Five Million Factory Robots, Two Speeds: Why Asia Is Automating and Europe Is Stalling

Every September, the International Federation of Robotics (IFR) publishes World Robotics, the reference census of industrial robots. It is built from data reported by robot suppliers and national robotics associations. The 2026 edition, released in Frankfurt on 24 September, crosses a symbolic line: more than five million industrial robots are now working in factories worldwide.

The headline is less useful than the split underneath it. Asia and the Americas grew strongly in 2025. Europe went the other way. This article looks at what the IFR figures actually say, using only the IFR executive summary, the IFR global press release and the IFR’s separate EU-27 release. It then gives our own reading of where artificial intelligence fits. The two are kept apart on purpose, because the IFR statistics count robots, not AI.

What the numbers show

A new record after four flat years

According to the IFR, 603,307 industrial robots were installed worldwide in 2025, an 11% increase on the previous year. For four years, installations had hovered around 550,000 units a year. The previous record, 552,946 units in 2022, was beaten by 9%. The IFR credits strong demand from the electronics industry, the automotive industry and most of general industry.

The installed base grew with it. The IFR puts the global operational stock at 5,079,078 units at the end of 2025, up 9%. In the press release, IFR President Jane Heffner notes that this is more than double the figure of seven years earlier. The IFR values the 2025 market for robot installations at US$18.6 billion, an 11% increase. That figure covers the robots themselves. It does not include software, peripherals or system integration, which usually cost much more over a project’s life.

Asia: three-quarters of the world market

Asia installed 457,315 robots in 2025, up 14%. That is 76% of all new robots worldwide, two percentage points more than in 2024. Three of the world’s five largest robot markets are in Asia.

China dominates the picture. The IFR counts 354,226 installations there in 2025, up 20%. That is 59% of global deployments, and almost 60,000 units more than China’s own previous annual record. Supplier nationality has also shifted. The executive summary says domestic suppliers held around 30% of the Chinese market from 2019 to 2021, then gained ground steeply from 2022. Since 2024, Chinese brands have supplied more than half of the robots installed in China. The press release puts installations by Chinese manufacturers at 195,000 units in 2025, up 15%. That is a domestic share of 55%, slightly down from 57% in 2024.

Elsewhere in Asia the picture is mixed:

• Japan installed 36,219 units, down 19%. It dropped from the world’s second-largest market to third. The IFR expects only slight growth and sees a major recovery as unlikely in the short to medium term.
• Republic of Korea installed 30,000 units, down 1%. It stays the fourth-largest market. The IFR notes that new automotive investments due in 2027–2028 could lift demand.
• India installed almost 10,500 units, up 15%, and ranks sixth worldwide. The IFR says annual installations grew by 27% a year on average between 2020 and 2025.

The Americas: a record year, and the US moves to second place

Installations in the Americas reached a record 57,044 units, up 14%. The United States accounted for 67% of the region, with 38,428 units installed (+12%). The press release calls this the third-highest US figure in the history of the statistics. With Japan shrinking, the US is now the world’s second-largest market, behind China.

The rest of the region was uneven. Mexico fell 7% to fewer than 5,200 units, its third year of decline. Canada rose 5% to almost 4,000 units. Brazil jumped 38% to almost 4,300 units, driven by automotive demand that roughly tripled (+212%). The IFR links this most likely to investments by Chinese carmakers in Brazil.

Europe: the second-largest region, and the only one shrinking

Europe is still the world’s second-largest region for robot installations, but it was the weak spot of 2025. The IFR records 73,054 units, down 14%. The executive summary gives two forces pulling in opposite directions. Europe has benefited from nearshoring, but it has recently suffered from weak demand in the automotive industry.

The IFR’s EU-27 release adds detail, and it is worth keeping the two geographies apart. “Europe” in the global figures is wider than the European Union. For the EU-27, the IFR reports about 60,500 installations in 2025, down 11%. The EU’s operational stock still reached a record of around 712,000 units, and installations have grown 3% a year on average since 2020. In the short term, though, the EU is contracting. In the release’s own words, Europe “is not investing as fast as Asia and North America”.

Within the EU:

• Germany remains the largest market in Europe and fifth in the world. Installations fell 8% to fewer than 25,000 units (24,800 in the EU release). That is 41% of EU installations. Automotive is still Germany’s largest robot customer, but its share is shrinking.
• Italy fell 11% to about 7,800 units.
• France returned to third place in Europe with almost 4,500 units (−8%).
• Spain installed about 4,300 units (−15%).

The automotive sector explains much of the EU decline. The IFR’s EU release puts automotive installations at 14,900 units in 2025, down 25%. Even after that drop, automotive is still about a quarter of EU installations and the largest customer industry.

Electronics leads, automotive close behind, metal and machinery rising

Globally, the electrical/electronics industry has been the largest customer of industrial robots since 2024 and kept that position in 2025. The IFR reports 142,235 installations (+10%) in electronics. Automotive was second with 138,429 units, also up 10% and a new peak, about 3,800 units behind electronics. Metal and machinery came third with 108,571 units, up 22% to a new peak, and an 18% share of all installations.

Readers should note the pattern. For decades, automotive purchasing largely set the rhythm of the global robot market. The IFR says this is no longer true globally, nor in China, Japan and Korea, where electronics now leads. In the EU, by contrast, automotive is still the largest single customer. That explains a large part of why Europe’s curve looks different.

