessor instructions that the trenched operating systems and where price/
RISC processor can understand. The availability performance is critical, RISC has been accepted
of such optimized compilers is often cited as one relatively fast.
of the reasons why the RISC dream was realized. For markets like personal computers where
66 A.N. Afuah, N. Bahram / Research Policy 24 (1995) 51-76
end-users have not only invested heavily in learn- IBM's RS6000, have so far been mostly incre-
ing and writing their own applications software in mental innovations for all members of the inno-
DOS-based systems, but also built up positive vation value-added chain.
network externalities, RISC is a lot more of a
radical innovation than for the workstation or
embedded control market. Thus RISC has had 4. S u p e r c o m p u t e r s
great difficulties dislodging CISC. The weak ap-
propriability (Teece, 1986) of RISC concepts has Supercomputers are generally described as the
helped CISC designs to close the gap in p r i c e / most powerful computational systems available at
performance. Table 5 shows the hypercube in any given time. This would mean that the first
tabular form, listing all CISC and RISC innova- supercomputer dates back to Charles Babbage's
tions. mid-1800s 'analytical engine'. Most of today's in-
Later generations of RISC, like MIPS Corpo- stalled base of supercomputers, however, can be
ration's R3000, and R4000 or later versions of attributed to Seymour Cray who, in 1975, left
Table 5
The hypercube in tabular form -- CISC and RISC innovations
Innovator Customer
Incremental Incremental
• uVAX, u370 (CISC) • uVAX,u370
• 80xxxfamily(CISC)• 68x (CISC)
• 80xxxfamily
(CISC)
• 68xxx family
(CISC)
• Alpha
Radical Radical
• 8086(CISC) • 8086(CISC)
• 68000(CISC) • 68000(CISC)
• MIPS R2000(RISC) • MIPSR2000(RISC)
• SPARCI(RISC) • SPARCI(RISC)
• HP-PA (RISC) • HP-PA (RISC)
• RS6000(RISC) • RS6000(RISC)
• ALPHA(RISC)
Supplier Complementary innovator
Incremental Incremental
All micros of the • uVAX,u370
same generation (CISC)
• 80xxxfamily
(CISC)
• 68xxxfamily
(CISC) Radical
Radical • Alpha
• 8086(CISC)
• 68000(CISC)
• MIPSR2000(RISC)
• SPARCI(RISC)
• HP-PA (RISC)
• RS6000(RISC)
A.N. Afuah, N. Bahram /Research Policy 24 (1995) 51-76 67
Table 6
Some supercomputer product innovations
Machine Year Manufacturer Bits CPU Technology Operating system
IBM700 1954 IBM 36 Sequential Vacuum T
IBM7000 1959 IBM 36 Sequential Transistor
CDC6600 1965 CDC 60 Sequential
CDC7600 1969 CDC 60 Scalar Transistor
Star-100 1972 CDC 64 Vector processor ICs
ILLIAC IV 1972 Burroughs 64 64 Process ICs
Cray-1 1976 CDC 64 Vector/scalar LSI
CDC 205 1976 CDC 64 Vector LSI
Univac LARC 1960 Univac
IBM7030 1960 IBM
IBM360/195 1971 IBM
TIASC 1974 TI 32 Vector LSI
Denelcor HEP-1 1977
Cyber205 1981 CDC 64 Vector LSI
Hitachi $810 20 1983 Vector/scalar LSI
Fujitsu VP Vector
ETA-30
Cray X-MP 1982 Cray 64 Vector 1st multiprocessor
Denelcor HEP-1 Denelcor Multiprocessor
ETA- 10 ETA Systems Vector, 8
processors
Cray-2 1985 Cray 64 4 CPUs
Cray Y-MP 1988 Cray 8-16 CPUs, UNICOS, COS,
vector CTSS
Cray C-90 1991 Cray Vector, 16 ECL UNICOS, COS,
processors CTSS
Cray-3 GaAs UNICOS, COS
IBM3090/600S VF 1988 IBM Vector, 1-6 MVS, AIX,
VM/CMS
Fujitsu VP-2600/20 1991 Fujitsu Vector Proprietary OS,
UTS/M
Fujitsu FACon VP-2000 1984 Fujitsu ECL
NEC SX-2 1985 NEC ECL
Hitachi $820/80 1988 Hitachi Vector ECL Proprietary OS,
HIUX
NEC SX-3 1992 Multiprocessor, UNIX
vector
Control Data Corporation (CDC) where he had Neumann architecture t. In 1976, Cray Research
designed supercomputers, to start his own super- shipped its first supercomputer, the Cray-1, the
computer company, Cray Research Inc. At CDC, first commercially available vector supercom-
Cray had designed the CDC 7600 supercomputer, puter. Vector computers, for the most part, need
a so-called scalar supercomputer because it had a
scalar processor (the 'engine' or brain of the
computer). Scalar processors have to issue an i The architecture used in most of today's computers is often
instruction for every single operation (e.g. addi- attributed to John Von Neumann's mid-1940s architecture. In
tion of two numbers) so that even vector data that architecture, the CPU of the computer fetches an instruc-
tion (data) from a central store (main memory), operates on it
would have to be broken down and an instruction
(for example, add or subtract), and returns the results into the
issued for operation on each element of the vec- main memory. Only one CPU is used, and that one CPU can
tor. The 7600 was also of the traditional Von do only one thing at a time.
