Stanford University Press

Stanford, California

© 2013 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved.

No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying and recording, or in any information storage or retrieval system without the prior written permission of Stanford University Press.

Special discounts for bulk quantities of Stanford Business Books are available to corporations, professional associations, and other organizations. For details and discount information, contact the special sales department of Stanford University Press. Tel: (650) 736-1782, Fax: (650) 736-1784

Printed in the United States of America on acid-free, archival-quality paper

Library of Congress Cataloging-in-Publication Data

Funk, Jeffrey L., author.

Technology change and the rise of new industries / Jeffrey L. Funk.

pages cm. — (Innovation and technology in the world economy)

Includes bibliographical references and index.

ISBN 978-0-8047-8385-9 (cloth : alk. paper)

ISBN 978-0-8047-8492-4 (e-book)

1. Technological innovations. 2. Industries. 3. Technology—Economic aspects. I. Title. II. Series: Innovation and technology in the world economy.

HC79.T4F86 2012

338′.064—dc23

2012014298

Typeset by Newgen in 10/12.5 Electra

Technology Change and the Rise of New Industries

JEFFREY L. FUNK

Stanford Business Books

An Imprint of Stanford University Press

Stanford, California

ADVANCE PRAISE FOR: Technology Change and the Rise of New Industries

In this important book, Jeff Funk examines what it will take to realize the potential of new technologies for innovating out of the economic challenges that we face. He argues that many theories of innovation are incomplete, outdated, or just plain wrong, and that new insights are sorely needed. The basic message is that we have to think of disruption and discontinuity as positive, productive forces that must be embraced.

—Anita M. McGahan, University of Toronto and Author of How Industries Evolve

Jeff Funk is one of the few scholars in the field of technological change who has tried to see the technology system as a whole. This book should be required reading for academics and government decision makers who need to acquaint themselves with this view.

—Robert U. Ayres, INSEAD and International Institute for Applied Systems Analysis

Without resorting to singular case studies, Funk takes a sophisticated approach to characterizing technological emergence and change, and the role that governments play in developmental trajectories. He builds a descriptive model using a historical approach to reinterpret data from a host of industries. Based on this model, the book takes a daring step to speculate on future technological developments in energy and electronics, providing sobering advice to those who think that government intervention is the panacea for national innovation.

—Phillip Phan, The Johns Hopkins Carey Business School

In this vitally important advance in the analysis of innovation, Funk explores the limits of learning curves as a mere function of forced or subsidized volumes. Learning has to be real, integrated, and multifaceted in order to benefit from different paths to improvement in cost and performance. Trenchantly demonstrating the need for multidimensional, supply-side innovation in the case of clean energy, he shows the futility of our current demand-side focus.

—George Gilder, Venture Capitalist and Author of the forthcoming Knowledge and Power: The Information Theory of Capitalism

Jeff Funk’s provocative elaboration on Giovanni Dosi’s notion of technology paradigms calls for a fundamental reexamination of conventional management wisdom about technologies and technology evolutions. In particular, Funk’s clear exposition of the supply-side technology dynamics that drive disruptive innovations provides a long overdue corrective to the demand-side story that has been widely advanced by Clayton Christensen, for example. More generally, Funk’s framework for analyzing and predicting future technology trajectories establishes a new and essential perspective for both technology strategists and technology policymakers.

—Ron Sanchez, Copenhagen Business School

Explaining much about innovation that others have ignored, Funk helps us to better understand how improvements in costs and performance occur with new technologies. While the conventional wisdom suggests that costs fall as cumulative production increases, Funk shows us that the reality of this relationship is different and more interesting. For example, technologies that benefit from reductions in scale (e.g., integrated circuits) have seen dramatic advances; finding these kinds of technologies (and products based on them) is a major task for R&D managers.

