AI Labor Tracker
Gartner Says AI Will Add 1.7 Million Jobs Worldwide by 2030. About 1.5 Million of That Net Gain Is in the Tech Industry.
Gartner's new forecast reads as good news: more jobs created by AI than destroyed. Break its own numbers down and almost all of the net gain lands in technology, the public sector comes out behind, and 2026 so far shows more jobs lost than gained. The biggest effect it describes is the one layoff counts never see: jobs that simply never get created.
About 1.5 million
technology industry's share of Gartner's 1.7 million net AI job gain by 2030 (1.9 million created minus 0.4 million lost; Gartner, Nov 10, 2026; PLE arithmetic) Derived estimate
On Tuesday in Barcelona, at its big annual conference for chief information officers, Gartner put out a jobs forecast with a cheerful headline. AI, the research firm said, will create 10.6 million jobs worldwide by 2030 and remove 8.9 million. Net: 1.7 million more jobs.
The most useful sentence in the release wasn’t the headline. It came from John Lovelock, Gartner’s chief of research, talking about a third category Gartner counts separately. “Deferred jobs are not the same as lost jobs,” he said. “They simply aren’t created.”
Gartner estimates executives deferred the creation and hiring of 1.3 million positions in 2026 alone. By 2030, it says, 5 million people will have experienced delayed hiring because of it.
So I went through the release line by line. My question: once you split Gartner’s own numbers by industry and by year, what kind of future is it actually forecasting, and for whom?
Split it by industry
Gartner breaks its forecast into three kinds of industries. The numbers are all in the release. The subtraction is ours.
Asset-intensive industries, the ones that make physical things or run physical infrastructure, are the biggest employers. Gartner has them creating 6.4 million jobs and losing 6 million by 2030. That nets to about 400,000.
The public sector comes out behind. Gartner forecasts 0.4 million job gains and 0.9 million losses there. That’s a net loss of about half a million, and Gartner adds that AI investment there is “unlikely to be sufficient to mitigate future severe skills shortages driven by other factors.”
Then technology. Gartner says nearly 70% of the tech workforce has high AI exposure, and it forecasts 1.9 million jobs created against 0.4 million lost. That nets to about 1.5 million.
Add the three and you get roughly 1.4 million. The rest of the 1.7 million comes from industries outside those three groups, which, by subtraction, net out to about 300,000.
Put plainly, about nine-tenths of Gartner’s net gain is in the technology industry. Outside tech, the forecast is close to a wash: some industries a little ahead, government a little behind. That’s not how “AI will create more jobs than it eliminates” sounds when you hear it on the news.
Nearly all of Gartner's net AI job gain is in tech
Millions of jobs, worldwide, cumulative 2025 to 2030. The 2026 panel is one year.
How to read thisBars start at zero except where the net is negative, which runs left of zero. Teal is the technology net. Gray is every other bar. Nets and other industries are our arithmetic from Gartner's published created and lost figures.
About 1.5 million
technology industry's share of the 1.7 million net, 1.9 million created minus 0.4 million lost
Bar chart. Use the up and down arrow keys to move through 16 categories. A data table follows.
Created, asset-intensive: Jobs, millions 6.4
- The World Economic Forum's 78 million is a different count, all structural drivers, and is not on this chart.
Figure 1. Created, lost, and net by industry. Other industries and every net are Post Labor Economics calculations. 2026 deferred is the 1.3 million named in the release.
