TL;DR
Get everyday essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
A November 8 report tracks factory employment in the United States, Germany and China, finding that U.S. factory jobs peaked in 1979 while output continued to rise until 2007. It says automation and productivity help explain the gap, while Chinese import competition contributed to some U.S. job losses; the source does not quantify AI’s specific role.
A report published by Thorsten Meyer AI on November 8 traces the rise and retreat of factory employment in the United States, Germany and China, finding that U.S. manufacturing employment fell from a peak of 19.6 million in June 1979 to a preliminary 12.7 million in October 2026 even as factory output remained nearly twice its 1979 level. The figures show a long-term split between employment and production; the report discusses technology and trade, but does not isolate how much of the change is due specifically to artificial intelligence.
The report uses U.S. Bureau of Labor Statistics payroll data to track manufacturing jobs. The series begins in January 1939, when factories employed about 9.1 million people, or 30.3% of nonfarm payroll workers. The employment share reached its high in November 1943, when factories accounted for 38.8% of payroll jobs. The number of factory jobs kept rising after that share began to fall, reaching 19,553,000 in June 1979, the series’ highest count.
By October 2026, the preliminary manufacturing payroll count was 12.7 million, about 35% below the 1979 peak and 8.0% of nonfarm payroll employment. The steepest decline described in the report came from January 2000 through February 2010: employment fell from 17.3 million to 11.4 million, a loss of nearly six million jobs. The count recovered somewhat during the 2010s, fell during the 2020 pandemic shock, and later stood at the preliminary October 2026 level.
Production followed a different path. Using the Federal Reserve’s manufacturing industrial production index, the report says output in September 2026 was 1.95 times its June 1979 level, while output per factory job was about three times as high. That per-job comparison is the report author’s calculation and does not account for hours worked. The report also says output has been roughly flat since its 2007 high and was about 9% below that high in September 2026.
How factory work changed jobs and output diverged
U.S. factories produced more with fewer payroll jobs over the long run. The report traces technology and trade, but does not measure AI’s specific effect.
Jobs and production split
Manufacturing jobs fell while output rose, showing that production alone does not reveal how many people factories employ.
- About 35% below the 1979 peak by October 2026.
- Manufacturing’s share of payroll jobs fell from 21.7% in 1979 to 8.0% in 2026.
- Nearly twice its 1979 level in September 2026.
- About 9% below its 2007 high.
Different countries, different measures
The comparisons describe distinct employment measures and should not be treated as directly equivalent.
| Place | Reported peak or comparison | Measure |
|---|---|---|
| United States | 19.6 million in 1979 | Manufacturing jobs |
| West Germany | 49.2% in 1965 | Industry and construction share |
| Germany | 22.9% in 2025 | Industry and construction share |
| China | 232 million in 2012 | Industry and construction employment |
What the evidence says about AI
The historical figures show a changing relationship between production and employment, not AI’s direct impact.
- The report does not quantify jobs created or eliminated by AI.
- Output per factory job was about three times its 1979 level, without adjustment for hours worked.
- Technology, trade, demand and changing production may all contribute.
Trade was part of the story
The report says Chinese import competition contributed to some later U.S. job losses, but does not explain the full decline.
More Output, Fewer Factory Jobs
The figures show why factory output and factory employment cannot be treated as the same measure of industrial strength. U.S. manufacturers produced substantially more than in 1979 with fewer payroll jobs, a pattern consistent with increased output per worker. The report links this divergence to productivity and technology, but the figures alone do not establish which technologies caused the change or how much each contributed.
For workers and communities, the distinction matters because rising production does not necessarily restore the number or kinds of jobs lost. The report places the U.S. decline within a wider change in the composition of work: manufacturing represented 8.0% of U.S. payroll jobs in October 2026, compared with 21.7% at the 1979 employment peak. That share measures manufacturing payroll jobs against all nonfarm payroll jobs; it is not a measure of total output or of all people doing industrial work.
The episode’s title foregrounds AI, but its supplied evidence is primarily a historical account of manufacturing employment, output and trade. It gives no estimate of how many jobs AI has created or eliminated. Readers should therefore treat the long-run numbers as evidence of a changing relationship between production and employment, rather than a measurement of AI’s direct effect.
Products Worth Considering
Apple MacBook Air 13.6-inch Laptop with M5 Chip, 16GB Memory, 512GB SSD, Sky Blue
Logitech MX Creative Console with 9 Customizable Keys, Graphite
Samsung Odyssey G7 40-Inch WUHD Curved Gaming Monitor (G75F)
As an affiliate, we earn on qualifying purchases.
Three Countries, Different Peaks
The U.S. employment peak came after the peak in manufacturing’s share of payroll jobs. The share was highest in 1943, during the Second World War, while the total number of factory jobs reached its high in 1979. The report presents that separation as an example of how an economy can add industrial jobs even while factories account for a shrinking proportion of a growing workforce.
The report also compares the United States with Germany and China. It says West Germany’s industry and construction share reached 49.2% in 1965, while Germany’s share was 22.9% in 2025. These figures cover industry and construction, a broader category than manufacturing alone, so they are not directly comparable to the U.S. manufacturing payroll share.
For China, the report puts the peak in industry and construction employment at 232 million in 2012. It uses the comparison to challenge the simple claim that U.S. factory jobs all moved to China. The U.S. decline began 22 years before China joined the World Trade Organization, although the report says Chinese imports contributed to losses later on.
““The proportion of the working population engaged in secondary industry appears in every country to rise to a maximum and then to begin falling.””
— Colin Clark, as quoted in the Thorsten Meyer AI report
The figures presented do not establish how many factory jobs were lost specifically because of AI, robotics or other automation. The report describes output per job as roughly three times its 1979 level, but that calculation does not adjust for hours worked and does not separate technology from changes in products, processes, demand or the mix of factories.
The trade estimate also has limits. The report says economists estimate that rising Chinese imports cost up to about one million U.S. factory jobs from 1999 to 2011, compared with almost six million factory jobs lost from January 2000 to February 2010. These are estimates with different time windows, and the source does not identify the studies or methods behind them in the supplied material. The October 2026 employment count is preliminary and may be revised.
Watch the Revised Employment Data
The next near-term check is whether the preliminary October 2026 employment count changes when updated payroll data are released. Future monthly employment and industrial production figures can show whether manufacturing jobs and output continue along separate paths, but they will not by themselves identify AI’s contribution.
The report is part of “Work After Work,” an episode series, and is also offered as a roughly seven-minute video. The supplied material does not name a date for a follow-up episode or a specific next release. A clearer assessment of AI’s effects would require evidence that separates AI adoption from other sources of productivity change and tracks resulting job changes over time.
Key Questions
When did U.S. factory employment peak?
The report places the peak at 19,553,000 jobs in June 1979, based on seasonally adjusted manufacturing payroll data.
How many U.S. factory jobs were there in October 2026?
The preliminary figure was 12.7 million, or 8.0% of nonfarm payroll jobs. The report says that was about 35% below the June 1979 peak.
Did factory output fall along with employment?
No. The report says U.S. manufacturing output in September 2026 was 1.95 times its June 1979 level. It also says output had been roughly flat since its 2007 high and was about 9% below that high in September 2026.
Does the report show how many jobs AI eliminated?
No. It discusses AI and technology in the context of factory work but does not quantify AI’s specific effect on employment. Its output-per-job comparison does not separate AI from other influences or adjust for hours worked.
Did Chinese imports account for all U.S. factory job losses?
The report says no. It notes that the U.S. decline began before China joined the WTO and cites estimates that rising Chinese imports cost up to about one million U.S. factory jobs from 1999 to 2011, within a larger period of manufacturing losses.
Source: Thorsten Meyer AI