68 A.N. Afuah, N. Bahram / Research Policy 24 (1995) 51-76
only one instruction to execute each operation on through the computer's electrical circuitry at the
vectors, and this greatly improves processing time speed of light, and no matter how much these
(for applications that lend themselves to lists) computers with 1-16 processors speed up each
compared to scalar processors. Vector processing processor, they would never attain some of the
was a key innovation in supercomputers, espe- speeds that many compute-intensive jobs need
cially since a lot of data on which supercomputers (for example, supercomputers still cannot synthe-
operate are either vector-like or could be vector- size a protein from its gene) because of the
ized. The first vector supercomputer was actually physical limit imposed by the speed of light. This
the CDC Star-100 but was not commercially is where massively parallel computers (MPC)
available until after the Cray-1, when it was re- come in.
leased as the Cyber 205. In massively parallel computers, hundreds or
One thing which the Cray-1, CDC7600, Cyber thousands of processors are put on one job, with
205, and previous supercomputers (vector or each processor simultaneously tackling an as-
scalar) had in common was that they each had signed stage of the job to get the whole job done
only one processor that could be put on any one faster than one processor operating sequentially
processing job at any one time. Cray Research - the structure of the job permitting. So rather
changed all that in 1982 when it introduced its than trying to speed up one or a few processors
multiprocessor Cray X-MP, the first commercially to do the job, MPCs put very many processors on
successful supercomputer to apply more than one the job to perform it in parallel. Now, the speed
processor to the same problem at any one time of light is no longer the physical limit, and execu-
(the ILLIAC IV, developed at the University of tion of inherently parallel jobs can be speeded up
Illinois, was the first parallel supercomputer). In considerably. Thinking Machines' CM5 uses hun-
the years that followed, Cray Research intro- dreds of 32-bit SPARC CMOS (complementary
duced many other multiprocessor supercomput- metal oxide semiconductor) microprocessors and
ers with the Cray Y-MP C90 its latest with 16 runs at 128 GFLOPS peak. The physical limit to
processors in 1991 that delivers 16 GFLOPS the speed of MPCs will eventually be the ability
(gigaFLOPS = billion floating point operations of the processors to communicate with each other.
per second) compared to the Cray-l's 100 MPCs use readily available CMOS (a proven
MFLOPS (million floating point operations per technology) chips that consume less power than
second). In 1992, NEC introduced its 4-processor the ECL chips used in conventional (Cray-like)
SX3 that gives 25 GFLOPS. Table 6 lists some of supercomputers, and these CMOS chips do not
the key supercomputers that have been intro- have to be as fast as ECL chips since it is not the
duced over the years. speed of each one that matters (at this stage of
Most of the gains in supercomputer perfor- the technology) in MPCs but their combination.
mance have come as a result of innovations in And because they consume less power, they are
semiconductor technology, from the transistor to air-cooled and do not need the elaborate liquid
VLSI circuits. NEC's four-processor supercom- cooling systems of ECL-based systems.
puter, for example, was able to deliver the 25 MPCs can be divided into two groups: multi-
GFLOP primarily because of its advanced ECL processors and multicomputers. Multiprocessor
(emitter-coupled logic) semiconductor technology MPCs like Kendal Square Research's KSR 1 have
and premier packaging techniques. numerous processors that share one memory
A key goal of these traditional Cray supercom- bank. The KSR 1 has 1088 64-bit microprocessors
puter designs that use few (1-16) processors is to that share the same memory bank. Multicom-
make each processor as fast as possible. But puter MPCs are interconnected microprocessors,
despite all the dramatic improvements in mi- each with its own memory, that communicate via
crochip and packaging technology, these kinds of message passing. Examples are Thinking Ma-
supercomputer designs are reaching a physical chines CM5, Intel's Paragon, and supercomputer
limit - the speed of light. Computer signals travel MPCs from Ncube, Ametek, and Transputer.