—Christopher L. Magee, Center for Innovation in Product Development, Massachusetts Institute of Technology

INNOVATION and TECHNOLOGY in the WORLD ECONOMY

Editor

MART IN KENNEY

University of California, Davis/Berkeley

Round Table on the International Economy

Other titles in the series:

The Evolution of a New Industry: A Genealogical Approach

ISRAEL DRORI, SHMUEL ELLIS, AND ZUR SHAPIRA

The Science of Science Policy: A Handbook

KAYE HUSBANDS FEALING, JULIA I. LANE, JOHN H. MARBURGER III, AND STEPHANIE S. SHIPP, EDS.

Restoring the Innovative Edge: Driving the Evolution of Science and Technology

JERALD HAGE

The Challenge of Remaining Innovative: Insights from Twentieth-Century American Business

SALLY H. CLARKE, NAOMI R. LAMOREAUX, AND STEVEN W. USSELMAN, EDS.

How Revolutionary Was the Revolution? National Responses and Global Technology in the Digital Era

JOHN ZYSMAN AND ABRAHAM NEWMAN, EDS.

Global Broadband Battles: Why the U.S. and Europe Lag Behind While Asia Leads

MART IN FRANSMAN, ED.

Ivory Tower and Industrial Innovation: University-Industry Technology Transfer Before and After the Bayh-Dole Act in the United States

DAVID C. MOWERY, RICHARD P. NELSON, BHAVEN N. SAMPAT, AND ARVIDS A. SIEDONIS

To Yvonne

CONTENTS

List of Illustrations

Acknowledgments

1. Introduction

PART I. WHAT DETERMINES THE POTENTIAL FOR NEW TECHNOLOGIES AND THUS NEW INDUSTRIES?

2. Technology Paradigm

3. Geometrical Scaling

PART II. WHEN DO TECHNOLOGICAL DISCONTINUITIES EMERGE?

