🔍 Read the full analysis: What The Retreat Of Factory Work Says About AI on ThorstenMeyerAI.com
Get the little things that make your day delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
A November 8 report traces factory employment’s long decline in the United States, Germany and China. U.S. manufacturing employment fell from a June 1979 peak of 19.6 million to a preliminary 12.7 million in October 2026, while factory output nearly doubled from 1979 to September 2026—but remained below its 2007 peak. The figures describe industrial change; they do not establish that AI caused the retreat in factory jobs.
The report uses the Bureau of Labor Statistics’ Current Employment Statistics, a payroll-job series that excludes farming, self-employment and private household work. In June 1979, the series recorded 19,553,000 manufacturing jobs, its highest level. By October 2026, the preliminary count was 12.7 million—about 35% below the peak and 8% of all U.S. payroll jobs.
The share of payroll work in factories peaked much earlier than the number of jobs. It reached 38.8% in November 1943, during World War II, then fell to 21.7% by June 1979 and 8.0% in October 2026. The job count continued to rise for decades after the factory share began declining, as the broader economy expanded.
Output followed a different path from employment. The Federal Reserve’s manufacturing production index was 1.95 times its June 1979 level in September 2026, according to the report. But output peaked in December 2007; the September 2026 level was about 9% below that high. The report’s rough comparison suggests about three times as much output per factory job as in 1979, but it compares job counts, not hours worked, and does not isolate the causes of the change.
Jobs Fell as Factory Output Changed
The comparison matters because employment and production are not interchangeable measures of industrial health. The long-run U.S. figures show that factories can account for a smaller share of jobs while producing more than they did decades earlier. That pattern is relevant to current debates about automation and AI, but it cannot by itself measure how many jobs any particular technology removed.
The more recent figures also complicate a simple story of steadily rising output with fewer workers. U.S. manufacturing output remains below its 2007 peak, even as employment has recovered somewhat from its 2010 low. Workers, policymakers and businesses assessing AI’s possible effects need to distinguish changes in head count from changes in production—and from the causes behind either trend.
Products Worth Considering
Hyperice Hypervolt Go 3 Lightweight Percussion Massage Gun with Carry Case
Garmin Forerunner 935 GPS Running and Triathlon Watch
Garmin Forerunner 165 GPS Running Smartwatch, 43 mm, Black
As an affiliate, we earn on qualifying purchases.
A Long Shift Beyond U.S. Factories
The report traces U.S. manufacturing employment back to January 1939, when factories had about 9.1 million payroll jobs, or 30.3% of the total. Employment rose sharply during World War II, then the factory share declined even as the number of jobs climbed to its 1979 record. Recessions in the early 1980s and later economic changes brought further losses. The largest drop in the account ran from January 2000 to February 2010, when employment fell from 17.3 million to 11.4 million.
The report says the contraction was not limited to the United States. In Germany, the share of workers in industry and construction fell from 49.2% in West Germany in 1965 to 22.9% in unified Germany in 2025. Those figures use a broader sector definition that includes construction, energy, water and mining, and the report cautions that classification changes affect comparisons across time.
China’s industry and construction employment peaked at 232 million in 2012, according to the report. It also challenges the idea that U.S. factory jobs simply “went to China”: the U.S. decline began 22 years before China joined the World Trade Organization. Citing economists’ estimates, the report says rising Chinese imports cost up to about one million U.S. factory jobs from 1999 to 2011, within a period when almost six million were lost.
““The number to remember is 19.6 million.””
— Thorsten Meyer, ThorstenMeyerAI.com
What the Figures Cannot Establish
The report documents employment and output trends but does not quantify AI’s contribution to factory job losses. The historical counts span decades in which technology, trade, recessions, changing demand and other forces also changed. The report’s figures alone cannot separate those effects or show what share of past losses came from automation.
Some details also remain provisional or limited by measurement. The October 2026 U.S. employment count is marked preliminary, and the output comparison is through September. The estimate of output per factory job is a rough ratio based on job counts, not hours worked or worker productivity. The report’s supplied material does not give comparable output-per-worker figures for Germany or China.
Track Jobs, Output and AI Separately
The next useful updates will be revised U.S. payroll data and new manufacturing output readings, which can show whether employment or production has shifted from the reported levels. Further evidence would be needed to connect specific factory job changes to AI, such as industry-level adoption data matched with employment, hours and production over time.
For now, the report’s central distinction is between a documented long-term decline in factory employment and an unproven explanation for that decline. The U.S. factory workforce is far below its 1979 peak, while output remains far above its 1979 level but below its 2007 high. Whether AI changes that relationship in the years ahead remains an open question.
Key Questions
How many people worked in U.S. factories at the 1979 peak?
The report gives a seasonally adjusted peak of 19,553,000 manufacturing payroll jobs in June 1979, commonly rounded to 19.6 million.
How many U.S. factory jobs were there in October 2026?
The preliminary count was 12.7 million, or 8.0% of all payroll jobs, according to the report.
Did factory output keep rising as employment fell?
Over the full period from June 1979 to September 2026, manufacturing output nearly doubled. However, it peaked in December 2007 and was about 9% below that peak in September 2026.
Does the report show that AI caused factory job losses?
No. It documents employment and output trends but does not calculate how much of the decline was caused by AI. The figures do not distinguish AI from other technological, economic and trade-related forces.
Did U.S. factory jobs simply move to China?
The report says that explanation is incomplete: the U.S. decline began 22 years before China joined the WTO. It cites estimates that rising Chinese imports cost up to about one million U.S. factory jobs from 1999 to 2011, out of almost six million lost during that period.
Source: ThorstenMeyerAI.com