A.N. Afuah, N. Bahram / Research Policy 24 (1995) 51-76 69
Most of the manufacturers of traditional su- called minisupercomputers. They utilize the same
percomputers (those with 16 or fewer very fast vector processing of traditional supercomputers,
processors, and elaborate cooling systems) like but with some important differences. They are
Cray Research Inc., IBM, etc. have either already cheaper, provide 25-35% the performance of tra-
started MPC programs or announced that they ditional supercomputers (Kelley, 1988), offer
will do so. But their quest to improve traditional lower price for the performance provided and
supercomputers has not stopped. When, in 1989, lend themselves to those low-end applications
Seymour Cray left Cray Research to start Cray that do not need the higher performance of higher
Computer, his answer to getting a faster super- power supercomputers, let alone their prices.
computer was to use gallium arsenide (GaAs) They use proven CMOS chips that are less ex-
chips which can be two and half times as fast as pensive and consume less power than the power-
conventional silicon chips and also consume a lot demanding but faster ECL chips used in tradi-
less power. Gallium arsenide is a relatively new tional designs. This results in cheaper systems
technology that is still in its infancy compared to that are air-cooled.
the silicon semiconductor technology that now
provides chips for computers. The introduction of 4.1. The hypercube model and supercomputers
the Cray-3 has been delayed primarily because of
the difficulties in getting GaAs chips to work. In this section we use the hypercube of innova-
Supercomputer Systems Inc. (SSI), another su- tion model to examine the supercomputer indus-
percomputer start-up, is also having difficulties try that we have just described. However, this is
delivering its first supercomputer because it was not a comprehensive treatment of innovations in
banking on GaAs chips. supercomputers.
Another viable set of computers are the so- Tables 7 and 8 list key supercomputer innova-
Table 7
Key innovations in supercomputers
Innovation Machine Year Firm
Vector processing Star-100 1973 a CDC
Vector processing Cray-I 1976 Cray Research
Multiprocessing (traditional) Cray X-MP 1982 Cray Research
MPC SIMD CM-2 1986 Thinking Machines
MPC multiprocessor KR 1 1992 Kendal Square Research
MPC multicomputer Intel Paragon 198x Intel Corp
Minisupercomputers Convex-2 198x Convex Computers
TIASC 1974 TI
Denelcor HEP-1 1977 Denelcor
Cyber205 1981 CDC
Hitachi $810 20 1983 Hitachi
Fujitsu VP Fujitsu
Cray-2 1985 Cray
Cray Y-MP 1988 Cray
Cray C-90 1991 Cray
Cray-3
IBM3090/600S VF 1988 IBM
Fujitsu VP-2600/20 1991 Fujitsu
Fujitsu FACon VP-200 1984 Fujitsu
NEC SX-2 1985 NEC
Hitachi $820/80 1988 Hitachi
NEC SX-3 1992 NEC
a Designed in 1969 but became operational in 1973. Never shipped. Features maintained in the Cyber 205 (1982).
