Diffusion, innovation and the tech-politics underlying power transitions
March 1, 2025
Introduction
Klaus Schwab, founder of the World Economic Forum, first popularised the term ‘Fourth Industrial Revolution’ (Schwab 2015).1 Schwab identified artificial intelligence (AI), robotics and biotechnology as some of the key emerging technologies of the unfolding Fourth Industrial Revolution (4IR). Recent developments have elevated one technology over all others as being the defining feature of the Fourth Industrial Revolution — AI.
The great power rivals of the twenty-first century — the United States and China — have both intensified their efforts to become the global leader in AI. Previous industrial revolutions have been associated with great power transitions. Will the geopolitical winners and losers of this century be decided by who leads the battle for AI supremacy? Will China eclipse the US as the preeminent global power?
Authors: Lokendra Sharma & Arindam Goswami are research analysts working with the High-Tech Geopolitics Programme at the Takshashila Institution in Bengaluru.
Technology and the Rise of Great Powers: How Diffusion Shapes Economic Competition (Princeton University Press, 2024) by Jeffrey Ding attempts to answer the above questions. Challenging alarmist accounts that have highlighted China’s lead in various AI innovation indicators, Ding argues that diffusion of general-purpose technologies (GPTs)2 such as AI matters more than just innovation. Highlighting the uphill challenges China faces in unseating the US, he argues that the US is ahead of China as far as GPT diffusion metrics related to AI are concerned.
In this book, Ding applies the GPT diffusion theory to provide a fresh retelling of historical industrial revolutions. His argument has relevance beyond the great power rivalry between the US and China; it has lessons for rising powers such as India.
Unpacking Ding’s GPT and Leading-Sector Theories
The central theme that motivates Ding’s scholarly inquiry is one that has inspired a lot of international relations scholarship3 — what causes the rise and fall of great powers? More specifically, what role do technological revolutions play in power transitions?
In answering this, Ding challenges the dominant idea of a leading-sector (LS) mechanism that ‘stresses a country’s ability to dominate innovation in leading sectors’ (p.15). Building on decades-long scholarship on GPTs4, Ding argues that it is the diffusion of GPTs, as opposed to the LS model, that best explains the historical winners and losers during power transitions.
Contrasting Causal Mechanisms
Demonstrating the analytical rigour that is present in the entire volume, Ding describes the contrasting causal chain for the LS and GPT mechanisms:
Leading-Sector (LS) Mechanism: - Domination of one great power in certain leading sectors - Leads to monopoly profits (following classic first-mover advantage) - Catapults the great power to economic pre-eminence
General-Purpose Technology (GPT) Mechanism: - ‘Some great powers sustain economic growth at higher levels than their rivals do because, during a gradual process spanning decades, they more intensively adopt GPTs across a broad range of industries’ (p.16)
While not discounting the significance of innovation in cutting-edge sectors, Ding favours the diffusion of GPTs over LS mechanism because the former leads to economy-wide productivity growth.
Empirical Evidence
Japan’s Experience (1983-1991): - Achieved 2.4 per cent annual total factor productivity (TFP) growth through leadership in high-tech sectors like semiconductors - However, reliance on LS mechanisms led to stagnation (0.2 per cent TFP growth in the 1990s) - Resulted in a widening GDP gap with the US, which prioritized GPT diffusion
Second Industrial Revolution (1870-1913): - The US surged ahead of Britain by diffusing electricity broadly - US GDP grew 5.3 times versus Britain’s 2.2 times - By 1912, US per capita electricity production was double Germany’s and five times that of Britain - This diffusion enabled the US to establish a per capita GDP lead over Britain by 1900
The Three Industrial Revolutions
Ding analyses in a very nuanced manner the First, Second, and Third Industrial Revolutions, that were defined by Britain’s rise, the US’s ascent, and Japan’s challenge, respectively. He compares and assesses LS and GPT theories to find out which one provides a compelling understanding of how technological leadership and transformation helps states achieve economic leadership. In Ding’s view, all three case studies provide support for the GPT theory.
