Jack Dorsey just cut 4,000 employees from Block - nearly half the company’s workforce - and said the quiet part out loud: “Within the next year, the majority of companies will reach the same conclusion and make similar structural changes.”
The Block announcement capped a brutal February for workers. The layoff tracker at TrueUp counts 49,318 tech workers cut in 2026 so far, averaging 850 people per day. And increasingly, companies are pointing to AI as the reason.
But here’s what the headlines miss: the picture is far more complicated than “AI takes jobs.” Some roles are disappearing. Others are paying more than ever. And a surprising number of those “AI layoffs” might end up rehired - just not on the same terms.
The Big Cuts This Month
Block (February 26): The payments company behind Square and Cash App slashed from 10,000 to under 6,000 employees. Dorsey’s memo insisted the business was “strong” and profits were growing - this wasn’t about survival, but about betting that AI makes smaller teams more productive. Investors agreed, sending the stock up 24% after hours.
Baker McKenzie (February): The global law firm announced cuts of 600 to 1,000 employees - roughly 10% of its workforce. The firm cited AI in its restructuring statement, targeting support roles in research, know-how, marketing, and secretarial work. Legal industry insiders are watching closely - this may give other BigLaw firms permission to make similar moves.
Meta (January-February): Reality Labs shed 1,500 positions, about 10% of the division, as Meta pivots from VR headsets to smart glasses and AI. After $80 billion in cumulative metaverse losses since 2020, Zuckerberg is redirecting spending.
Commonwealth Bank: Australia’s largest bank cut 300 jobs, explicitly preparing workers for what they called an AI-driven “shift.”
Pinterest and Dow: Both companies announced layoffs attributing the cuts “in part” to AI adoption.
Is AI Actually Doing the Work?
Here’s the twist that deserves more attention: Harvard Business Review reports that companies are laying off workers because of AI’s potential, not its actual demonstrated performance. They’re making bets on what AI will be able to do, not what it’s proven to do.
And according to Forrester Research, half of those AI-attributed layoffs will be “quietly rehired - but offshore or at significantly lower salaries.” The jobs don’t disappear. They get restructured, relocated, and repriced.
Block’s severance package offers some insight: 20 weeks salary plus one week per year of tenure, six months healthcare, and $5,000 transition assistance. Generous by tech standards - which suggests even Dorsey knows this isn’t as clean as “AI does it now.”
Who’s Actually Hiring
Despite the bloodletting, AI-related job postings are growing even as overall tech hiring weakens. Indeed’s January data shows jobs mentioning AI continue to rise while broader hiring slows.
The hottest roles right now:
- MLOps Engineers: $160,000 to $350,000+. Keeping AI systems running in production is harder than building them.
- AI Data Annotation and Labeling: Up 154% year-over-year. Models need clean data, and someone has to label it.
- NLP Specialists: 155% increase in job postings. Voice interfaces and document processing drive demand.
- Agentic AI Specialists: Designing autonomous AI workflows, £100k-180k in the UK.
- AI Prompt Engineers: $111,000 to $170,000 average, with top earners hitting $250,000+.
Here’s the salary bump that matters: candidates with AI skills earn 23% more on average than comparable candidates without them, according to World Economic Forum research. They’re also 8-15% more likely to get interview invitations.
What Skills Actually Matter
Forget the breathless takes about prompt engineering being a gold rush. The skills paying off in 2026 are mostly the same ones that paid off before - plus familiarity with AI tools:
Technical skills in demand:
- Deep learning frameworks (TensorFlow, PyTorch)
- Natural language processing
- Computer vision
- MLOps and deployment
- Traditional software engineering (still essential)
The real differentiator: Domain knowledge plus AI familiarity. A financial analyst who can use AI tools effectively beats a “prompt engineer” who doesn’t understand finance. Same for legal, healthcare, and other fields.
Entry-level pathways are narrowing - recruiters consistently report that junior roles are hardest to fill in the traditional sense, because companies want AI-augmented productivity from day one.
What This Means
Three things are happening simultaneously:
-
Companies are using AI as cover. Some layoffs attributed to AI are really about cutting costs, reshoring, or restructuring. The AI excuse sounds better than “we overhired in 2021.”
-
Genuine displacement is beginning. Support roles - research, documentation, basic analysis - are genuinely being compressed. Not eliminated entirely, but needing fewer people.
-
New roles are emerging. But they require different skills, and the people being laid off from support roles aren’t automatically qualified for MLOps positions.
The uncomfortable truth: 55% of hiring managers surveyed expect layoffs in 2026, and 44% expect AI to drive them. This isn’t a one-month story. It’s the new normal.
What You Can Do
If you’re worried about your job:
- Learn the tools your company uses. Being the person who actually knows how to get value from AI tools beats being the person AI is supposed to replace.
- Document your domain expertise. The stuff in your head about how things actually work is harder to replicate than you think.
- Watch for restructuring signals. When companies start talking about “efficiency” and “AI transformation,” that’s the time to update your resume, not after the layoffs hit.
If you’re hiring:
- Be honest about what AI can actually do. Overpromising and underdelivering on AI productivity gains is expensive.
- The roles that matter are different. You probably need fewer content writers and more people who can keep AI systems running reliably.
February 2026 won’t be remembered as the month AI took everyone’s job. But it might be remembered as the month companies stopped being shy about saying that’s where they think this is headed.