4. Computers

5. Audio and Video Recording and Playback Equipment

6. Semiconductors

PART III. OPPORTUNITIES AND CHALLENGES FOR FIRMS AND GOVERNMENTS

7. Competition in New Industries

8. Different Industries, Different Challenges

PART IV. THINKING ABOUT THE FUTURE

9. Electronics and Electronic Systems

10. Clean Energy

11. Conclusions

Appendix: Research Methodology

Notes

References

Index

ILLUSTRATIONS

FIGURES

2.1   Improvements in maximum efficiency of engines and turbines

2.2   Improvements in energy storage density

2.3   Improvements in luminosity per watt for various technologies

2.4   Improvements in luminosity per watt for LEDs and OLEDs

2.5   Improvements in MIPS (million instructions per second) per price level (in U.S. dollars)

2.6   Improvements in data rates

3.1   Maximum scale of engines and turbines

3.2   Examples of geometrical scaling with engines

4.1   Minimum thresholds of performance in computer components

4.2   Innovation frontiers for computers

5.1   Innovation frontiers for audio and video equipment (quality versus size)

5.2   Innovation frontiers for audio and video equipment (quality versus price)

5.3   Minimum thresholds of performance for magnetic tape in systems

6.1   Innovation frontier for the early years of the semiconductor industry

6.2   Minimum thresholds of component/equipment performance for IC discontinuities

6.3   Minimum thresholds of equipment performance in terms of number of transistors per chip

6.4   Innovation frontier for the semiconductor industry (after 1970)

8.1   Typology of industry formation

10.1   Cost per peak watt of solar cells

TABLES

2.1   Technology paradigms for engine technologies

2.2   Technology paradigms for transportation technologies

2.3   Technology paradigms for electricity generation technologies

2.4   Technology paradigms for lighting and display technologies

2.5   Technology paradigms for information technologies

2.6   Technology paradigms for telecommunication technologies

3.1   Types of geometrical scaling

3.2   Extent of geometrical scaling (approximate figures)

3.3   Current prices per capacity for large- and small-scale oil tankers, freight vehicles, and aircraft

3.4   Geometrical scaling and rates of improvement for selected components

3.5   Cost reductions for ICs, LCDs, and solar cells

4.1   Classification of selected discontinuities in computers

4.2   Changes in users, applications, sales channels, and methods of value capture

5.1   Technological discontinuities in recording and playback equipment

5.2   Impact of design-related decisions on key dimensions of performance

6.1   Technological discontinuities in the semiconductor industry

7.1   Standards for discontinuities/systems in the IT sector (1950–1995)

7.2   Percentage of de novo semiconductor firms

7.3   Differences between de novo and de alio semiconductor firms in 1995

8.1   Examples of critical choices in the early years of complex industries

8.2   Examples of critical choices in the early years of complex network industries

10.1   Relevant equations for wind turbines

10.2   Best solar cell efficiencies and theoretical limits

10.3   Expected benefits from larger solar cell substrates

ACKNOWLEDGMENTS

This book benefited from the assistance of many people. Early versions of various chapters benefited from comments by anonymous reviewers on individual papers, feedback on presentations at various conferences and at universities such as Carnegie Mellon and Case Western Reserve, and informal conversations with many people, including Ron Sanchez, Chris Tucci, Phil Phan, Brian Arthur, and Dick Lipsey. For the actual book manuscript, Nuno Gil, Jeroen van der Bergh, and Chris Magee provided written and verbal feedback. Special thanks to Chris for closely reading several chapters and identifying many key issues. Three anonymous reviewers provided many insightful comments, and some of these led to very significant improvements to the manuscript. I would also like to thank Margo Fleming and Jessica Walsh, at Stanford University Press, and Martin Kenney, as the series editor, for their help in turning the manuscript into a final book. Most of all, I would like to thank my wife Yvonne for her constant support.

1

INTRODUCTION

The U.S. and other governments spend far more money subsidizing the production of clean energy technologies, such as electric vehicles, wind turbines, and solar cells, than they do on clean energy research and development (R&D).¹ Why? A major reason is that many believe that costs fall as a function of cumulative production in a so-called learning or experience curve, and thus stimulating demand is the best way to reduce costs. According to such a curve, product costs drop a certain percentage each time cumulative production doubles as automated manufacturing equipment is introduced and organized into flow lines.² Although such a learning curve does not explicitly exclude activities performed outside of a factory, the fact that learning curves link cost reductions with cumulative production focuses our attention on the production of a final product and implies that learning gained outside of a factory is either unimportant or is driven by that production.

But is this true? Are cumulative production and its associated activities in a factory the most important sources of cost reductions for clean energy or any other technology for that matter? Among other things, this book shows that most improvements in wind turbines, solar cells, and electric vehicles are being implemented outside of factories and that many of them are only indirectly related to production. Engineers and scientists are increasing the physical scale of wind turbines, increasing the efficiencies as well as reducing the material thicknesses of solar cells,³ and improving the energy storage densities of batteries for electric vehicles, primarily in laboratories and not in factories. This suggests that increases in production volumes, particularly those of existing technologies, are less important than increases in spending on R&D (i.e., supply-side approaches)—an argument that Bill Gates⁴ and other business leaders regularly make. Although demand and thus demand-based subsidies do encourage R&D,⁵ only a small portion of these subsidies will end up funding R&D activities.

Should this surprise us? Consider computers (and other electronic products such as mobile phones⁶). The implementation of automated equipment and its organization into flow lines in response to increases in production volumes are not the main reasons for the dramatic reduction in the cost of computers over the last 50 years. The cost of computers dropped primarily for the same reason that their performance rose: continuous improvements in integrated circuits (ICs). Furthermore, improvements in the cost and performance of ICs were only partly from the introduction of automated equipment and its organization into flow lines. A much more important cause was large reductions in the scale of transistors, memory cells, and other dimensional features, where these reductions required improvements in semiconductor-manufacturing equipment. This equipment was largely developed in laboratories, and these developments depended on advances in science; their rate of implementation depended more on calendar time (think of Moore’s Law) than on cumulative production volumes of ICs.⁷

NEW QUESTIONS AND NEW APPROACHES

We need a better understanding of how improvements in cost and performance emerge and of why they emerge more for some technologies than for others, issues that are largely ignored by books on management (and economics). While most such books are about innovative managers and organizations, and their flexibility and open-mindedness, they don’t help us understand why some technologies experience more improvements in cost and performance than do others. In fact, they dangerously imply that the potential for innovation is everywhere and thus all technologies have about the same potential for improvement.

Nothing can be further from the truth. ICs, magnetic disks, magnetic tape, optical discs, and fiber optics experienced what Ray Kurzweil calls exponential improvements in cost and performance in the second half of the 20th century, while mechanical components and products assembled from them did not.⁸ Mobile phones, set-top boxes, digital televisions, the Internet, automated algorithmic trading (in hedge funds, for example), and online education also experienced large improvements over the last 20 years because they benefited from improvements in the previously mentioned technologies. A different set of technologies (e.g., steam engines, steel, locomotives, and automobiles) experienced large improvements in both cost and performance in the 18th and 19th centuries. An understanding of why some technologies have more potential for improvements than do others is necessary for firms, governments, and organizations to make good decisions about clean energy and new technologies in general.

We also need a better understanding of how science and the characteristics of a technology determine the potential of new technologies. Although there is a large body of literature on how advances in science facilitate advances in technology in the so-called linear model of innovation,⁹ many of these nuances are ignored once learning curves and cumulative production are considered. For example, improvements in solar cell efficiency and reductions in material thicknesses involve different sets of activities, and the potential for these improvements depends on the types of solar cells and on levels of scientific understanding for each type. Lumping together the cumulative production from different types of solar cells causes these critical nuances to be ignored and thus prevents us from implementing the best policies.

Part of the problem is that we don’t understand what causes a time lag (often a long one) between advances in science, improvements in technology that are based on these advances, and the commercialization of technology. And without such an understanding, how can firms and governments make good decisions about clean energy? More fundamentally, how can they understand the potential for Schumpeter’s so-called creative destruction and new industry formation? A new industry is defined as a set of products or services based on a new concept and/or architecture where these products or services are supplied by a new collection of firms and where their sales are significant (e.g., greater than $5 billion). According to Schumpeter, waves of new technologies (which are often based on new science) have created new industries, along with opportunities and wealth for new firms, as they have destroyed existing technologies and their incumbent suppliers.

This is a book about why specific industries emerge at certain moments in time and how improvements in technologies largely determine this timing. For example, why did the mainframe computer industry emerge in the 1950s, the personal computer (PC) industry in the 1970s, the mobile phone and automated algorithmic trading industries in the 1980s, the World Wide Web in the 1990s, and online universities in the 2000s? On the other hand, why haven’t the personal flight, underwater, and space transportation industries emerged, in spite of large expectations for them in the 1960s?¹⁰ Similarly, why haven’t large electric vehicle, wind, and solar industries yet emerged, or when will such industries emerge that can exist without subsidies?

Parts of these questions concern policies and strategies. When did governments introduce the right polices and when did firms introduce the right strategies? But parts also involve science and technology, and, as mentioned previously, they have been largely ignored by management books on technology and innovation,¹¹ even as the rates of scientific and technological change have accelerated and the barriers to change have fallen.¹² When was our understanding of scientific phenomena or the levels of performance and cost for the relevant technologies sufficient for industry formation to occur? We need better answers to these kinds of questions in order to complement research on government policies and firms’ R&D strategies. For example, understanding the factors that impact on the timing of scientific, technical, and economic feasibility can help firms create better product and technology road maps, business models, and product introduction strategies. They can help entrepreneurs understand when they should quit existing firms and start new ones.¹³ And they can help universities better teach students how to look for new business opportunities and address global problems; such problems include global warming, other environmental emissions, the world’s dependency on oil and minerals from unstable regions, and the lack of clean water and affordable housing in many countries.

Some of the problems that arise when firms misjudge the timing of economic feasibility can be found in the mobile phone industry. In the early 1980s, studies concluded that mobile phones would never be widely used, while in the late 1990s studies concluded that the mobile Internet was right around the corner. Some would argue that we underestimated the importance of mobile communication, but I would argue that these studies misjudged the rate at which improvements in performance and cost would occur. The 1980s studies should have been asking what consumers would do when Moore’s Law made handsets free and talk times less than 10 cents a minute. The 1990s studies should have been addressing the levels of performance and cost needed in displays, microprocessor and memory ICs, and networks before various types of mobile Internet content and applications could become technically and economically feasible.¹⁴

Chapters 2 and 3 (Part I) address the potential of new technologies using the concept of technology paradigm primarily advanced by Giovanni Dosi.¹⁵ Few scholars or practitioners have attempted to use the technology paradigm to assess the potential of new technologies or to compare different ones.¹⁶ One key aspect of this paradigm is geometrical scaling, which is a little-known idea initially noticed in the chemical industries (and in living organisms).¹⁷ Part I shows how a technology paradigm can help us better understand the potential for new technologies where technologies with a potential for large improvements in cost and performance often lead to the rise of new industries. Part I and the rest of this book also show how implementing a technology and exploiting the full potential of its technology paradigm require advances in science and improvements in components.

One reason for using the term component is to distinguish between components and systems in what can be called a nested hierarchy of subsystems.¹⁸ Systems are composed of subsystems, subsystems are composed of components, and components may be composed of various inputs including equipment and raw materials. This book will just use the terms systems and components to simplify the discussion. For example, a system for producing integrated circuits is composed of components such as raw materials and semiconductor-manufacturing equipment.

TECHNOLOGICAL DISCONTINUITIES AND A TECHNOLOGY PARADIGM

A technology paradigm can be defined at any level in a nested hierarchy of subsystems, where we are primarily interested in large changes in technologies, or what many call technological discontinuities. These are products based on a different set of concepts and/or architectures from that of existing products, and they are often defined as the start of new industries.¹⁹ For example, the first mainframe computers, magnetic tape–based playback equipment, and transistors (like new services such as automated algorithmic trading and online universities) were based on a different set of concepts than were their predecessors: punch card equipment, phonograph records, and vacuum tubes, respectively. On the other hand, minicomputers, PCs, and various forms of portable computers only involved changes in architectures.

Building from Giovanni Dosi’s characterization and using an analysis of many technologies (See the Appendix for the research methodology), Chapter 2 and the rest of this book define a technology paradigm in terms of (1) a technology’s basic concepts or principles and the trade-offs that are defined by them; (2) the directions of advance within these trade-offs, where advance is defined by a technological trajectory (or more than one);²⁰ (3) the potential limits to trajectories and their paradigms; and (4) the roles of components and scientific knowledge in these limits.²¹ Partly because this book is concerned with understanding when a new technology might offer a superior value proposition, Chapter 2 focuses on the second and third items and shows how there are four broad methods of achieving advances in performance and cost along technological trajectories: (1) improving the efficiency by which basic concepts and their underlying physical phenomena are exploited; (2) radical new processes; (3) geometrical scaling; and (4) improvements in key components.

In doing so, Chapter 2 shows how improvements in performance and/or price occur in a rather smooth and incremental manner over multiple generations of discontinuities. While some argue that these improvements can be represented by a series of S-curves where each discontinuity initially leads to dramatic improvements in performance-to-price ratios,²² this and succeeding chapters show that such dramatic changes in the rates of improvement are relatively rare. Instead, this book’s analyses suggest that there are smooth rates of improvement that can be characterized as incremental over multiple generations of technologies, and that these incremental improvements in a technological trajectory enable one to roughly understand near-term trends in performance and/or price/cost for new technologies.

GEOMETRICAL SCALING

Chapter 3 focuses on geometrical scaling as a method of achieving improvements in the performance and cost of a technology. Geometrical scaling refers to the relationship between the geometry of a technology, its scale, and the physical laws that govern it. As others describe it, the scale effects are permanently embedded in the geometry and the physical nature of the world in which we live.²³

As a result of geometrical scaling, some technologies benefit from either large increases (e.g., engines or wind turbines) or large reductions (e.g., ICs) in physical scale. For example, consider the pipes and reaction vessels that make up chemical plants, which benefit from increases in scale While economies of scale generally refer to amortizing a fixed cost over a large volume, at least until the capacity of a plant is reached, geometrical scaling refers to the fact that the output from pipes varies as a function of one dimension (radius) squared whereas the costs of pipes vary as a function of this dimension (radius) to the first power. Similarly, the output from a reaction vessel varies as a function of one dimension (radius) cubed whereas the costs of the reaction vessels vary as a function of one dimension (radius) squared. This is why empirical analyses have found that the costs of chemical plants rise only about two-thirds for each doubling of output and thus increases in the scale of chemical plants have led to dramatic reductions in the cost of many chemicals.²⁴

Other technologies benefit from reductions in scale. The most well-known examples of this type of geometrical scaling can be found in ICs, magnetic disks and tape, and optical discs, where reducing the scale of transistors and storage regions has led to enormous improvements in the cost and performance of these technologies.²⁵ This is because reductions in scale lead to improvements in both performance and costs. For example, placing more transistors or magnetic or optical storage regions in a certain area increases speed and functionality and reduces both the power consumption and size of the final product, which are typically considered improvements in performance for most electronic products; they also lead to lower material, equipment, and transportation costs. The combination of increased performance and reduced costs as size is reduced has led to exponential improvements in the performance-to-cost ratio of many electronic components.

Like Chapter 2, Chapter 3 and other chapters show how geometrical scaling is related to a nested hierarchy of subsystems. Chapter 3 demonstrates that benefiting from geometrical scaling in a higher-level system depends on improvements in lower-level supporting components,²⁶ and that large benefits from geometrical scaling in a lower-level key component can drive long-term improvements in the performance and cost of a higher-level system. In the second instance, these long-term improvements may lead to the emergence of technological discontinuities in systems, particularly when the systems do not benefit from increases in scale. Part II shows how exponential improvements in ICs and magnetic storage densities led to discontinuities in computers and magnetic recording and playback equipment, as well as in semiconductors. Chapter 9 explores this for other systems.

In fact, most of the disruptive innovations covered by Clayton Christensen, who many consider to be the guru of innovation,²⁷ benefit from geometrical scaling (and experience exponential improvements) in either the system or a key component in the system. This suggests that there is a supply-side aspect to Christensen’s theory of disruptive innovation that is very different from his focus on the demand side of technological change. While his theory suggests to some that large improvements in performance and costs along a technological trajectory automatically emerge once a product finds a low-end niche, and so finding the low-end niche is the central challenge of creating disruptive innovations,²⁸ Chapters 3 and 4 show how geometrical scaling explains why some low-end technological discontinuities became disruptive innovations and why these low-end technological discontinuities initially emerged. A search for potentially disruptive technologies, then, should consider the extent to which a system or a key component in it can benefit from rapid rates of improvement through, for example, geometrical scaling.

Some readers may find the emphasis on supply-side factors in Chapters 2 and 3 (Part I) to be excessive and thus may classify the author as a believer in so-called technological determinism. Nothing could be further from the truth. I recognize that there is an interaction between market needs and product designs, that increases in demand encourage investment in R&D, and that the technologies covered in this book were socially constructed.²⁹ The relevance of this social construction is partly reflected in the role of new users in many of the technological discontinuities covered in Part II, where these new users and changes in user needs can lead to the rise of new industries.³⁰ For example, the emergence of industries represented by microbreweries and artisanal cheeses is more the result of changes in consumer taste than of changes in technology. Some of these changes come from rising incomes that have led to the emergence of many industries serving the rich or even the super rich. When the upper 1 percent of Americans receives 25 percent of total income, many industries that cater to specialized consumer tastes will naturally appear.³¹

This book focuses on supply-side factors because industries that have the potential to significantly enhance most lives or improve overall productivity require dramatic improvements in performance and cost. As Paul Nightingale says about Giovanni