Source: Gartner, 'Gartner Forecasts AI Will Drive a Net Increase of 1.7 Million Jobs by 2030,' Nov 10, 2026. Net figures and 'other industries' are Post Labor Economics calculations. Chart: Post Labor Economics Download CSV Methodology
Data table
Other industries is the residual: total created 10.6 minus 6.4, 0.4 and 1.9. The same subtraction produces lost and net.
| item | Jobs, millions |
|---|---|
| Created, asset-intensive | 6.4 |
| Created, public sector | 0.4 |
| Created, technology | 1.9 |
| Created, other industries | 1.9 |
| Lost, asset-intensive | 6.0 |
| Lost, public sector | 0.9 |
| Lost, technology | 0.4 |
| Lost, other industries | 1.6 |
| Net, asset-intensive | 0.4 |
| Net, public sector | −0.5 |
| Net, technology | 1.5 |
| Net, other industries | 0.3 |
| 2026 gained | 1.6 |
| 2026 lost | 2.2 |
| 2026 net | −0.6 |
| 2026 deferred | 1.3 |
Split it by year
The release also gives Gartner’s estimate for this year. In its AI job count for 2026, Gartner puts 2.2 million jobs lost and 1.6 million gained, alongside the 1.3 million deferred.
Do the subtraction and 2026 comes out about 600,000 jobs behind before you count deferrals. For Gartner’s cumulative total, which starts in 2025, to finish 1.7 million ahead, the remaining years through 2030 have to make up that gap and then some.
Maybe they will. Technology has a long record of destroying jobs early and creating them later, once companies figure out what to build with it. But it means the positive number is a forecast about the future, while the negative number is Gartner’s estimate of the present.
What “at most 1.8%” means
The release’s summary says, “At Most, Job Losses Will Impact 1.8% of Global Workers.” The body is more specific. It says that by 2030, “between 0.4% and 1.8% of workers will experience job loss directly attributable to AI.”
Those two endpoints match the industry figures. Gartner puts AI-driven job losses at 0.4% of the asset-intensive workforce, 0.4% of the public sector workforce, and 1.8% of the global technology workforce. So the 1.8% isn’t a worst case for all workers. It’s the rate for the industry Gartner thinks is most exposed. That’s our reading of the release; Gartner’s full report, which is for clients, may explain it differently.
There’s one more number that tells you how big the denominator is. Gartner says “just over 20% of workers (435.5 million employees worldwide)” will have to adapt to a changed job. Working backward, that implies a base of a little over 2 billion employees. Set 8.9 million lost jobs against that and you get roughly 0.4%. Small, if you’re looking at the planet. Not small if you’re one of them.
Why this forecast is much smaller than the one you’ve heard
If 1.7 million sounds tiny, it’s because you’ve probably seen a bigger number. In January 2025, the World Economic Forum’s Future of Jobs Report projected 170 million jobs created and 92 million displaced by 2030, for net growth of 78 million.
Some coverage at the time described that as AI’s effect. Ars Technica’s headline read “AI could create 78 million more jobs than it eliminates by 2030.” But the report itself attributes those numbers to “structural labour-market transformation,” every force it surveyed, from technology to demographics to the green transition. It was built from a survey of more than 1,000 employers with over 14 million workers.
Gartner is counting something narrower: jobs it attributes to AI. Two forecasts with different denominators can both be honest. Put side by side in a headline, they’re useless. Whenever you see a big jobs-by-2030 number, the first question is what it’s counting.
What a deferred job looks like
Here’s why I think Lovelock’s line matters more than the headline. A deferred job leaves no trace in the usual places. There’s no layoff, no WARN notice, no severance, no line in a layoff tracker. A position that would have existed next year just doesn’t.
US Foods gave a clean example on Nov. 5. The food distributor, which supplies restaurants, hospitals and hotels, told investors on its third-quarter call that its AI tools are now generating more than $75 million in adjusted earnings a year. Then CEO Dave Flitman said what that means for hiring.
“As a result of the AI-driven technology advancements we have deployed across the sales force over the last two years, going forward, we now expect seller headcount to grow at a low single-digit rate versus our prior mid-single-digit target,” he said.
Nobody at US Foods lost a job in that sentence. An analyst asked whether the slower pace was temporary. “For the foreseeable future,” Flitman said. He was careful to add, “the sales force is the lifeblood of this company, and it always will be,” and that the change also reflects a switch to all-commission pay. But a lower growth target means fewer sales reps hired each year than the company had planned. That’s a deferral in Gartner’s sense, in a single company.
