Three of the industries that use AI most, information, professional services and finance, lifted their output per payroll job by 5.0% a year since the end of 2022, up from a 2.2% trend over 2005-22, EPB Research said on 25 October. The rest of the private economy went from 0.9% to 1.4%.
That makes the three sectors’ recent trend about 3.6 times the rest’s, and their speed-up 2.8 percentage points against 0.5, by our arithmetic on EPB’s figures. EPB puts the three at about 15% of jobs and 25% of output.
Business Pill · The Jevons Paradox Explained
A short explainer on why cheaper, more efficient inputs can raise total use. It teaches the general idea only and says nothing about EPB’s data.
The key insight: As we read EPB Research’s chart, the productivity gain from AI so far shows up in a narrow band of the economy: the three sectors EPB calls the heaviest adopters lifted output per job to a 5.0% yearly trend, while the rest moved from 0.9% to 1.4%, a gap of about 3.6 times by our arithmetic.
EPB’s chart tracks real output per payroll job: value added from the Bureau of Economic Analysis divided by payroll jobs from the Bureau of Labor Statistics, indexed to 100 in the first quarter of 2005 and running through the second quarter of 2026.
Its title reads “Three of the Heaviest AI Adopters Doubled Their Productivity Growth”. In the left panel, the three sectors’ index ends at 179, against 159 where the 2005-22 trend line extends to. That puts them about 12.6% above their old trend path, by our arithmetic.
The right panel, for all other private industries, ends at 120, with the trend moving from 0.9% to 1.4% a year.
Trend growth in real output per payroll job, 2005-22 against since the end of 2022: 2.2% to 5.0% a year for information, professional services and finance, and 0.9% to 1.4% for the rest of the private economy (EPB Research chart, 25 October 2026, from BEA and BLS data). The 3.6x and point changes are our arithmetic.
What Was Posted on X
AI productivity gains are showing up, but only in ~15% of jobs.
The three heaviest AI adopters (information, professional services, finance) went from a 2.2% productivity trend to 5.0% since 2022.
EPB Research’s post of 25 October, with the chart this piece draws on: the three heaviest AI adopters, information, professional services and finance, went from a 2.2% productivity trend to 5.0% since 2022, against 0.9% to 1.4% for the rest of the economy.
Checking the 15% of Jobs
Payroll data from the Bureau of Labor Statistics for September 2026, which are preliminary, show 2.739 million jobs in information, 10.84 million in professional, scientific and technical services and 6.64 million in finance and insurance, out of 135.7 million private payroll jobs.
That is 14.9% of private jobs, by our arithmetic, in line with EPB’s figure if its sector labels map to those BLS categories. That mapping is our assumption; we did not read EPB’s full report, which is for subscribers.
On EPB’s own shares, 25% of output from 15% of jobs, a job in the three sectors produces about 1.9 times the output of a job elsewhere in the private economy, by our arithmetic.
Where This Fits
Our recent coverage on AI and work includes three related findings. A DeepMind-backed survey found most workers value their jobs for more than pay, Goldman found AI-displaced workers stay jobless longer, and the New York Fed’s data show computer engineering graduates’ jobless rate at 7.8%.
The Business Lever: Output per Job
EPB’s measure is output per job, the denominator a business sees when it asks what each hire produces. As we read the chart, the gain so far sits in the sectors EPB calls the heaviest adopters, while most jobs show a smaller change in trend.
The Structural Read
As we read the measure, it divides output by payroll jobs, so faster growth can come from more output, fewer jobs or both; EPB’s chart does not split the two.
As we read the shares, the three sectors already produced about 1.9 times the output per job of the rest of the private economy, by our arithmetic on EPB’s 25% and 15%, and the gap between their trends widened from 2.2 against 0.9 to 5.0 against 1.4.
As we read the index, the three sectors sit about 12.6% above where their 2005-22 trend would have put them, by our arithmetic, while the rest of the private economy ends at 120.
EPB Research, on X, 25 October 2026
“AI productivity gains are showing up, but only in ~15% of jobs.”
Three Implications
FOR EMPLOYERS EPB’s data show output per payroll job rising 5.0% a year since the end of 2022 in information, professional services and finance.
FOR THE REST OF THE ECONOMY The other 85% of private jobs show a trend of 1.4% a year, up from 0.9%, in EPB’s chart.
FOR READING THE CLAIM The chart shows a correlation with heavy AI use, not proof that AI caused it, and EPB’s full report is for subscribers.
The framework opens: “In any product cycle, there are two curves to follow.” One is what the technology can do; the other is what users can actually do with it.
As we read it, EPB’s split is that second curve in the economic data: the sectors that use AI most show the faster output per job. The framework is the owner’s argument and does not discuss EPB’s data.
EPB’s chart shows a correlation between heavy AI adoption and faster output per job; it does not show that AI caused the change. Sector mix, prices and the post-2020 swings in hiring can move a ratio of output to jobs. The trend lines and index levels are EPB’s; we read them from the published chart, not from EPB’s data, and the 2026 payroll figures are preliminary.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
EPB Research’s data show output per payroll job in information, professional services and finance growing 5.0% a year since the end of 2022, against 1.4% for the rest of the private economy. That trend is about 3.6 times faster, by our arithmetic, in sectors that hold about 15% of private jobs.
A note on sourcing. The trend rates, shares and index levels come from EPB Research’s published chart and post of 25 October 2026, built on BEA and BLS data; we did not read EPB’s subscriber report. The September 2026 payroll figures are preliminary BLS data. The 3.6x, point changes, 12.6%, 14.9% and 1.9x are our arithmetic. We did not contact EPB.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
