KPMG U.S. CEO Says AI Is No Y2K as 53% Plan Cuts in Some Areas

KPMG U.S. Chair and CEO Tim Walsh said artificial intelligence will not repeat the Y2K scare as U.S. executives plan workforce reductions in some areas while also restructuring roles and hiring for AI skills.

“I don’t think AI is Y2K,” Walsh said.

The 2026 survey covered 400 U.S. CEOs at large companies, all of which had annual revenue above $500 million. The broader survey included 1,400 CEOs across 12 markets and ran from Aug. 11 through Sept. 21. The results describe large-company respondents, not all U.S. employers.

Over the next two to five years, 53% plan workforce reductions in some areas, while 86% plan to redesign roles, career paths and team structures for human-AI collaboration. Another 83% are focused on retaining and retraining high-potential employees, and 79% plan to hire workers with AI and technology skills.

These figures capture stated plans rather than completed workforce changes. The survey does not identify affected roles, give a count of jobs companies expect to remove or add, or specify when reductions might take place within the two-to-five-year window. The 53% figure refers to respondents planning reductions, not the share of jobs expected to disappear.

Separately, 89% expect their companies’ total headcount to increase over the next three years. That figure measures overall staffing over three years, while the reduction figure concerns some areas over two to five years. The results do not establish a net job forecast.

Cybersecurity also weighs on CEOs’ expectations for business prosperity and AI deployment. 85% expect cybercrime and cyber insecurity to affect their organizations’ prosperity over the next three years. 48% plan to increase cybersecurity investment during that period, and 56% identify cybersecurity as a challenge to deploying AI. Those measures address different questions: anticipated business impact, planned spending and difficulty deploying AI.

Among CEOs planning higher cybersecurity spending, 78% point to AI-enabled social engineering or targeted intrusions, 76% to malware or phishing aided by AI, and 73% to data-protection and privacy concerns tied to AI agents.

“AI and cyber strategies have become inseparable,” Walsh said.