The gap between “mid” and “low” single digits sounds like a rounding error. It compounds. As an illustration, not US Foods’ numbers, a team growing 5% a year ends up about 28% bigger after five years. At 2% a year, it ends up about 10% bigger. The difference is the hires that never happen.
Revolut’s CEO, Nik Storonsky, described the same thing on a podcast this week, as City AM reported on Tuesday. He said team sizes at the fintech haven’t changed yet, and that once its automation work is done, he hopes Revolut can hire fewer people.
Can you see deferrals in US data?
Not directly. No government survey asks employers how many jobs they decided not to create. But you can see the shape a lot of deferrals would leave behind.
In the Bureau of Labor Statistics’ job openings and turnover survey, the hires rate, hires as a share of employment, was 3.4% in September 2026. It has stayed between 3.1% and 3.5% every month since mid-2024, by our pull of the BLS data. In 2019 it averaged about 3.9%.
Layoffs, meanwhile, stayed low. The layoff and discharge rate was 1.0% in September. A labor market where few people get fired and few people get hired is what you’d expect if companies are holding headcount flat and letting attrition do the work. It’s also what you’d expect from high interest rates, tariff uncertainty and ordinary caution. The data fit the deferral story without proving it.
There’s also a tension with Gartner’s tech number. In the US, the information sector, which includes software publishing, telecom and data processing, had about 119,000 fewer jobs in October 2026 than a year earlier, according to the BLS payroll survey. Gartner’s technology industry is global and defined differently, so the two don’t contradict each other outright. But if most of the net gain is supposed to come from tech, the US part of tech isn’t showing it yet.
Who pays for a deferred job
A layoff hits people who have a job. A deferral hits people who don’t have one yet: the graduate who would have filled the entry-level seat, the career-changer who would have been the new sales rep, the worker trying to move up from a smaller company.
That’s a different set of people, and it’s harder to see. We’ve written about it from other angles, including buyouts that never appear in layoff counts and how AI-exposed workers who lose a job take longer to find the next one. Deferred hiring is the same pattern. The damage shows up as slower entry and longer searches, not as a spike in firings.
Gartner’s 5 million deferred by 2030 is a modeled estimate, and Gartner doesn’t publish its method in the free release. I wouldn’t lean on the exact figure. I would lean on the idea. When a company like US Foods says it will grow a job category more slowly because software makes each person more productive, it’s telling you where the effect lands.
My take
My read: Gartner’s forecast is more interesting than its headline and less reassuring. Net of everything, AI adds jobs mostly inside the technology industry, takes a few from government, and roughly breaks even everywhere else, after a year Gartner itself estimates went the wrong way. The most important effect it names isn’t destruction or creation. It’s delay.
I also want to be fair to the forecast. Gartner’s analysts are saying something I agree with: that the companies furthest along don’t treat AI only as a way to replace labor. Helen Poitevin put it this way in Barcelona: “They ensure AI’s economic value goes beyond replacing labor.” US Foods fits that. Its CFO said most of its AI benefit comes from extra sales and better margins, not cost cuts.
The limit of the evidence: this is one firm’s model, with a full report behind a paywall. The industry split is in the public release, but the assumptions behind it are not.
Back to Barcelona
Lovelock’s point was that a job that’s never created doesn’t look like a loss. That’s exactly why it’s easy to miss. The deferral at US Foods took one sentence on an earnings call, and it’ll show up only as a slower rise in the number of sales reps, a few years from now, if anyone checks.
My recommendation: if you want to know how AI is changing hiring in your industry, stop watching layoff announcements and start reading hiring targets. Listen to the earnings calls of the biggest employers in your field for phrases like “headcount growth,” “hiring pace” and “seller headcount.” When those numbers come down and revenue doesn’t, that’s the deferred job, and it’s the opening you were planning to apply for.