Silicon Valley’s sweeping layoffs demonstrate that tech CEOs are experiencing ‘AI psychosis,’ Box CEO says

(SeaPRwire) –   A widening gap is emerging in Silicon Valley between executive suites and employee workstations.

Aaron Levie, the chief executive of content management firm Box, recently voiced on X what many are thinking privately, noting that his fellow tech leaders do not fully comprehend the extent of work required for AI.

“CEOs are uniquely susceptible to AI psychosis because they are far enough removed from the final mile of work necessary to extract the most value from AI,” Levie stated on X.

He continued: “Consequently, when they experiment with AI, they witness ideal outcomes, frequently overlooking the subsequent 10 or 20 steps required to achieve sustainable results from agents.”

Essentially, executives only see the technology’s strengths, remaining detached from the glitches, hallucinations, and other errors that employees handling the daily tasks face.

This insight aligns with current data. A 2025 survey by AI company Rev revealed that intensive AI users encounter three times as many hallucinations and take nearly ten times longer to obtain answers. These employees are “tokenmaxxing,” or exhausting their AI token quotas. This is an aspect of the technology that some CEOs completely overlook as they prepare to terminate thousands of employees to substitute them with AI.

Furthermore, the return on investment outlook is even less clear. A recent Gartner study involving 350 global executives from companies earning at least $1 billion annually discovered that although 80% of those who tested AI or autonomous tech reported staff reductions, businesses eliminated positions irrespective of whether the technology was yielding actual returns. Essentially, layoffs linked to AI may be more performative than effective. Numerous business leaders, including OpenAI CEO Sam Altman, suggest that companies are engaging in “AI washing,” using the tech as a convenient excuse for downsizing when, in truth, the cuts are driven by other financial issues.

The cost of tokens is climbing

Levie provides an illustration of how this AI predicament manifests in the office environment.

“‘Look, I generated a contract,’” he suggested an employee might claim. He went on: “But you didn’t verify all the terms before sending it to the counterparty, nor did you integrate all previous contracts to ensure compatibility.” A significant gap may still exist between the perceived final product and the effort needed to reach that goal.

This is becoming an increasingly costly factor as AI expenses are accumulating. Microsoft recently transitioned from Claude to GitHub Copilot CLI. Uber CTO Praveen Neppalli Naga disclosed to The Information that the company exhausted its entire 2026 budget for AI coding tools in merely four months. Additionally, a client of an AI consultant interviewed by Axios reported accidentally spending half a billion dollars after neglecting to set usage limits on employee Claude licenses.

Certainly, although AI executives remain confident that the technology will be profitable—investing nearly $700 billion in AI infrastructure development—prominent figures like Sam Altman and Anthropic CEO Dario Amodei have walked back their earlier assertions regarding AI’s capacity to replace significant portions of white-collar roles.

“I’m delighted to ⁠be wrong about this,” Altman remarked during a recent AI interview with Commonwealth Bank of Australia CEO Matt Comyn. “I believed there would have been a greater impact on the elimination of entry-level white-collar jobs by now than ​has actually occurred.”

Nevertheless, technology companies continue to reduce their staff. Website builder Wix announced on Thursday that it had let go of 1,000 employees, representing 20% of its workforce, partly due to AI-driven efficiencies. Meta recently cut 10% of its staff, with CEO Mark Zuckerberg stating in a memo that “AI is the most consequential technology of our lifetimes.” According to outplacement firm Challenger, Gray & Christmas, 49,135 layoffs have been attributed to AI year-to-date, nearing the 55,000 total AI-related layoffs recorded in 2025.

Levie provides guidance to CEOs on how to better understand the current state of AI capabilities.

“Use AI a *ton,*” he advised, “to determine the true implications of agents within the enterprise, and emerge with an understanding of both the benefits and the actual effort required to implement them.”