The world installed 542,000 industrial robots in 2024, double the number from ten years ago. This doubling has not translated into equivalent gains everywhere: Asia captures 74% of new installations, while in some major European countries, manufacturing employment is close to its pre-pandemic level or has declined since its 2018 peak; Western European countries display 267 robots per 10,000 manufacturing employees, compared to 131 in Asia.
The Essentials
- A single price signal—the falling cost of robots—produces opposite effects depending on whether a country’s institutional architecture allows it to integrate technology into a coherent production chain.
- In 2024, 542,000 robots were installed globally, 54% of them in China alone; Europe, with 267 robots per 10,000 employees versus 131 in Asia, demonstrates that high density does not guarantee employment or production gains (IFR World Robotics 2025).
- Creative destruction as described by Philippe Aghion requires framework conditions—training, industrial policy, access to financing—without which the dividend from automation remains captured by a few firms and fails to diffuse throughout the productive fabric.
- The concentration of installations between 2024 and 2028 could lock in manufacturing competitive advantages for an entire generation.
Europe Has More Robots and Fewer Results: What Density Does Not Tell
Western European countries have 267 robots per 10,000 manufacturing industry employees in 2024. Asia has 131. At first glance, the old continent should dominate automated production. It does not.
In some major European countries, manufacturing employment is close to its pre-pandemic level or has declined since its 2018 peak. Market shares in exposed sectors—consumer electronics, batteries, telecommunications equipment—are declining. The apparent paradox is not one at all: density measures an accumulated stock, not a dynamic. What matters is the speed at which new robots integrate into new production lines, in new sectors, driven by growing companies. On this criterion, Europe represents only 16% of global installations in 2024.
The stock tells us what was done yesterday. The flow tells us what is being built today.
China illustrates this gap in extreme fashion. Its operational stock is five times that of the United States according to the IFR. Its dominance stems from its capacity for massive and rapid deployment, from coordination between public investment, technical training, and sector strategy, so that each installed robot fits into a production chain that moves upmarket. Competitive advantage precedes density; density follows.
Price Signal Alone Is Not Enough: The Lesson of the Ten-Year Doubling
Industrial robots cost less. The price of a standard articulated arm varies depending on its characteristics and acquisition conditions. The evolution of robot costs can influence the pace of installations. But it has not produced the same effects everywhere.
This is where Aghion’s thesis becomes operative. In The Power of Creative Destruction, Aghion, Antonin, and Bunel document that long-term growth requires endogenous technological renewal: available technology is not enough. The economic system must be organized to absorb it, so that inefficient firms make way for innovative ones and displaced workers find credible retraining rather than permanent exit from the labor market. Adopting robots alone does not guarantee systemic growth.
Germany’s case illustrates the limit. The German automotive industry massively robotized its assembly lines starting in the 1990s. It maintained its skilled jobs, high wages, and export advantage. But this model was calibrated for an era when the combustion engine remained the product’s core. Automation optimized a value chain that is now challenged at its very foundation by electrification.
German robotization accompanied the maintenance of existing production chains. It is a short-term success and a medium-term structural risk.
Conversely, South Korea coupled its waves of automation with explicit sector bets: memory, displays, then batteries. South Korean robotic investments accompanied the development of multiple industrial sectors. In 2024, the Republic of Korea has 1,220 robots per 10,000 manufacturing industry employees according to the IFR.
China Builds Productive Architecture
Reducing Chinese dominance to state subsidies or cost advantage would be inaccurate. The issue is above all architectural.
China has simultaneously invested in training maintenance technicians for robots, in developing local suppliers of equipment (SIASUN, Estun, Inovance), and in specialized industrial zones where suppliers, assemblers, and customers are within trucking distance. This network reduces transaction costs, accelerates organizational learning cycles, and allows an installed robot to reach maximum efficiency much faster than an isolated machine in a sparsely dense productive environment.
This model has a documented limitation: it tends to reproduce the captures that articles on robotization and retraining needs identify in other contexts. Automation gains can concentrate in certain companies and regions. Automation effects can vary by territory, company, and worker category. The territorial distribution of automation gains can be unequal.
International comparison alone does not allow us to rank the effects of different institutional architectures.