The IFR outlook

The IFR expects the long-term growth trend to continue. It forecasts installations rising 9% to 655,000 units in 2026 and reaching 806,000 units in 2029, an average of 7% a year from 2026 to 2029. It expects the operational stock to pass six million units in 2028. The IFR names China, the US and India as strong growth markets. It also cites relocation of manufacturing driven by supply-chain resilience and trade and industrial policy, plus demographic change, as structural drivers.

What the IFR statistics do not measure

The IFR counts robot units by country, region and customer industry. It does not report how many of those robots run AI-based perception, learned control or any other AI function. Nothing in the installation figures shows how much of the 2025 growth came from AI, and this article does not claim it does. The IFR’s comments on AI sit in a separate technology-trends section and are qualitative.

That section is still useful as a maturity map. According to the executive summary, machine vision for detection and robot guidance is “already established and industry-grade”. Predictive maintenance, process optimisation and AI-based inspection are increasingly used. General-purpose physical AI and learning from simulation and experience are “not yet industry-grade”. On humanoids, the IFR says most remain prototypes or pilot-stage technologies. It adds that conventional industrial robots are currently economically superior for industrial production tasks.

Our analysis: where AI fits

Everything in this section is AndBeyond.Tech’s interpretation. It is not an IFR finding.

The two-speed split is mostly about industry mix and investment cycles

The main reason Asia automates faster than Europe does not look like an AI story. China’s growth sits on a very large electronics and general-industry base. Its domestic suppliers are gaining share, which tends to bring prices down and widen access. The EU depends heavily on automotive. When car plants cut capital spending, as the IFR’s −25% automotive figure for the EU shows, the whole European market feels it. Germany’s dependence on automotive amplifies the effect. Robot demand follows capital spending, and capital spending follows order books, energy costs and confidence. None of those are software problems.

The US case points the same way. Its +12% growth and new second place fit the IFR’s account of relocation and industrial policy pulling manufacturing capacity into high-wage economies. Where new capacity is being built, robots get installed. Where it is not, they do not.

Where AI plausibly matters: lowering the cost of each new cell

AI becomes relevant one step further down: in what each deployment costs and how long it takes. The IFR itself points to “standardized and modular cells, preconfigured application packages, simpler programming, and faster commissioning” as levers to reduce integration cost and project risk. It also says AI, vision and sensing are expanding the range of applications that are technically feasible and economically viable.

We read this as AI working mainly on the integration bottleneck. A vision-guided pick that copes with variation removes fixtures. AI-based inspection replaces a manual quality gate. Easier programming cuts engineering hours per cell. These are the applications the IFR calls mature or maturing, and they match where we have seen factories report returns first. We covered this in AI Cobots and Vision: Where Factories See ROI First.

For Europe, this matters for a specific reason. The EU’s industrial base has many small and medium-sized manufacturers with high product mix and short runs. For them, integration cost per cell often decides whether automation pays off, more than the robot’s price. If AI-assisted tools make high-mix cells cheaper to commission and easier to change over, they could widen the European market beyond automotive. That is a hypothesis, not a measured trend. The 2025 data cannot confirm or refute it.

Where AI is not yet the answer

The IFR’s own caution deserves weight. General-purpose physical AI and learning from simulation are not industry-grade, and humanoids are mostly at pilot stage. A European manufacturer that delays proven automation while waiting for those technologies to mature risks falling further behind. In 2025, the robots that moved the numbers were conventional industrial robots in electronics, automotive and metal-working plants.

Competitive pressure on European suppliers

There is also a supply-side angle that has little to do with AI. The IFR’s EU release notes that Chinese suppliers are developing strategies to enter the EU market, and some have already set up European distribution hubs. Chinese brands already supply more than half of their home market, so they arrive with scale. For European buyers this could mean more choice and lower prices. For Europe’s own robot and integration industry, it adds competition at a time of shrinking domestic demand. In our view, software (programming tools, AI-based perception, fleet management, safety integration) is one of the areas where European suppliers can still differentiate, though no statistic yet shows it.

What to watch in the 2027 edition

Based on the IFR’s framing, a few indicators will show whether the two-speed pattern is cyclical or structural:

1. EU automotive installations. Does the −25% drop stabilise as model programmes and electrification investments restart?
2. General industry in Europe. Do plastics and chemicals (−3% in the EU release) and food and beverage (+4%) grow enough to reduce dependence on automotive?
3. Supplier mix in Europe. How quickly do Chinese brands gain share outside China?
4. The US trajectory. Does relocation keep US installations near record levels?
5. AI maturity. Does the IFR’s technology assessment move learning-based methods from “not yet industry-grade” towards industrial use?

Bottom line

The IFR figures describe a world market that has left its plateau. 603,307 robots were installed in 2025, more than five million are in operation, and the IFR forecasts 655,000 installations in 2026. The growth is concentrated: China alone takes 59% of installations, and Asia 76%. The US now ranks second. Europe fell 14% and the EU 11%, mainly because of automotive weakness.

AI is not what the data measures. The statistics count robots, not algorithms. Our reading is that AI’s near-term role is to lower the cost and risk of each new deployment: vision, inspection and simpler programming first, general-purpose physical AI later. Whether that helps Europe close the gap will depend first on investment, and only then on software.

Related reading

• Physical AI Data Factories: How Synthetic Data Is Accelerating Industrial Robots
• Sim-to-Real in 2026: Closing the Gap Without Endless Real-World Data
• From Pilot to Every Plant: The Race to Make Robot Skills Portable

Images: original illustrations generated for AndBeyond.Tech; no logos or real people depicted.