First Industrial Revolution (1780-1840): Britain’s Rise
Britain became the world’s most advanced economic power during the (first) Industrial Revolution not through breakthrough innovations, as claimed by the LS theory, but because it was able to spread mechanical skills across industries, which aligns with the GPT theory.
Key Institutional Advantages: - France produced elite engineers, but failed in disseminating technical knowledge widely - Britain developed a flexible apprenticeship system and created institutes which helped spread mechanical expertise - Mechanics in Britain had superior access to technical publications and training opportunities - The GPT skills infrastructure contributed immensely to Britain’s success
Second Industrial Revolution (1870-1914): US Ascent
The US emerged as the technological and economic leader, again not so much through technological innovations as through systematic technological diffusion.
Candidate Technologies: - Leading sectors: steel, electrical equipment, chemicals, and automobiles - GPTs: interchangeable manufacturing, electrification, chemicalisation, and internal combustion engines
US Institutional Innovations: - Created comprehensive engineering education systems emphasising practical, experience-based learning - Land-grant schools, technical institutes, and strong university-industry linkages - In chemical engineering, pioneered the ‘unit operations’ concept, breaking down complex processes into standardized, transferable components - While Germany led in chemical breakthroughs, the US systematised, standardised, and created institutional mechanisms for spread of technological understanding
Third Industrial Revolution (1960-2000): Japan’s Challenge
The Information Revolution serves a different purpose in Ding’s analysis. Japan and the US were engaged in close economic competition, spurred by developments in information and communications technologies (ICT). During the 1970s and 1980s, scholars and policymakers in the US anticipated Japan eclipsing the US as an economic powerhouse. However, such fears were not realised as Japan’s growth story stalled in the 1990s.
Why Japan Failed to Eclipse the US: - While Japan dominated innovation in key technology sectors (semiconductors, consumer electronics, computers) - The US was ahead in diffusing ICTs during the same period - The US led in diffusing computerisation (the chosen candidate GPT for IR-3) - ‘Institutional adaptations that widened the base of computer engineering skills and knowledge proved crucial to the enduring technological leadership of the United States’ (p.148)
Therefore, the IR-3 case study disconfirms the LS mechanism while supporting the GPT theory.
Critical Assessment
Shortcoming: Inadequacy in Threat-Based Explanations
Ding also examines varieties of capitalism (VoC) and threat-based explanations alongside GPT and LS theories. While he convincingly demonstrates that VoC does not explain Japan’s failure to eclipse the US, he provides insufficient grounds to discard threat-based explanations for IR-3.
Varieties of Capitalism (VoC) highlights differences among developed democracies in labour markets, industrial organization, and interfirm relations, separating them into coordinated market economies (CMEs) and liberal market economies (LMEs).
The Problem with Ding’s Analysis: - The book fails to acknowledge that Japan and the US did not face threats on a similar scale - The US was embroiled in an ideologically-fuelled great power rivalry with the Soviet Union at the global level - Japan’s threat environment was primarily regional - There is scope to strengthen the threat-based explanations approach, especially for the IR-3 period
GPT and LS Are Not Either-Or
There is a risk of policymakers misinterpreting the book’s argument to mean that it is necessarily an either-or choice between LS and GPT theories. States can focus on both: - Gaining first-mover advantage in critical sectors - Building up GPT skill infrastructure5 to help diffuse knowledge and technological expertise
In the highly charged geopolitical environment of the present, gaining expertise in critical sectors is essential, and research in academic institutions and laboratories could yield significant benefits.