70 A.N. Afuah, N. Bahrain/Research Policy 24 (1995) 51-76
Table 8
Classifications of some of Cray Research's innovations
Year Product Innovation Operating system Microchip Cray Research
Technology
1976 Cray-1 Vector processing COS (Cray Op System) ECL Radical
innovation
1979 Cray-1/S COS
1982 Cray-1/M COS MOS memory
1982 Cray X-MP Multiprocessing COS
1985 Cray-2 4-CPUs UNICOS
1988 Cray Y-MP 8-CPUs
1990 Cray Y-MP 2E Air/water-cooled
1991 Cray Y-MP 8E
1991 Cray Y-MP UNICOS
199x Cray-3
Table 9
The hypercube in tabular form - - supercomputer innovations
Innovator Customer
Incremental Incremental
• Cray Y-MP 2E • Cray X-MP
• Cray Y-MP 8E • Cray-2
• Convex-2
• Cray Y-MP C90
• Cray-1 Radical Radical
• Star-100 • Illiac IV • Illiac IV
• Cray X-MP • CM-2 • Cray-1
• Cray-2 • Paragon • Star-100
• Convex-2 • KR1 • CM-2
• Cray Y-MP C90 • CM-5 • CM-5
• Cray-3 • Paragon
• SSI • KR 1
Supplier Complementary innovator
Incremental Incremental
• Illiac IV • Cray X-MP
• Cray-1 • Cray-2
• Star-100 • Convex-2
• Cray X-MP • Cray Y-MP C90
• Cray-2
• Convex-2
• Cray Y-MP C90
• CM-2
• CM-5
• Paragon
• KR 1
Radical Radical
• Cray-1 • Illiac IV
• SSI • Cray-1
• Star-100
• CM-2
• CM-5
• Paragon
*KR1
A.N. Afuah, N. Bahram / Research Policy 24 (1995) 51-76 71
tions, while their impact on the capabilities and convert some of the old software written for
assets of their innovators, suppliers, customers, scalar machines to forms in which vector ma-
and complementary innovators is shown in Tables chines could crunch. So the impact of the radical-
9 and 10 and Fig. 10. ness of the innovation on customers was not that
Using the Henderson and Clark categorization important. Cray Research's multiprocessor, Cray
criteria, vector processing was an architectural X-MP, can also be considered an architectural
innovation to CDC (Star-100) and Cray Research innovation for reasons similar to those just listed
(Cray-1) when they designed these systems. The above. For customers, its impact was more incre-
main components of the supercomputer-mem- mental than radical since it still used the same
ory, CPU, input/output ( I / O ) - and core design Cray operating system (COS).
concepts had not changed radically; the key The Cray-3 and SSI's machine are examples of
change was the provision of vector processing. machines that are radical to suppliers and facing
But the linkages between these components and problems because of it. Both machines are multi-
core concepts were being altered. To many cus- processor but with no more than 16 processors
tomers and suppliers of applications software, and not radically different from previous designs.
however, this was a radical innovation because They are, however, depending on GaAs chips to
they had to learn how to program with vector make major contributions to the planned speed
processors. Luckily, for the Cray, most users of improvements. But GaAs technology is still in its
supercomputers then were scientists who wrote infancy compared to the proven silicon technol-
their own software and were more interested in a ogy that other computers use and is thus a radical
number-crunching engine than a complete data innovation to any computer. Cray Computer's
processing solution. This relatively small segment solution to reducing this uncertainty was to ac-
of the market would remain small for this very quire Gigabit Logic, a GaAs chip manufacturer.
reason. More importantly, a complementary inno- That has still not worked. While the problems
vation, the vectorizer was developed that could with the Cray-3 and SSI's machine may not be
Table 10
The hypercube in tabular form - - Cray Research Inc.
Innovator Customer
Incremental Incremental
• Cray Y-MP 2E • Cray X-MP
• Cray Y-MP 8E • Cray 2
• Cray Y-MP C90
• Cray XMS
• Cray-1 Radical Architectural Radical
• Cray X-MP • Cray-1
• Cray-2
• Cray Y-MP C90
• Cray XMS
Supplier Complementary innovator
Incremental Incremental
• Cray-1 • Cray X-MP
• Cray X-Mp • Cray-2
• Cray-2 • Cray Y-MP C90
• Cray Y-MP C90 • Cray XMS
• Cray XMS
Radical Radical
• Cray-1 • Cray-1
72 A.N. Afuah, N. Bahram / Research Policy 24 (1995) 51-76
Complementary /
Innovator /
.Cray-3 I
.SSI !
Unchanged .Convex !
Linkages between core . . . . . . I. . . . . .
concepts & components FCmy-1
l.,S)ar-I it0
Suppfier Changed tCM-2
.Cray-1 ICM-5
.Star-100 t
.Cmy X-MP I
Unchanged .Convex-2 I
.CM-2 I Reinforced Overturned
Llaluqlee betweon core .~5 . . . . !. . . . . .
~ c e p t s & components .Paragon I .Cray-3 Cote Concepts
.KI~1 ! .SSI
Changed I
I
!
Reinforced Overturned
Core Concepts
Innovator /
Supplier
I
Incremental i Modular
Unchanged Innovation I Innovation
I
L h t k q m I~twteo t o r t . . . . . . . I. . . . . . .
concepts & components I
Architectural j Radical
Changed Innovation i Innovation
I
I
Reinforced Overturned
Core Concepts
!
