Key points
- McKinsey models a net gain of 5 million jobs
- About 11 million workers may still have to switch careers
- AI is this year's top reason for announced job cuts
McKinsey estimates that about 11 million US workers may need to change occupations by 2035, even if employment grows overall. Challenger, Gray & Christmas says AI was the leading reason employers cited for announced job cuts through September. The question I keep coming back to is whether the people losing jobs can move into the ones being created.
McKinsey's net number is positive
In an 82-page report published Sept. 29, the McKinsey Global Institute estimates that automation could reduce labor demand by the equivalent of about 36 million US jobs by 2035, while growth could add demand for roughly 41 million. The resulting gain of about 5 million needs some context: McKinsey anchors overall employment growth to Bureau of Labor Statistics projections, then models how work could shift between occupations. It isn't an independent forecast that AI will create more jobs than it removes.
Much of the additional demand comes from broader economic changes. McKinsey attributes about 18 million jobs to rising living standards and another 8 million to an aging population. Entirely new occupations emerging from AI could add between 500,000 and 2 million.
The number I'd focus on is 11 million. McKinsey estimates that growth within existing occupations could offset automation's effects for about 25 million workers, allowing them to stay in the same field even as their work changes. Another 11 million, about 7% of today's workforce, may need to change occupations. That estimate ranges from 6 million to 16 million depending on how quickly automation spreads and how much it reduces demand for labor. "The next decade's challenge is mobility, not scarcity," the report says.
Most displaced workers don't have an easy path
McKinsey groups career moves into three paths. A direct path leads to an occupation where demand is growing, with little retraining and no pay cut. A winding path requires more retraining or accepting lower pay. An unpaved path involves larger skill gaps, lower wages, or lengthy credential requirements.
Only about one in seven workers has a direct path, according to the report, while almost half may face an unpaved one. Roughly 85% of growing jobs require a credential or certification, creating another barrier for workers who may otherwise have the right skills.
The shift also divides workers by pay. More than 70% of declining jobs could fall in the bottom two wage tiers, concentrated in office and administrative support, retail and sales, and transportation and logistics. About 60% of growing jobs could fall in the top two tiers, led by healthcare, construction, and management. McKinsey estimates that AI and automation could automate or substantially assist work accounting for about 80% of current office and administrative hours by 2035, the highest share of any occupational group it studied. That doesn't mean 80% of those hours would disappear.
Challenger's layoff count shows the other side
Employers announced 573,195 job cuts through September, down 39% from the same period last year, Challenger reported Thursday. They cited AI in 120,136 cuts, about 21% of the total, making it the leading reason for the year so far. In September alone, AI ranked fifth, accounting for 3,961 announced cuts.
"Companies are in a wait-and-see period right now," said Andy Challenger, the firm's chief revenue officer. He cited high energy costs, the war in Iran, and a rate hike that could make hiring more expensive.
Announced job cuts are down overall, but AI remains the reason employers cite most often this year. That's consistent with a reshuffling of work, though it doesn't establish how McKinsey's longer-term projections will play out. For workers facing those cuts, the distinction offers little reassurance.
The broader job market isn't helping. The US economy added just 29,000 jobs in September, the Labor Department said Friday.
The new jobs are already showing up
We've been tracking where the hiring is going, and it lines up with McKinsey's list. Bank job postings for AI roles rose 49% this year, according to hiring data firm Draup, and mentions of agent skills rose the fastest. On the physical side, AI data center jobs like electrician can pay more than $100,000 without a four-year degree.
Those are exactly the kinds of jobs McKinsey says will grow. They're also exactly the kinds that need a license, a certification, or new training first. An office admin can't walk into an electrician's job next month.
McKinsey found demand for AI fluency has risen elevenfold since 2022. A positive jobs total would still leave the hardest question unanswered. Can displaced workers afford the training, time, and lost income involved in changing careers? Only about one in seven has an easy path.
