China’s 14th five-year plan makes quantum a priority sector, while American investments mobilize numerous agencies and 14 centers within a coordinated federal framework. American entities cooperate within a federal coordination structure, even if industrial coordination can remain incomplete. Carl Benedikt Frey, a researcher at Oxford, devotes his latest book to exactly this type of tension: over the past five centuries, nations that centralize deploy quickly, those that decentralize innovate better, and few have managed to sustain both durably.
The Essentials
- Frey examines over a very long historical span, from antiquity to the contemporary era, a structural dilemma: centralization accelerates the scaling of innovations, while decentralization produces breakthroughs, each limiting the other.
- The quantum race offers a site to observe this tension: China industrializes in a centralized way, Western ecosystems combine competition between actors and public coordination mechanisms, and the outcome remains open.
- The book draws on the long history of empires, industrial revolutions, and twentieth-century technology policies to construct this thesis.
- The tension between the two models appears in patent trajectories, deployment timelines, and funding decisions.
- Frey shows that the challenge is to invent a hybrid architecture and that time is pressing, without settling in favor of one model.
Who is Carl Benedikt Frey
Carl Benedikt Frey directs the Future of Work program at the Oxford Internet Institute. He entered public debate in 2013 with a study on automation, the one estimating that nearly half of American jobs were exposed to substitution by machines, and has never left the core question: how economies absorb, or fail to absorb, waves of technological change. How Progress Ends extends this work into long-form history. Frey reaches back to the Song empires, to Italian merchant republics, to the British Industrial Revolution, and to Cold War space programs to seek what determines the fate of nations in the face of innovation.
The Thesis: A Five-Century-Old Dilemma
Frey raises a simple question, hard to refute: some societies invent without industrializing, others industrialize without inventing. The mechanism he constructs throughout the book rests on this opposition. Decentralization—free markets, competition between states, fragmented institutions—creates the conditions for breakthrough innovation. It multiplies experiments, tolerates failure, lets outsiders thrive. Centralization, meanwhile, concentrates resources on existing technologies and reduces the costs of large-scale deployment.
It executes fast. It is less suited to radical innovation and exploration.
Gunpowder is generally placed before the Song, with evidence of primitive powder under the Tang; the Song developed its military uses. The printing press and the compass are Song innovations. But it was fragmented Europe, its rival cities, its competing monarchies, its merchant guilds, that transformed these inventions into world power. Watt conceived a decisive improvement and Boulton’s financing helped its deployment, within a British institutional context that Frey analyzes more broadly. The USSR deployed its industrial technologies at an unparalleled speed, put Sputnik into orbit before anyone else.
According to Frey, Soviet decline stemmed in part from the limits of planning, established interests, and the inability to reform in an era when innovation became more crucial.
Frey formulates a diagnosis of condition, not a pessimistic forecast. He shows that the institutional balances that make progress possible are fragile and reversible. Nations that allow these balances to degrade lose their competitive edge, often before becoming aware of it.
The Central Argument: Scale Kills Invention
The heart of the book is the complementarity and succession between decentralization to explore and centralization to exploit and diffuse. Frey analyzes patent data, research investments, and total factor productivity. From this he draws that industrial concentration produces ambivalent effects, beneficial for scale but risky when it eliminates competition. Below a certain threshold, firms are too small to finance fundamental research. Above another threshold, they are too large to tolerate disruptions that threaten their existing activities.
This mechanism applies to firms. It also applies to states. Centralized industrial planning can allocate massive resources to defined targets, as China does with quantum. It cannot program creative accident, unforeseen discovery, the marginal researcher whose work overtakes the official roadmap. The history of major scientific breakthroughs is largely a history of anomalies the system had not anticipated.
Frey mobilizes Bell Labs here as an instructive counter-example. Bell Labs produced or developed the transistor, the laser, and Unix; between 1947 and 1987, its work was associated with five Nobel Prizes, while its total historical record today stands at ten, within a framework that was neither fully centralized nor fully decentralized. Bell Labs had resources guaranteed by AT&T’s regulated monopoly, but its researchers operated with intellectual freedom that few public or private institutions have been able to reproduce since. Frey sees it as a model, and a warning: AT&T was dismantled in 1984, Bell Labs suffered splits, restructurings, and budget cuts, but retained fundamental research functions, and few institutions have reproduced a comparable model.
The Quantum Race as a Real-Time Test
Frey does not write a book about quantum computing, but his thesis can serve as a general interpretive framework for quantum. China programmed the quantum project into its 2016 plan; in 2017, it documented the launch of the first phase of a national laboratory and the preparation of a “quantum communications and quantum computing” program. The publicly stated objective is building toward 2030 a globally extensive network of quantum communications. Official documentation consulted indicates approximately 7 billion yuan for the first phase of the national quantum information laboratory.
On the Western side, the American National Science Foundation, the Department of Energy, IBM, Google, dozens of startups, and several European agencies work on quantum, each with its own success criteria, its own timelines, its own incentives. Fragmentation generates inventive diversity. It slows industrial coordination.
Neither model has yet won, and that is exactly what Frey predicts. China can cross deployment thresholds that the West will reach later. It can also stumble on an unforeseen technical rupture that its planning system failed to finance because it was not on the roadmap. The comparison of quantum patents is a signal to watch: patents aimed at industrialization, specific applications, communication protocols, materials, reveal which model is actually advancing toward scaling, beyond funding announcements.
