The Economist's own count puts the number of jobs created by the artificial-intelligence boom at roughly 1 million, against about 200,000 layoffs that companies have attributed to AI over the past few years.
Each monthly payroll report is now read for evidence of an AI jobs apocalypse. Alex Domash went looking for the jobs the boom has created instead, and found them in data-center construction, electrical work and software.
"We did a sort of simple back of the envelope calculation and estimate that somewhere in the neighborhood of 1 million jobs have been created because of the artificial intelligence boom."
Domash is an economics correspondent for The Economist, and built the estimate himself — five occupations tied to the data-center build-out, measured against the rest of construction and manufacturing to get a counterfactual.
I listened to the full segment so you can skip it.
Here are the 7 numbers that matter.
👤 Guest: Alex Domash, an Economics Correspondent for The Economist
🎙️ Host: Jason Palmer, a co-host of The Intelligence
📰 Published: 10 September 2026 on The Intelligence from The Economist
🟢 Spotify | 🟣 Apple Podcasts | 🔗 Episode page | ⏱️ 20 min
Key Takeaways
August payrolls came in at 162,000, far above what forecasters expected, and youth unemployment stayed low
Youth unemployment is the series economists have been watching for AI's effect on young workers
Around 200,000 layoffs have been attributed to AI, which is small next to the labor market's monthly churn
The data-center build-out has created roughly 350,000 jobs, many of them open to workers without a degree
Electricians, HVAC installers and data-center technicians, with wage gains large enough to suggest shortages
Occupations closest to AI added almost 750,000 jobs since 2023 beyond what the trend alone would have produced
The white-collar jobs said to be first in line — paralegals, financial analysts, translators — are still growing
Customer service representatives and office and administrative assistants are the two groups actually losing ground
Their job growth was already falling before 2023 and the decline has accelerated since
1. The August Jobs Report
The segment opened on the payroll print rather than on a forecast. The show's own framing: "Last week, the American Bureau of Labor Statistics came out with its monthly jobs report and it found that in August there were 162,000 jobs that were created in the economy," Palmer said.
Domash said the number beat expectations outright. "This far exceeded expectations, a lot of people thought that the number was going to be much lower."
The series economists watch for AI damage to young workers did not move either. "At the same time, the unemployment rate continued to be very low and youth unemployment, which is something that people have been tracking as it relates specifically to the effect of AI on the labor market, also continues to be quite low, despite all of the fears around the effects of AI on young workers."
Palmer put the conclusion as a question — that the "much ballyhooed AI jobs apocalypse is not nigh" — and Domash agreed with it.
2. Layoffs vs Jobs Created
Domash would not credit the whole jobs boom to AI, and said so before making his case. "And I don't want to attribute all of the boom to AI specifically, but it does sort of cut across the narrative that AI is having a significantly negative effect on jobs."
The AI layoff tally is real but small. "Overall, we've seen over the past few years, about 200,000 layoffs that have been attributed to artificial intelligence by companies."
Measured against how many layoffs happen in the US economy in any given month, he said, that level "is actually quite small"
Against it he put a far larger number on the other side of the ledger. "We did a sort of simple back of the envelope calculation and estimate that somewhere in the neighborhood of 1 million jobs have been created because of the artificial intelligence boom."
3. 350K Data-Center Jobs
Most of that million, on Domash's account, is physical work: chip investment, data-center construction, and then the racks, cooling systems and power to service them.
The build-out's hiring is concentrated in jobs that do not require a degree. "And this has created a lot of demand for workers without bachelor's degree in the US economy, both on the construction side, so the electricians, the HVAC installers, but also in terms of operations."
Once the buildings are running, he said, "they do need a lot of technicians to make sure that the servers are running"
He put the infrastructure count at about 350,000. "We estimate that it has created somewhere around 350,000 new jobs in America."
The method was a counterfactual, not a headcount. "So what we did is we looked at five occupations that are closely tied to the infrastructure boom and the data center build-out." Electricians, workers across the utilities industry and data-center technicians are among the five; the comparison group was construction and manufacturing excluding those occupations.
Since roughly 2023 the five categories have outgrown the rest of that sector and are "up roughly 360,000 jobs since 2023"
Wages are the corroborating evidence. "You also do see this in wage growth in these specific occupations for workers such as electricians, such as data center technicians." He read the very large gains as a sign of genuine labor shortages — employers trying to attract more workers and struggling to find enough of them.
4. 750K White-Collar Jobs
The same method applied to the professional occupations closest to AI produced a bigger number. "And we estimate that the total jobs in these occupations have increased almost 750,000 since 2023, relative to what you would expect in the absence of the AI boom."
The occupations he named: software developers, engineers, data scientists, and a broad range of mathematical and science roles
AI does two different things to white-collar work, he said — it makes existing jobs more productive, and it creates classes of work that did not exist before.
An outside estimate lands in the same place. Domash cited Gad Levanon at the Burning Glass Institute, who uses LinkedIn data to identify professional jobs that would not exist without the AI boom: "He estimates that about 1% of professional jobs in the US can now be classified as AI jobs, which corresponds to roughly 1 million jobs."
5. The Jobs Said To Be Doomed
The occupations named as AI's first victims are growing. "We are actually seeing quite strong job growth among paralegals and among financial analysts. You even see job growth roughly flat for translators."
He noted the prediction being tested: that within a couple of years of law firms adopting AI there would be no need for paralegals
His explanation is a demand effect, not a reprieve. Making a worker more productive lowers the cost of the service they sell, which raises demand for it — and the technology can also shift what the worker does. "AI can be an augmenting technology rather than a displacing technology."
Domash called the evidence on these exposed occupations "mixed so far," rather than settled in either direction.
6. What Is Actually Shrinking
Two groups have seen job growth fall away. "Here, the standouts are customer service representatives and office and administrative assistants."
He was explicit that AI did not start this. A lot of those occupations had seen job growth declining even before 2023, but there has been an acceleration over the past couple of years — "and this is something that policymakers and economists should take seriously."
7. The Historians' Case
Domash said none of this should be a surprise, and put it on the economic historians. General-purpose technologies — he named the internet, computers and electrification — hurt particular occupations while creating many new jobs overall.
The longer pattern is the argument. "The long-term trend of the labor market is towards creating all sorts of new classes of jobs." Every few decades, he said, most people are working in jobs that did not exist decades earlier, and he expects the same of AI.
Bonus Insights
Palmer opened the segment with a century of the same worry — the seamstress and the mechanical loom, the horse driver and the car, then steam engines, telephone exchanges and the typewriter. A technology arrives, fear of job losses spreads, and then industries transform and new kinds of jobs demanding new kinds of skills appear.
His conclusion for anyone watching AI do it again: "So if, here in the turn of the AI century, you're fretting about your job, you might wanna relax a little."
Domash's bottom line is that the AI build-out is currently adding far more jobs than AI automation is taking away — around 350,000 in data-center infrastructure and almost 750,000 in the occupations closest to the technology — and that the damage so far is confined to customer service and administrative work rather than spread across the white-collar economy.
Products, Companies & Tools Mentioned
Bureau of Labor Statistics (Its monthly jobs report is the 162,000-job August print the segment was built on)
The Burning Glass Institute (Where Gad Levanon estimates that about 1% of US professional jobs are now AI jobs)
LinkedIn (The data Levanon works from to identify professional jobs that would not exist without the AI boom)
If this was worth your time, send it to someone closer to the industry than you are.
Get the latest market chatter as it happens: