80,000 Tech Workers Lost Their Jobs in 3 Months. Nearly Half Because of AI.

The first quarter of 2026 is now over. The numbers are in. And they are the worst since the post-pandemic collapse of 2023.

According to a report by Nikkei Asia, citing analysis from RationalFX, 78,557 workers in the tech industry lost their jobs between January 1 and April 2026. That figure alone would be significant. What makes it historically unusual is the breakdown of why.

Of those 78,557 layoffs, 37,638 — exactly 47.9%, nearly half — were directly attributed to AI implementation and workflow automation. Not restructuring. Not poor quarterly results. Not a post-pandemic correction. AI.

Bloomberg reported that tech employers announced 18,720 job cuts in March alone — a 24% increase from March 2025 — bringing the first-quarter total to the highest since 2023. The trend is not slowing. Analysts project the full-year total could approach 265,000 global tech layoffs if the current pace holds.

76.7% of those cuts happened in the United States.

The names attached to these numbers are not abstractions. Oracle sent a 6 AM email on April 1 — April Fool’s Day — informing an estimated 10,000 to 30,000 employees that their final working day was immediate. Block eliminated 4,000 jobs — 40% of its workforce — with CEO Jack Dorsey explicitly citing AI capability as the reason. Amazon cut 14,000 corporate roles. Microsoft had already cut 9,000 in 2025.

Each company had its own specific rationale. But the common thread running through Q1 2026 is the one that Bloomberg described without equivocation: AI adoption is catalysing leaner staffing levels, and the technology sector is moving fastest.

There is a debate worth taking seriously about how much of this is genuine AI displacement versus AI being used as cover for cuts that were coming anyway.

Sam Altman, the CEO of OpenAI — whose technology is cited as a direct driver of many of these layoffs — said recently that there is “some AI washing where people are blaming AI for layoffs that they would otherwise do.” Babak Hodjat, Chief AI Officer at Cognizant, told Nikkei that the layoffs may reflect “the expectation that AI will improve productivity, rather than actual data reflecting this.” His estimate: it will take another six to twelve months before companies start seeing real productivity gains.

This is a meaningful caveat. It is also cold comfort to 37,638 people who have already lost their jobs.

The distinction between “AI caused this layoff” and “AI was used as a convenient justification” does not change much for the person who received the 6 AM email. The job is gone either way.

Anthropic’s research published in March 2026 provides the clearest picture yet of which tech roles are most exposed. Computer programmers have the highest observed AI exposure of any occupation measured: 74.5% of their daily tasks are already being performed or assisted by AI in professional settings. Customer service representatives sit at 70.1%. Data entry workers at 67%.

These are not fringe roles. They are the backbone of the tech workforce as it existed five years ago. They are also, increasingly, the roles that are disappearing.

What is replacing them is not nothing. IBM announced this week that it has tripled its entry-level hiring in 2026 — a notable exception to the general trend, and a signal that some companies understand the long-term risk of eliminating the pipeline that produces future senior workers. But IBM is one company moving against a tide that is running strongly in the other direction.

Last week Goldman Sachs put a number on the net effect: 16,000 net jobs lost per month across the US economy, with entry-level workers and those under 30 bearing the sharpest impact. The Q1 tech figures are consistent with that estimate and, in some months, exceed it.

The honest question for anyone working in tech — or considering a career in tech — is not whether AI is affecting the sector. It clearly is, in numbers that are now too large to dismiss or contextualise away. The question is what the Q1 data actually tells you about where the risk is concentrated.

The companies cutting hardest are eliminating roles that involve primarily structured, repeatable work: customer support, data processing, routine coding, administrative functions. The roles that involve genuine system architecture, security, complex debugging, and the kind of engineering judgment that requires years of contextual experience are less affected — and in some cases, more in demand as AI tools make skilled engineers dramatically more productive.

The structural problem, which the IBM exception highlights rather than resolves, is that the entry-level and mid-level roles being eliminated were the training ground for the senior roles that remain. You cannot skip the years of foundational work and arrive directly at the judgment required for complex engineering. We explored this in detail in The Last Junior.

The Q1 numbers make that problem more acute, not less. 37,638 people absorbed the first wave of genuine AI displacement in the tech sector this year. If the pace holds, the second wave will be larger.

There is one thing worth saying clearly, because the coverage of these numbers tends to oscillate between panic and reassurance without landing anywhere useful.

The Q1 figures are significant but not catastrophic at a macroeconomic level. 78,557 tech layoffs represents a small fraction of the total US workforce. The unemployment rate for workers over 25 in AI-exposed occupations has not risen dramatically. The economy is not in free fall.

What is happening is structural and directional rather than sudden and total. The floor beneath certain categories of tech work is being lowered, slowly but measurably, and the people most exposed are those doing the most routine work at the most junior levels. This is not the end of tech employment. It is a sustained reorganisation of who gets hired, for what work, at what price.

Understanding that reorganisation clearly — before the layoff email arrives at 6 AM — is the only practical response to it.