Application of Ding’s Framework to India
While the book is primarily concerned with the rise and fall of great powers, it has lessons for rising powers such as India. India did not feature as a power of global significance during any of the previous industrial revolutions:
- IR-1 and IR-2: India was a British colony
- IR-3: India was a newly-independent state with extreme poverty and sluggish economic growth
- 1990s onwards: India’s economy started growing faster, coinciding with the spread of computers, telecom, and internet
India’s Computerisation Success
India undertook significant diffusion of the GPT of IR-3—computerisation—beginning in the 1980s, and accelerating through the 1990s and 2000s:
IT Industry Growth: - 1990: India’s IT and IT-enabled services (ITeS) sector was about USD 100 million - 2024: India’s IT-BPM industry is expected to reach USD 254 billion
Employment: - 1997: 160,000 people employed in the software industry - 2024: IT-ITeS industry employs more than five million people
Education Infrastructure: - 1993: About 100 engineering colleges offered computer science degrees; ~3,000 graduates - 2022-23: 2,461 AICTE-approved institutions offering computer science education; 376,048 students enrolled
Strategy for IR-4: AI Diffusion
Given that AI has all the characteristics of GPTs, following Ding’s GPT Diffusion Theory, AI’s diffusion across different sectors is necessary for economic growth. IR-4 offers India an opportunity to increase its relative power by focusing on:
GPT Diffusion Mechanism: - Expanding AI engineering and knowledge - Creating necessary skill infrastructure - Significant financial resource allocation
Trade-offs to Consider: - Potential lack of focus on other important technologies due to resource constraints - Increased job displacement - Managing computational resource constraints - These must be weighed against productivity gains and opportunity costs of inaction
Balancing Innovation and Diffusion
Notwithstanding the focus on diffusing GPT such as AI during IR-4, India must also work on pioneering, breakthrough AI innovation. Current challenges include:
- Trade wars and supply chain instability
- National security concerns affecting global value chains
- Difficulty in sourcing high-tech innovation products/services
Strategic Recommendations:
R&D Investment: - For LS theory: R&D for staying cutting-edge in leading sectors - For GPT theory: R&D can happen anywhere as long as technology can be sourced and diffused - Focus on open-source AI models as a GPT diffusion mechanism
Resource Optimization: - India’s relative resource constraints (compared to China and US) require optimized deployment - Short-to-medium term impact focus needed - GPT diffusion takes decades to become significantly impactful
Dual Strategy: - LS mechanisms for short-to-medium term returns - GPT mechanisms for long-term returns
As Ding argues: ‘if the lessons of past industrial revolutions hold, the key driver of a possible US-China economic power transition will be the relative success of these nations in diffusing AI throughout their economies over many decades’ (p.189).
Why the Book Matters
The work featured in this book started as Ding’s dissertation at the University of Oxford and continued to develop during his stints at Stanford University and George Washington University. This dense academic treatise presents new ideas regarding:
- Technological innovation
- Economic development
- Global power dynamics
Key Contributions: - Provides a framework for examining ongoing technological transformations - Helps understand how technologies actually transform economies - Challenges readers to think beyond individual inventions - Considers broader institutional and skill-related systems that drive technological and economic progress
The book’s central message — that diffusion of GPTs (as opposed to innovation in leading sectors) during industrial revolutions decides the winners and losers of power transition — has profound implications for policymakers navigating the Fourth Industrial Revolution.
References
- Gopalakrishnan, Kris S. 2016. Indian IT and ITeS journey: Liberalization and beyond. Mint, April 26. https://www.livemint.com/Opinion/fNjocJ9cwlGCDqLWt2OjXP/Indian-IT-and-ITeS-journey-Liberalization-and-beyond.html
Footnotes
Schwab, Klaus. 2015. “The Fourth Industrial Revolution: What It Means and How to Respond.” Foreign Affairs, December 12. https://www.foreignaffairs.com/world/fourth-industrial-revolution↩︎
Note that this is different from the ‘GPT’ in ChatGPT.↩︎
Kennedy 1987; Yang 2013; Brooks and Wohlforth 2015; Chen and Evers 2023↩︎
Bresnahan and Trajtenberg 1995; Petralia, 2020↩︎
Ding defines GPT skill infrastructure as ‘education and training systems that widen the pool of engineering skills and knowledge linked to a GPT’ (p.8)↩︎