Customer /
.Cmy X-MP !
~SI I
Unchanged .Convex-2 I
Cray-3 I
Linkages between core . . . . . . |. . . . . .
¢o~p~ & ¢ompo~n~ i .Cray-1
I .SUtr-I00
Core Concepts
entirely due to GaAs chips, it is true that GaAs, a MPCs are a radical innovation for all members
radical innovation to most suppliers of chips, has of the innovation value-added chain except sup-
contributed to the problems of the two machines. pliers. Their design is conceptually very different
A.N. Afuah, N. Bahrain/Research Policy 24 (1995) 51-76 73
from that of traditional supercomputer designs. impact of the innovation will be on the innovating
Writing software for them is even trickier. Users entity's capabilities and assets, but also on those
of the installed base of traditional Cray-type de- of suppliers, customers, and complementary inno-
signs would prefer machines that allow them to vators. We suggest that innovators should think
keep some of the skills and knowledge acquired twice about innovations that destroy skills, knowl-
with the Cray-like machines, and especially any edge and positive network externalities at any of
applications programs that they may have written. the stages of the value-added chain, especially at
Their operating systems are also different. Appli- the customer level. They should avoid the red
cations, as well as systems programmers for the zone (of the mapping of innovations along the
new machines, are also not easy to find. Hardcore innovation value-added chain) and go with inno-
supercomputer users (scientists and academics) vations that reinforce key concepts and linkages
can write their own software. But for MPCs to all along the value-added chain (innovations that
diffuse into general purpose applications that will fall in the green or yellow zone).
greatly increase their success, they need lots of We also note some criteria for innovating in
software. In particular, MPCs need to be pro- the red zone. Specifically, we suggest that the red
grammable in existing programming languages zone should be avoided unless a subset of the
like FORTRAN, C and c ÷+. This would mean that following is true. (1) The p r i c e / p e r f o r m a n c e ra-
the end-user only sees the change in speed. tio of the innovation, as viewed by all the levels of
the value-added chain especially the customer,
outweighs any losses incurred as a result of com-
5. Summary and implications of the hypercube petence or positive network externality destruc-
tion. This happens, for example, when the physi-
Using several examples, we have shown that an cal limit of an older technological trajectory has
innovating entity that only looks at the impact of been reached and the only way to overcome this
its innovation on its competence and existing physical limitation is to move to a new technolog-
products, and does not critically examine the ical trajectory - - a move that often means de-
impact of that innovation on the competence and struction of competence acquired during the evo-
capabilities of its suppliers, customers, competi- lution along the older trajectory but great im-
tors and complementary innovators, may be mak- provement in some key parameter. (2) New mar-
ing a mistake. Dvorak's DSK keyboard failed to kets where customers have not yet had time to
diffuse because it was an architectural innovation build any innovation-specific skills and knowl-
to Dvorak but a radical innovation to its cus- edge, and competence destruction is not an issue.
tomers. O S / 2 was a radical innovation to IBM (3) Complementary innovations, that allow cus-
but an incremental innovation to DOS users ver- tomers (or other members of the innovation
sus Microsoft Windows which was an incremental value-added chain) to keep their competence and
innovation to Microsoft and an incremental inno- positive network externalities exist. (4) When in-
vation to users of DOS. This allowed Microsoft to stitutional requirements mandate the innovation.
enter the market early and so far, Microsoft We analyzed RISC and CISC chips, and su-
Windows is winning. Similarly, Lotus, that did not percomputers using the model. In particular, we
pay attention to the momentum of Windows soft- analyzed the impact of key innovations in CISC
ware, lost some ground in its spreadsheet market chips, RISC chips, and supercomputers on the
share to Microsoft's Excel. The case of the elec- capabilities of suppliers, customers, and comple-
tric car - which is a radical innovation to the mentary innovators. In C1SC, we suggested that
innovating firms, suppliers of components and Intel's foresight in designing the 8088 micropro-
complementary products, but an incremental in- cessor in response to the inertia of complemen-
novation to users - was also discussed. tary 8-bit chips may have contributed to its being
The model forces innovation managers to look chosen by IBM over competitors to provide the
at their innovations not only in terms of what the microprocessor architecture for the now very
74 A.N. Afuah, N. Bahram / Research Policy 24 (1995) 51-76
popular IBM PC and PC compatibles. This may adopted R I S C / U N I X technology instead of the
have been Intel's most important decision ever. mostly CISC/proprietary operating system solu-
We also found that although RISC is an archi- tions of the past. There are still, however, many
tectural innovation as far as chipmakers like Mo- manufacturers of proprietary systems.