Carl Frey and the Locking in of Trajectories
A competing interpretation sheds different light on this finding. Carl Benedikt Frey, in his work on technology and development trajectories, emphasizes that automation’s effects on employment depend heavily on when and in what context it occurs. The effects of automation on productivity and employment depend on the moment and context in which it intervenes.
This interpretation applies directly to the current situation. In 2024, electronics and automobiles are the two main customer sectors, with roughly 24% each; IFR data do not allow us to assert that a majority of installations concern electronics, electric vehicles, and battery equipment. These are precisely the sectors where Asia concentrates its installations. The automation of existing manufacturing sectors can be accompanied by restructuring.
The divergence in reading between Aghion and Frey is real, but it concerns the remedy, not the facts. Aghion emphasizes framework conditions for innovation (pro-incumbent regulation, research funding, training) as a lever governments can activate. Frey places greater emphasis on sequencing and timing, suggesting that certain trajectories close off and the challenge becomes managing the transition rather than merely stimulating innovation.
Available data do not allow us to attribute manufacturing stagnation to a deficit in framework conditions rather than other factors, nor to oppose this to a presumed excess of automation. And this deficit has a direct human cost. The article on factories struggling to recruit technicians capable of running their robots documents the same rupture from inside production sites: the equipment is there, the human capital to operate it is missing.
Does the 2024-2028 Deployment Lock in Advantages for a Generation?
This is the question that conditions everything else. And it deserves to be asked honestly, without excessive projection.
The hypothesis of durable lock-in rests on a credible mechanism. Companies installing robots today in growing sectors accumulate several advantages simultaneously: an organizational learning curve, a production database that feeds optimization through AI, suppliers calibrated to their processes, and workers trained on these specific systems. These advantages can vary by company, technology, and sector. They can make entry into certain markets more difficult.
The ongoing deployment could durably influence competitive positions in the sectors where it concentrates. This scenario would be consistent with Frey’s framework on locked-in trajectories and with Aghion’s observation that the first-mover advantage in a wave of process innovation is difficult to catch up through regulation alone.
An alternative scenario is nonetheless defensible. Robotic technologies are converging toward increasing standardization; the arms from Universal Robots, Fanuc, or KUKA are programmed on increasingly interoperable platforms. The next generation of humanoid robots, if it delivers on its flexibility promises, could reduce the advantage of specialized installations by making productive conversion less costly. In this scenario, lock-in would be temporary, and a country or region capable of rapidly mobilizing its framework conditions—financing, training, sector strategy—could catch up part of the gap in the 2030s.
The difference between the two scenarios depends on observable signals in the coming years. The speed at which robotic equipment suppliers standardize their interfaces. The pace at which countries currently underrepresented in installations—India, Mexico, Central Europe—gain strength. And above all, the capacity of Western governments to build framework conditions that neither the market alone nor regulation alone produces.
This analysis converges with a concrete political issue. Europe has real assets: a solid scientific base, top-quality equipment suppliers (KUKA, Comau, ABB in part), and a technical workforce whose training level remains high. IFR data show a European share of 16% of global installations in 2024; they do not alone allow diagnosis of a sector strategy deficit or exclude a technology deficit.
Europe can robotize further. The question is whether European and national institutions are capable of building conditions in which a new wave of automation translates into growing sectors rather than mere cost reduction in declining ones.
This question remains open, and the time to answer it is shortening as installations accumulate elsewhere.
The geographic concentration of industrial robotics, as the IFR documents it for 2024, is also a signal about economies’ place in the global value chain: countries that assemble without designing risk finding themselves outside the learning loops that create the value of technical progress over the long term.
Sources
- International Federation of Robotics, Global Robot Demand in Factories Doubles Over 10 Years (IFR World Robotics 2025)
- IFR Robot Density Report, April 2026, International Federation of Robotics
- Our World in Data, Industrial Robots Dataset (from IFR data)
- Philippe Aghion, Céline Antonin, Simon Bunel, The Power of Creative Destruction: Innovation, Growth, and the Future of Capitalism, Odile Jacob, 2020
- Carl Benedikt Frey, The Technology Trap: Capital, Labor, and Power in the Age of Automation, Princeton University Press, 2019
