Box founder warns of 'AI psychosis' as tech layoffs surge in 2026
As ClickUp axes 22% of staff and 2026 tech layoffs near last year's total, Box's Aaron Levie says bosses are cutting jobs they barely understand. India's IT giants are next.
The News
Aaron Levie, the founder of cloud firm Box, has given a name to a pattern he sees spreading through corporate boardrooms: "AI psychosis." Speaking on a TechCrunch podcast published on 29 May 2026, Levie argued that the executives deciding which jobs artificial intelligence can absorb are often the people furthest removed from the daily reality of those roles. The result, he suggested, is confident cost-cutting built on a shaky grasp of what workers actually do.
His warning lands during a punishing stretch for technology staff. Layoffs across the sector in 2026 are already close to matching the entire total for 2025, and employers increasingly frame the cuts as an automation story rather than a budget story. Project-management firm ClickUp recently shed 22% of its workforce, pinning the reduction on AI agents that now handle tasks once owned by people.
The same discussion tracked where investment is still flowing. Snowflake signed a six-billion-dollar, five-year agreement with AWS. Logistics group Stord raised 250 million dollars at a three-billion-dollar valuation. Model-routing startup OpenRouter pulled in 113 million dollars, doubling its valuation to 1.3 billion dollars inside a single year. DuckDuckGo, meanwhile, reported a 30% rise in installs as users resisted Google weaving AI into search by default.
Why It Matters
The picture is not a market in retreat. It is one spending heavily on the infrastructure layer while quietly thinning the human one, faster than anyone can prove the trade works. There is little independent evidence yet that AI agents fully replace eliminated roles, rather than absorbing a slice while the rest migrates to whoever remains. When a chief executive announces a fifth of staff gone "because of AI," the claim doubles as a productivity boast to investors.
History offers a caution. During the 2000 to 2002 downturn, firms that gutted teams on the promise of web-era efficiency often spent the following upturn rehiring at a premium for the knowledge they had discarded. The picks-and-shovels enthusiasm behind the Snowflake, Stord and OpenRouter cheques mirrors that earlier faith in the platform layer. The open question is whether the headcount logic stacked on top survives reality.
Indian Angle
For Indian readers the stakes are unusually direct. The country's IT services giants, from TCS and Infosys to Wipro and HCLTech, employ several million people whose work sits squarely in the path of agentic automation: testing, support, maintenance and routine coding. Entry-level hiring has already cooled, and the "AI psychosis" framing maps neatly onto an industry whose leadership is under pressure to show AI savings to global clients.
The global capability centres that multinationals run in Bengaluru, Hyderabad and Pune face the same crosswind. They have been a quiet engine of white-collar job creation, yet the ClickUp template, cut deep and credit the agents, is exactly the playbook their parent companies may import. Indian policymakers and NASSCOM will be watching whether 2026 marks a structural shift or a cyclical scare.
An upside is worth naming. Indian model builders such as Sarvam and Krutrim sit on the supply side of the same boom funding OpenRouter and Snowflake. If India moves up from executing the automation to designing it, the disruption Levie describes becomes an opening, not a threat.
FAQ
What does "AI psychosis" actually mean here?
It is Aaron Levie's label for a specific blind spot: senior leaders confidently deciding AI can replace a role while having little firsthand understanding of what that role involves. It is a critique of decision-making, not a claim that the technology itself is faulty or overhyped.
How bad are the 2026 tech layoffs?
By late May 2026, job cuts across the technology sector were already nearly matching the full-year total for 2025. A growing share are being attributed to AI agents, with ClickUp's 22% workforce reduction cited as a prominent example of the trend.
Is AI investment actually slowing down?
No. The same period saw Snowflake commit to a six-billion-dollar AWS deal, Stord raise 250 million dollars at a three-billion-dollar valuation, and OpenRouter raise 113 million dollars at a 1.3-billion-dollar valuation. Spending is concentrating in infrastructure even as headcount falls.
What should Indian IT workers take from this?
Roles centred on testing, support and routine coding face the most exposure, and entry-level hiring has already softened. The longer-term hedge is moving up the value chain toward designing and deploying AI systems, where Indian engineers and model builders are well placed to compete.
This story was reported by TechCrunch. Read the full original coverage at TechCrunch.
Sources & Citations
Frequently Asked Questions
What does "AI psychosis" actually mean here?
It is Aaron Levie's label for a specific blind spot: senior leaders confidently deciding AI can replace a role while having little firsthand understanding of what that role involves. It is a critique of decision-making, not a claim that the technology itself is faulty or overhyped.
How bad are the 2026 tech layoffs?
By late May 2026, job cuts across the technology sector were already nearly matching the full-year total for 2025. A growing share are being attributed to AI agents, with ClickUp's 22% workforce reduction cited as a prominent example of the trend.
Is AI investment actually slowing down?
No. The same period saw Snowflake commit to a six-billion-dollar AWS deal, Stord raise 250 million dollars at a three-billion-dollar valuation, and OpenRouter raise 113 million dollars at a 1.3-billion-dollar valuation. Spending is concentrating in infrastructure even as headcount falls.
What should Indian IT workers take from this?
Roles centred on testing, support and routine coding face the most exposure, and entry-level hiring has already softened. The longer-term hedge is moving up the value chain toward designing and deploying AI systems, where Indian engineers and model builders are well placed to compete.