torola and Intel are concerned, it is a radical In supercomputers, Cray Computer Corp. and
innovation to OEM customers who have been SSI are having difficulties introducing their new
using CISC chips to design personal computers GaAs chip-based supercomputers partly because
and sell to end-users. This is because with RISC, GaAs is a radical innovation to chip suppliers
these OEMs have to learn new assembly lan- relative to the mature silicon technology.
guages, establish new development systems, and Earlier versions of supercomputer innovations
retrain their engineers on how to design systems that were radical innovations to customers did
with the RISC chips. To personal computer end- not have the disastrous consequences predicted
users who have learned to use DOS, acquired or by the hypercube model because many of those
written their own applications programs, and es- early users were scientists and academics who
tablished positive network externalities on CISC- wrote their own programs, and could trade the
based machines, RISC is a radical innovation program writing for a more powerful computing
since in its present form, it destroys the compe- engine.
tence, capabilities and positive network externali- Massively parallel computers, despite being
ties of these customers. The promise of speed faster than the traditional Cray-like supercom-
alone is not enough to dislodge CISC in this puters may not be diffusing as fast as one would
particular market. For computer systems under expect because they are a radical innovation not
$25 000, annual sales of Intel's CISC chips alone only to the innovating entities but also to cus-
exceeded 20 million units while all RISC chips tomers and suppliers of complementary innova-
combined had sales of only 339 000 units in 1991 tions like software. It is, however, an incremental
(Business Week, 30 March, 1992) 2. We also sug- innovation for suppliers of hardware components
gest that it may be the realization of the inertia of like microchips and disk drives.
CISC vis-h-vis RISC that made firms like Com- The real breakthrough in supercomputer diffu-
paq pull out of the ACE consortium. All that sion will come when the parallel machines pene-
could change if a complementary innovation (e.g. trate the general purpose business applications
software) could be developed that allows all DOS that could use their compute power. This will
users to preserve their skills and old applications come only if the software is there, which in turn,
software when they use RISC machines. Mi- can only be developed if the current parallel
crosoft NT is intended to be this innovation. machines can be programmed with existing lan-
In newer markets like workstations, where the guages such as FORTRAN, C, C++, etc.
capabilities, competencies, and positive network
externalities have not been well-established yet,
RISC is doing very well. In the embedded control 6. Conclusion
market where speed is critical and the end-user is
not locked into CISC as in the PC market, RISC The common practice of classifying innova-
is also doing well. tions only according to the impact of the innova-
In the minicomputer and mainframe markets tion on the innovating entity's capabilities vis-a-vis
the price/performance advantages of RISC have its existing technology and markets is not ade-
been sufficiently compelling that manufacturers quate for high technology products that require
and customers of these classes of computers have critical input components and equipment from
suppliers, depend on complementary innovations
for success, require high levels of learning by
2 Our thanks to an anonymous referee who suggested this customers before use, and that lend themselves to
example. positive network externalities. For such products,
A.N. Afuah, N. Bahram / Research Policy 24 (1995) 51-76 75
the impact of the innovation on the capabilities Els, S.A., 1989, Configuration and competitiveness in the
and assets of suppliers, customers and comple- supercomputer Industry, Master's Thesis, Massachusetts
Institute of Technology.
mentary innovators may be just as critical as that Farrell, J. and G. Saloner, 1985, Standardization, compatibil-
on the innovating entity's competence and assets. ity and innovation. Rand Journal of Economics 16(1) 70-
The hypercube model forces managers at the 83.
innovating entity to evaluate their innovations in Fernbach, S., 1984, Supercomputers - - past, present,
terms of the impact of those innovations on the prospects, Future Generation Computer Systems, July.
