45,000 Tech Workers Lost Their Jobs to AI in March 2026. The Ones Who Kept Theirs Had This in Common.

Block laid off 40% of its workforce last week. Jack Dorsey said the quiet part out loud: "Within the next year, I believe the majority of companies will reach the same conclusion." Atlassian cut 1,600 people — then immediately posted 800 AI-focused roles. Meta is planning to axe up to 16,000.

The numbers for March 2026 are brutal. Over 45,000 tech workers lost their jobs, with roughly 9,200 of those cuts explicitly attributed to AI. If you're in tech, you've probably refreshed LinkedIn three times today already.

But here's the thing nobody's talking about: the layoffs aren't the whole story.

The Klarna Problem

Remember when Klarna was the poster child for AI replacement? They cut 700 customer service agents, replaced them with AI, and their CEO went on a victory lap. Efficiency up. Headcount down. Shareholders thrilled.

Then quality tanked. Customer satisfaction cratered. And quietly, Klarna started rehiring humans.

They're not alone. According to Orgvue's 2026 workforce report, 55% of companies that replaced workers with AI now regret the decision. Not because the AI didn't work — it did, technically — but because "working" and "delivering value" turned out to be very different things.

Gartner is now predicting that 50% of AI-attributed layoffs will result in rehires by 2027. That's not a prediction about AI failing. It's a prediction about companies learning, the hard way, that automation and augmentation are fundamentally different strategies.

Two Playbooks, Two Outcomes

The data is starting to paint a clear picture, and it splits cleanly into two camps.

Camp 1: Replace humans with AI. Cut headcount, automate workflows, celebrate the savings. Short-term wins on the balance sheet. Long-term problems everywhere else — quality drops, institutional knowledge evaporates, and the remaining team burns out trying to cover gaps the AI can't.

Camp 2: Augment humans with AI. Keep the people, give them better tools. According to Harvard Business Review's latest workforce study, companies taking this approach are seeing 3x performance improvements and 38% higher revenue growth compared to the replacement crowd.

The difference isn't subtle. It's not a rounding error. It's a completely different trajectory.

What the Survivors Have in Common

Here's where it gets interesting for individual workers. PwC's Global AI Jobs Barometer found that workers with AI skills command a 56% wage premium over their peers. Not 5%. Not 15%. Fifty-six percent.

The people keeping their jobs — and getting paid more for them — aren't the ones ignoring AI. They're the ones who learned to work alongside it.

HBR's research breaks it down further: AI is cutting about 17% of pure automation roles (data entry, basic reporting, routine processing). But it's simultaneously increasing demand by 22% for what they call "collaboration roles" — positions where humans direct, refine, and quality-check AI output.

The jobs disappearing are the ones that look like assembly lines. The jobs growing are the ones that look like partnerships.

The Entry-Level Crisis Nobody's Addressing

There's a darker angle to this story that deserves attention. The Dallas Federal Reserve published data showing that entry-level workers aged 22-25 are seeing a 13% employment decline in tech-adjacent roles. These are the junior positions that companies are most aggressively automating.

This creates a genuine pipeline problem. If you automate away the entry-level roles, where do your senior people come from in five years? The companies cutting junior positions today are borrowing against their own future talent pool.

The Augmentation Playbook

So what does "working with AI" actually look like in practice? It's not just "use ChatGPT sometimes." It means:

• Directing AI output rather than accepting it raw. The person who can prompt, evaluate, and refine AI work is worth more than the AI alone.
• Maintaining quality standards that AI can't self-assess. Klarna learned this one the expensive way.
• Building institutional knowledge that makes AI more effective over time. An AI assistant that remembers your processes, your clients, your edge cases — that's augmentation.
• Handling the exceptions that break automated workflows. Every business has them. The 20% of cases that take 80% of the judgment.

This is exactly the thesis behind what we're building at Geta.Team. AI employees that work alongside you — not instead of you. They remember context, handle routine tasks, and free you up for the work that actually requires a human brain. The goal was never to replace your team. It was to make your existing team unreasonably productive.

The Uncomfortable Math

Here's the bottom line: 45,000 people lost their jobs in tech this month. That number will likely grow. But the companies doing the cutting are increasingly discovering that wholesale replacement doesn't deliver what the spreadsheet promised.

The workers who thrive through this aren't the ones running from AI. They're the ones running toward it — learning to direct it, evaluate its output, and use it as leverage rather than competition.

The next twelve months will separate companies (and careers) into two categories: those that figured out augmentation, and those that learned why replacement doesn't scale.

Jack Dorsey thinks most companies will "reach the same conclusion." He's probably right. The question is which conclusion they reach — and how much it costs them to get there.