On this point, the work of Acemoglu and Johnson on the distribution of technological gains introduces a useful tension to Frey’s thesis: even if fragmented the West produces the quantum breakthrough, nothing guarantees that its gains will be shared widely rather than captured by the first actors to industrialize. Institutional architecture determines not only who invents and who deploys, but also who has access to the result.
The Book’s Blind Spots
Frey constructs his thesis on long-form history, and that is its strength. It is also its principal limit. The examples he mobilizes—the Song empires, Venice, the British Industrial Revolution—all have in common that they are resolved: we know who won. The thesis becomes harder to apply in real time, when balances are still in motion and data remain partial.
On quantum specifically, Frey acknowledges that technical uncertainty persists about dominant architectures. A breakthrough in one of these directions could significantly alter competitive advantages. This concentration of initial investment on a defined research trajectory carries a risk: if the winning trajectory is elsewhere, the advantage of concentration can turn into rigidity.
Another limit concerns the democratic dimension. Frey analyzes organizational models as variables of technological performance. He addresses less the question of whether quantum dominance obtained through authoritarian planning poses problems that exceed economic competitiveness, in matters of military cryptography, surveillance, strategic autonomy of democracies. The question of who controls the machines that decide is a dimension his analytical framework does not fully absorb.
Finally, Frey remains cautious on solutions. He diagnoses the dilemma with precision. He proposes Bell Labs as a model of balance, but he does not describe how to recreate such a balance when fundamental research resources are increasingly contested between underfunded public universities, startups with short exit horizons, and Big Tech whose research agendas follow commercial cycles.
Two Scenarios for 2035, and What Would Allow Distinguishing Them
Frey’s thesis leads to an open question: which model prevails over the horizon of a decade, and on what conditions.
The first scenario is dominance through industrialization. Chinese quantum deployment capacities could provide application advantages, but centralization can limit the emergence of unforeseen radical innovations. History offers precedents: the USSR maintained an edge in certain military domains for decades through concentration, even without generating the foundational innovations that would have allowed it to compete with the American civilian economy. A Chinese quantum advantage in government and military communications would have significant geopolitical implications, even if it did not cover the full range of civilian applications.
The second scenario is decentralized breakthrough. A Western actor, a startup, a university, a DOE laboratory, produces a technical rupture that centralized planning had not programmed. Decentralization demonstrates its inventive capacity. But without coordination of scaling, the delay between breakthrough and industrial deployment leaves the competitive edge unexploited. Fragmentation would win the scientific battle and lose the economic war.
What would allow distinguishing the two trajectories before they play out: the ratio between fundamental research patents and industrial application patents, measured separately for centralized and decentralized actors. An actor accumulating fundamental patents without application patents advances in science; an actor massively patenting applications advances toward market. The gap between the two sides on this measure will say, within three to five years, whether Chinese centralization actually industrializes or whether it saturates a trajectory without breakthrough.
A Hybrid Architecture Remains to Be Invented
Frey does not end on a verdict. He ends on a requirement. Frey argues that a balance between inventive decentralization and deployment capacities is necessary for sustainable progress, with clarity on what each mode contributes and what it destroys.
Several paths exist. A Western quantum coordination agency with strictly limited mandate, system interoperability, common industrial scaling, could close the deployment deficit without absorbing the fundamental research budgets of existing entities. Transnational public-private consortiums on specific applications could test whether partial coordination suffices, without forcing the institutional harmonization that national differences would render untenable. The risk is the opposite: that coordination creates additional bureaucracy that adds a layer of rigidity to existing fragmentation, without producing the industrial acceleration sought.
The question Frey leaves open, and it is his form of intellectual generosity, is whether liberal democracies are capable of this partial coordination. They have accomplished comparable things: CERN, Airbus in its early years, certain NATO defense programs. They have also let windows close, notably in semiconductors, where European lag on robots and on mass industrialization preceded costly structural dependence. Quantum opens a comparable window. The duration of this window is one of the unknowns no one can honestly quantify.
Who Should Read This Book
How Progress Ends addresses anyone wanting to understand why certain nations lose their technological edge. It is a book of structure, not an industrial policy treatise. Frey proposes a framework applicable to any technological race, without claiming to offer a recipe. Economists, public decision-makers, political science researchers, and readers attentive to the long view will find distinct analytical grids, which speaks to the solidity of the work.
Its reading is particularly useful at this precise moment, when debate on technology policy swings between two symmetrically naive positions: let markets decide alone, or plan as if the state knew what markets do not. Frey shows that both positions miss the mechanism. What matters is the institutional architecture that makes invention possible without preventing deployment. It is a less comfortable thesis than the two extremes, and it is precisely for that reason it deserves to be read.
Bibliographic Information
How Progress Ends: Technology, Innovation, and the Fate of Nations Carl Benedikt Frey Princeton University Press, September 2025
Sources
- Carl Benedikt Frey, How Progress Ends: Technology, Innovation, and the Fate of Nations, Princeton University Press, 2025, Oxford Internet Institute
- Carl Benedikt Frey, personal page
- MERICS, China’s Quantum Ambitions, report 2025 (Mercator Institute for China Studies)
- National Science Foundation, annual reports on quantum R&D 2024-2025
- NIST, quantum standardization programs, publications 2024-2025
- CSIS, reports on Sino-American technological competition in the quantum field, 2024-2025