Foster, R.A., 1986, Innovation: The attacker's advantage.
competence and assets of all the members of the (Summit Books, New York).
innovation value-added chain. Our examination Freeman, C., 1986, The economics of industrial innovation,
of the CISC, RISC and supercomputer industries 2nd edn. (MIT Press, Cambridge, MA).
suggests that the innovator should pursue innova- Gill, J., T. Gross, J. Hennessy, N. Jouppi, S. Przybylski and C.
tions that reinforce core concepts and compe- Rowen, 1983, Summary of MIPS instructions, Stanford
University Technical Note No. CSL-83-237.
tence along the innovation value-added chain, Henderson, R., 1993, Underinvestment and incompetence as
while being more cautious with those that do not. responses to radical innovation: Evidence from the pho-
The innovator should watch out for the inertia of tolithographic alignment equipment industry, RAND
older complementary innovations and the mo- Journal of Economics 24(2).
mentum of newer ones, and take advantage of Henderson, R. and K. Clark, 1990, Architectural innovation:
The reconfiguration of existing product technologies and
them. the failure of established firms, Administrative Sciences
Quarterly 35, 9-30.
6.1. Areas of future research Hewlett-Packard, 1986, Precision architecture and instruction
reference manual (Hewlett-Packard).
We have developed the hypercube model using Kane, G., 1987, MIPS R2000 RISC architecture (Prentice
Hall, Englewood Cliffs, NJ).
qualitative data. An area of further research Kartz, M.L. and C. Shapiro, 1985, Network externalities,
would be the collection of quantitative data to competition and compatibility, American Economic Re-
gain further insights to this concept. view 75(3), 424-440.
The second and equally important area of re- Kartz, M.L. and C. Shapiro, 1986, Technology adoption in the
search would be the extension of the hypercube presence of network externalities. Journal of Political
Economy 94(4), 822-841.
to include multi-innovator/competitor scenarios. Katevenis, M.G.H., 1985, Reduced instruction set computer
Such an expanded model could then be applied architectures for VLSI (MIT Press, Cambridge, MA).
to innovation-based international competition. Lunde, A., 1977, Empirical evaluation of some features of
instruction set processor architecture, Communications of
the ACM 20(3), 143-153.
Patterson, D.A. and D.R. Ditzel, 1980, The case for the
References reduced instruction set computer, Computer Architecture
News 8(6), 25-33.
Abernathy, W.J. and K. Clark, 1985, Mapping the winds of Patterson, D. and C. Sequin, 1981, RISC I: A reduced instruc-
change, Research Policy 14, 3-22. tion set VLSI computer, Proceedings of the 8th Annual
Abernathy, W.J. and J.M. Utterback, 1978, Patterns of inno- Symposium on Computer Architecture (ACM SIGARCH
vation in technology, Technology Review 80(7), 40-47. CAN 9.3) 443-457.
Afuah, A.N. and J.M. Utterback, 1991, The emergence of a Pratt. G.A., 1992, Electric cars: on the road again, Technology
new supercomputer architecture, Technology Forecasting Review, August/September.
and Social Change 40, 315-328. Radin, G., 1983, The 801 minicomputer, IBM Journal of
David, P.A., 1985, A contribution to the theory of innovation, Research and Development 27(3), 237-246.
Memorandum No. 71, Stanford Center for Research in Reinganum, J., 1983, Uncertain innovation and the persis-
Economic Growth, Stanford University. tence of monopoly, American Economic Review 73, 741-
David, P.A., 1985, Clio and the economics of QWERTY, 748.
American Economic Review: Papers and Proceedings Reinganum, J., 1984, Uncertain innovation and the persis-
75(2), 332-336. tence of monopoly: Reply, American Economic Review
Dobbs, C., P. Reed and T. Ng, 1988, Supercomputing on a 74, 243-245.
chip - - genesis of the 88000, VLSI System Design, May, Roberts, E. and C.A. Berry, 1985, Entering new businesses,
24-33. Sloan Management Review, Spring, 3-17.
76 A.N. Afuah, N. Bahrain/Research Policy 24 (1995) 51-76
Schumpeter, J.A., 1947, The creative response in economic tinuities and organizational environments, Administrative
history, Journal of Economic History 7, 149-159. Reprinted Science Quarterly, 31,439-465.
in: R.V. Clemence (Editor), Essays of J.A. Schumpeter, Utterback, J.M. and L. Kim, 1986, Invasion of a stable busi-
(Addison-Wesley, Cambridge, MA, 1951). ness by radical innovation, in: P.R. Kleindorfer (Editor),
Teece, D.J., 1986, Profiting from technological innovation: The management of productivity and technology in manu-
implications for integration, collaboration, licensing and facturing (Plenum, New York).
public policy, Research Policy, 285-305. Von Hippel, E., 1988, The sources of innovation (Oxford
Tushman, M.L. and P. Anderson, 1986, Technological discon- University Press, New York).