Last summer, 35% of global chief executives told KPMG they were planning workforce reductions due to AI within two to five years. Six months later, the story has flipped. In KPMG's latest U.S. CEO Outlook Pulse Survey, conducted January 26 through February 17 with 100 CEOs of companies pulling in over $500 million in revenue, just 9% said they plan to cut jobs because of AI this year. Meanwhile, 55% said AI will lead them to increase hiring.
That is a striking reversal, but the details underneath it tell a more complicated story than "AI creates jobs."
The Hiring Is Real, But the Jobs Are Different
KPMG U.S. Chair and CEO Tim Walsh put it bluntly: his own headcount at KPMG is not down, but the composition of who he's hiring "has changed fundamentally." The firm is bringing on technologists in ways it never did before. The new roles popping up across the survey - orchestration engineers, AI agent adoption strategists, operations managers for AI systems - didn't exist two years ago.
At the same time, Walsh was direct about which roles are in trouble: "Jobs that are repetitive tasks... That's a scary place to be right now."
So when 55% of CEOs say they're hiring more, they don't mean more of the same people. They mean different people.
$1 in Every $5 Going to AI
The spending numbers back up the confidence. Nearly 80% of the surveyed CEOs said they're allocating at least 5% of their total capital budgets to AI. Forty-one percent are putting in 10% or more, and 35% are spending between 11% and 20% of their entire capital budget on the technology. That rivals what companies spent during the peak cloud infrastructure transition, which took roughly a decade to reshape how businesses operate.
And 77% of those same CEOs agreed that generative AI was overhyped in the past year but will be under-hyped regarding its five-to-ten-year potential. They're spending now because they think the real impact hasn't started yet.
The Gap Nobody's Talking About
Here's the number that should get more attention: two-thirds of the CEOs surveyed admitted they have not yet redefined roles or career paths to account for AI. They're pouring money into the technology and hiring new types of workers, but they haven't restructured how existing employees grow, train, or advance.
That disconnect matters. If you're spending 15% of your capital budget on AI and hiring orchestration engineers, but your existing mid-career employees are still on career ladders designed for a pre-AI world, you're building a two-track workforce. The new hires get the interesting work. The legacy employees get... uncertainty.
Another 31% of CEOs flagged reduced early-career learning opportunities as their top workforce concern. That tracks with a real problem: if AI handles the routine work that junior employees used to cut their teeth on, how do those employees develop judgment and expertise? Nobody has a good answer yet.
Walsh framed the broader economic shift in terms of a single metric: labor cost margin. The share of labor costs in the mix will decline while technology costs rise, he said, and "at the end of the day, I'm going to be able to run a lot more volume." Translation: same revenue, fewer labor dollars, more tech dollars. That math works for shareholders. For individual workers, it depends entirely on which side of the transition you land on.
The optimistic read on this survey is that most large-company CEOs aren't planning mass layoffs over AI. The realistic read is that the workforce is being quietly reshuffled, the new jobs require different skills than the old ones, and most companies haven't figured out how to bridge the gap for their current employees.