Aaron Levie accuses CEOs of 'AI psychosis' over wave of layoffs

Box founder Aaron Levie has sharply criticized the wave of mass layoffs in the tech industry driven by AI optimization. He called the phenomenon 'AI psychosis', arguing that the executives deciding on layoffs often have no clear understanding of what their employees actually do.

What Levie means by AI psychosis

Levie points to a deep gap between the people who decide on AI automation and the people who actually understand how the organization works. When CEOs and investors read marketing materials about the capabilities of modern AI agents, they often fall into a trap of excessive optimism. They overestimate automation's ability to replace entire roles and functions, without grasping the real complexity of the work, its dependencies and its hidden value.

The result of this 'psychosis' is hasty decisions on mass cuts, which companies try to justify as efficiency gains. But these decisions often turn out to be a strategic mistake.

The numbers on layoffs

There are already examples. ClickUp recently announced it was cutting 22% of its team, betting on AI agents. On a broader scale, the tech industry in early 2026 is already approaching the number of layoffs recorded in all of 2025.

• ClickUp cut 22% of its staff, counting on AI agents
• Tech layoffs in 2026 are already comparable to the full year of 2025
• The wave has hit both young startups and established companies
• Many promise 'retraining', but rarely deliver it

Why this is a strategic mistake

«The people who decide to replace your work with AI are usually the least able to understand what you actually do,»

Levie notes.

The problem runs deeper than simple optimism. Positions that require deep knowledge of a company's context, history, customers and the nuances of the work are often the ones cut. Yes, AI can handle routine tasks and standard replies. But AI cannot fully replace complex decision-making, negotiation, relationship management and mentoring.

The result: companies lose critical expertise and institutional memory. A year or two later, when the problems surface, the remaining team struggles to fix them.

What this means

The wave of AI layoffs suggests the industry has entered a phase of excessive optimism similar to the dot-com boom. Before cutting staff, companies need to be more honest about which tasks AI can handle reliably and autonomously, and which require a human in the decision loop.