Workforce in motion: net job gains, eleven million moves
McKinsey Global Institute's new US workforce study finds AI and automation could remove about 36 million jobs by 2035 while the wider economy creates more than 40 million, but roughly 11 million workers would need to change occupations, and most growing jobs demand postsecondary credentials.
TL;DR
- McKinsey Global Institute's September 2026 report estimates that AI and automation could eliminate about 36 million US jobs by 2035, while growth in the AI value chain and the wider economy creates more than 40 million, a net gain of roughly 4 million.
- The headline is not the net number but the churn: about 11 million workers, roughly 7 percent of the labour force, may need to switch occupations, at a pace press coverage describes as three to four times the historical rate.
- Growth is concentrated in healthcare, construction, and professional and technical services, and 84 percent of growing jobs require postsecondary credentials, which puts the cost of moving on the workers who can least afford it.
Most public debate about AI and jobs is an argument about a single net figure: does AI destroy more work than it creates? The McKinsey Global Institute's new study, released on September 29 and covered widely by Bloomberg, CNN, Axios, and Forbes, answers that with a cautious yes-it-nets-positive, and then spends most of its length on the harder question. Even if the totals balance, who moves, how fast, and what stands in their way?
The net number hides the churn
| MGI, "Workforce in motion" (US, to 2035) | Estimate |
|---|---|
| Labour demand removed by automation | ~36 million jobs |
| Demand created by AI value chain and wider economic growth | 40+ million jobs |
| Workers who may need to change occupation | ~11 million (about 7% of the labour force) |
| Affected workers able to stay in their occupation because of sector growth | ~25 million |
| Office and administrative support: work hours exposed by 2035 | ~80% |
| Technology and analytics: work hours exposed | ~70% |
| Retail and sales: work hours exposed | ~68% |
| Healthcare professionals: work hours exposed | ~26% |
| Growing jobs requiring postsecondary credentials | 84% |
Read together, the figures describe an economy that ends up with more jobs than it has today, and most of them better paid, but where the path from a shrinking occupation to a growing one is neither short nor cheap. Exposure of work hours is highest in office support, technology, and retail and sales, while healthcare, the largest single source of new openings, sees technology arrive far more gradually. More than a third of the positions likely to open in growing occupations sit in healthcare support, healthcare professionals, construction, and management.
The bottleneck is the bridge
The technology may require the largest and most sustained workforce transformation in US history.
That is how the report's message was summarised in press coverage, and the credential figure explains why. If 84 percent of growing jobs demand postsecondary credentials, then the barriers to transition are time, tuition, and lost income during retraining. These fall hardest on lower-wage workers, who are also over-represented in the occupations with the highest exposure. About 70 percent of workers are expected to face some degree of role reinvention even where they keep their job title, so the reskilling challenge reaches well beyond the 11 million who actually move.
For employers, this reframes a familiar tension. The aggregate labour market may clear, but individual firms will still face skills shortages in growing roles and redundancy in shrinking ones at the same time, and the two rarely sit in the same postcode or pay grade.
What this means for leaders
- Plan for redeployment, not just headcount: map which roles in your organisation sit in the high-exposure categories and which adjacent growing roles their skills could reach.
- Treat credentials as a design choice. If 84 percent of growing jobs require postsecondary qualifications, audit which of your own requirements are genuine and where skills-based hiring and paid internal training can replace them.
- Fund transitions directly: tuition support, paid learning time, and internal mobility programmes lower the barrier that this research identifies as decisive.
- Prepare for role reinvention across the board; budget for AI upskilling for the roughly 70 percent whose jobs change, not only for those whose jobs disappear.
- Use the growth sectors as a signal: healthcare, construction, and professional services are where labour demand is expected to be tightest, which matters for wage pressure and hiring competition.
Related reading
- America's Next Workforce Challenge: Building Pathways To Better Jobs (Forbes)
- The reskilling deficit: training for today's AI, not tomorrow's jobs
Source: Workforce in motion: Skills and pathways to future jobs in the United States, McKinsey Global Institute.