AI in manufacturing 2027: robots, rules and the productivity dip

Executive summary

Manufacturing enters 2027 with more robots on the floor, a growing share of operations run with AI and, in Europe, new rules for machines that learn. Research on US plants also shows that the first years of AI adoption can cost productivity before they add to it. For plant and operations leaders, the task for 2027 is to plan for all three.

Robot deployment keeps growing. The International Federation of Robotics expects 655,000 industrial robot installations worldwide in 2026, up 9%, and 806,000 in 2029. That follows a record 2025, when factories installed more than 600,000 robots, 11% more than in 2024, and the number in operation rose 9% to 5 million. China installed 354,000 units, 59% of the global total. The United States overtook Japan as the second-largest market, with almost 38,500 installations, and Germany installed fewer than 25,000, still the most in Europe [1].
AI is moving into daily operations. In Rockwell Automation’s 2026 survey of 1,560 manufacturers in 17 countries, 59% said they were actively using smart manufacturing technologies and 18% were still piloting them. Respondents put the share of their operations augmented by AI at 34%, mainly in quality, cybersecurity and process optimisation, and expect more than half to be by 2030 [2].
Most data still goes unused. The same survey found that only 43% of the data manufacturers collect is used effectively, and that 46% of them had at least one cyber incident in the past year [2].
Returns often follow a dip. Using Census Bureau surveys of tens of thousands of US manufacturers in 2017 and 2021, researchers found that firms adopting AI saw productivity fall by 1.33 percentage points in the short run, after controlling for size, age, capital and IT. Over the four years covered by the data, adopters went on to outperform non-adopters in productivity and market share [3][4].
Management practice shapes the recovery. Older firms were hit hardest. Their use of structured management practices, such as tracking KPIs and production targets, declined after they adopted AI, and that alone accounted for nearly a third of their productivity losses. Firms that were already digitally mature saw the strongest gains [3].
New EU rules start in January. The EU Machinery Regulation applies to machinery placed on the market from 20 January 2027, with no transition period [5]. It lists machinery whose safety functions rely on machine learning with self-evolving behaviour among its highest-risk categories, which manufacturers cannot certify through internal checks alone [6]. Under the AI Act, obligations for high-risk AI built into regulated products such as machinery follow from 2 August 2028 [7].

Key findings

01Industrial robot installations are forecast to reach 655,000 in 2026 and 806,000 in 2029, after a record of more than 600,000 in 2025

02From 20 January 2027, EU machinery rules treat safety functions that rely on machine learning as high-risk, so they cannot be certified through internal checks alone

03US manufacturers that adopted AI saw productivity fall at first, with older plants hit hardest, before adopters outperformed their peers

Industrial robot installations by market, 2025

China

354k

United States

38.5k

Japan

36.2k

South Korea

30k

Germany

24.8k

Source: International Federation of Robotics, World Robotics 2026

Smart manufacturing in 2026: selected survey results

Manufacturers actively using smart manufacturing technologies

59%

Manufacturers still in pilot mode

18%

Share of operations augmented by AI

34%

Share of collected data used effectively

43%

Manufacturers with a cyber incident in the past year

46%

Source: Rockwell Automation, 2026 State of Smart Manufacturing Report

What leaders should do

Before 20 January 2027, identify any machine you build or buy whose safety functions rely on machine learning, and confirm how it will be certified under the new EU rules.
Plan for a productivity dip in the first year of an AI rollout, and agree in advance how it will be measured and how long it is expected to last.
Keep production management routines such as KPI reviews and output targets running through the rollout, since their decline drove much of the loss at older plants.
Raise the share of collected data that is actually used before adding sensors, and secure the IT and OT connections that AI depends on.

What this report covers

Robot deployment and the outlook to 2029
AI in daily operations
Data and security
The productivity dip after adoption
EU rules from 2027
What to do in 2027

Methodology

We reviewed public sources published between July 2025 and November 2026: the International Federation of Robotics’ annual statistics on industrial robots (World Robotics 2026), a global survey of 1,560 manufacturers in 17 countries by Rockwell Automation and Sapio Research, and research by Kristina McElheran, Mu-Jeung Yang, Zachary Kroff and Erik Brynjolfsson based on US Census Bureau surveys of manufacturers, together with official guidance and legal commentary on the EU Machinery Regulation and the AI Act. Each figure is cited to a dated source. Where we relied on a summary of a primary document, the source list says so.

Survey results are self-reported and are best read as indicators. Forecasts are those of the organisations that published them. The Census research covers US manufacturers in 2017 and 2021, and its findings may apply differently in other countries and in later years. We did not survey organisations directly.