Why Big Tech Is Using AI to Justify Mass Layoffs in 2025–2026
Something shifted in 2025. Tech layoffs weren’t new — they’d been rattling through Silicon Valley since the post-pandemic correction of 2022. But the language changed. Companies stopped saying “restructuring” or “cost optimisation.” They started saying AI.
The numbers are staggering. According to data from TrueUp’s Tech Layoffs Tracker, over 245,000 tech employees were laid off at 783 companies in 2025 alone — roughly 674 workers every single day. And 2026 has opened even faster, with approximately 52,000 jobs cut in the first two months at a rate of 839 people per day.
The question isn’t whether layoffs are happening. It’s whether AI is genuinely causing them — or simply being used as a convenient cover.
The Numbers Behind the Wave
To understand the scale, consider the major players. Intel planned to reduce its workforce from around 109,000 to 75,000 by the end of 2025 — a cut of roughly 34,000 roles. Microsoft eliminated around 15,000 positions. Amazon cut 14,000. Verizon 13,000. HP between 4,000 and 6,000. Salesforce’s CEO, Marc Benioff, publicly credited AI tools with handling work previously done by 4,000 customer support staff.
The pattern isn’t random. These aren’t companies struggling financially. Most are posting strong earnings. What they share is a narrative: AI is letting us do more with fewer people, so we are doing exactly that.
“Unlike earlier layoff waves driven by over-hiring, many of 2025’s reductions were permanent, with entire roles eliminated as companies rebuilt around AI-first operating models.” — RationalFX analysis
The Real Story Is More Complicated
Not everyone is convinced the AI narrative is the full picture. An Oxford Economics report released in early 2026 found that many layoffs CEOs labelled as “AI-related” were actually the result of post-pandemic over-hiring finally being corrected. The tech sector hired aggressively in 2020 and 2021 during a boom that wasn’t sustainable. The correction was always coming.
Still, even skeptics acknowledge that AI is accelerating what might have been a slower, more gradual correction. When a company can demonstrably automate 30% of its customer support tickets, the justification for large support teams becomes harder to maintain, whatever the board’s original motivation.
Which Roles Are Actually Being Replaced
The layoffs aren’t evenly distributed. Based on disclosed data from 2025’s cuts, the roles most affected include:
- Customer support and service — most directly replaced by AI chat and voice tools
- Mid-level product management — roles focused on documentation and coordination
- Marketing operations — content production, social scheduling, copywriting
- Data entry and processing — almost entirely automated
- Some software engineering tiers — junior roles where AI coding tools now produce working code
What’s notably not being cut at the same rate: senior engineering, product strategy, AI infrastructure roles, and anything requiring deep contextual judgment or client relationships. The pattern is consistent — AI is replacing execution, not strategy.
What Comes Next
Anton Korinek, an economist who studies AI’s economic impact, offered a sobering take in early 2026: “Whereas the job market effects of AI in 2025 were still quite ambiguous, AI capabilities have advanced rapidly. This may be the beginning of a new trend where white-collar jobs become threatened more seriously by AI. Once a few companies start the trend, competitive forces may induce others to follow suit.”
Jack Dorsey, whose company Block cut nearly 4,000 jobs in February 2026, was even more direct in a shareholder letter: “I think most companies are late. Within the next year, I believe the majority of companies will reach the same conclusion and make similar structural changes.”
Whether that’s prescient or self-serving depends on your perspective. What’s not in dispute is that the tech labour market has fundamentally changed — and the workers most at risk are those whose roles are most easily described to an AI in a sentence or two.