Two reports landed Tuesday. One looked ten years ahead. The other looked at this month. Put side by side, they expose a mismatch.
The forecast
The McKinsey Global Institute estimates that 11 million US workers may need an entirely different occupation by 2035. That equals about 6.5% of today’s labor force.
The base case runs on two numbers. Automation could cut labor demand by 36 million jobs. AI-related fields and broader growth could add demand for 40 million.
About 25 million affected workers can stay in their current field, because industry growth offsets the damage. The other 11 million cannot.
McKinsey’s framing stands out: the challenge is “mobility, not scarcity.” The report also expects more jobs than workers in 2035, since the population is aging.
The pace problem
The size of the shift matters less than its speed. McKinsey estimates about 770,000 workers a year would need to switch occupations, against a long-term average of roughly 215,000. That is more than triple the historical rate. The range is wide, too. The estimate runs from 6 million to 16 million workers, depending on how fast AI spreads.
A market that barely moves
Now check the labor data. The Bureau of Labor Statistics JOLTS report shows job openings at a five-month low in August. Voluntary quits sit near a six-year low. Layoffs fell for the second month running.
Hiring has picked up, but only from a very low base. Payrolls grew by about 80,000 a month this year. Last year’s pace stayed under 10,000.
Here is the tension. McKinsey’s forecast needs people to leave shrinking fields and land in growing ones. Workers do that through churn. Right now, the market has almost none.
The market has run in a low-hire, low-fire state for two years. Employers neither add nor cut much. Workers neither quit nor move.
Why workers stay put
Glassdoor chief economist Daniel Zhao says the freeze leaves workers stuck. Frustration and anxiety build, he says, with no healthy outlet.
The numbers back him up. Glassdoor’s Employee Confidence Index hit a record low in September, its third this year. Workers cited job security, economic uncertainty and AI.
That last factor feeds on itself. AI-driven workplace anxiety makes people hold tighter to the job they have, which slows the very movement McKinsey says they need.
The burden also falls unevenly. Lower-income workers may be nearly eight times more likely than higher earners to need a new occupation. Those workers often have the least savings to cover a gap in pay.
AI is not the only culprit
Be careful here. Other forces weigh on this market.
Baby Boomers keep retiring. Net immigration has slowed. The Conference Board’s confidence index fell to 81.9 in September, its lowest level in 12 years. Consumers blamed higher gas prices, the war in Iran and trade uncertainty.
Oxford Economics’ Grace Zwemmer adds a useful counterpoint. Low hiring hurts the unemployed, she notes, but the labor market is roughly in balance. Slow labor force growth matches weak payroll gains.
Research also cautions against blaming AI for everything. A Yale School of Management study found that demand fell for some AI-exposed occupations while others grew, because AI made those workers more effective. AI hits jobs unevenly.
Where the shift lands first
McKinsey’s occupational map shows the split. Healthcare, construction and management grow, while office support, retail and transportation decline.
Companies already prepare for that shift. Amazon’s warehouse automation push aims to raise output without raising headcount. Elsewhere, firms ask staff to document their own workflows so AI agents can take over the tasks.
Both moves work at the task level. Jobs rarely vanish in one step. They thin out quietly, through roles nobody posts.
What to watch next
The September jobs report lands Friday. FactSet’s consensus calls for 95,000 new jobs, down from 162,000 in August. Economists expect unemployment to hold at 4.1%.
The better signal sits in the next JOLTS release. Watch the quits rate. If it stays near a six-year low, McKinsey’s transition has a bigger problem than technology.
Retraining programs and AI tools matter. But a workforce that stays frozen in place cannot move into the jobs AI creates.
Related: China Is Giving Away AI. The Real Product Is Governance
