Four Years of Tech Layoffs Had Four Different Causes

"I got this wrong." That was Tobi Lütke, in the memo that went out with Shopify's layoffs in mid-2022. It's still one of the most candid lines of the whole correction.

Between Meta's first-ever mass layoff in November 2022 and Amazon's largest-ever one in October 2025, the industry cut more than half a million jobs. People talk about all of it as one event, "the tech layoffs." We don't think the data supports that. Sort the cuts by cause and you get four separate waves, with different reasons and different victims. If you want to know what a given layoff means for you, the wave it belongs to tells you more than its headcount.

Mid-2022: the pandemic bets unwound

The first companies to cut were the pandemic darlings. Netflix lost subscribers. Peloton's bikes stopped selling. Shopify had bet that the e-commerce surge would last forever, and that bet came apart, while Coinbase walked straight into crypto winter. These were ordinary demand shocks at businesses that had scaled for a world that partly went back to how it was.

What got underrated at the time was how narrow this wave was. The market punished particular bets, and tech as a whole was mostly fine. Companies with normal demand curves kept hiring through most of 2022.

Late 2022 into 2023: big tech cut what it over-hired

Next came the big names, one after another. Meta cut 11,000. Amazon cut 18,000 and Google 12,000. Then Microsoft cut 10,000 and Salesforce 8,000. Inside about ninety days, nearly every major tech company let go of somewhere between 5 and 13% of its staff.

The cause was arithmetic. Big tech had grown headcount by 40 to 60% during the pandemic years, on the assumption that pandemic-era growth was the baseline from here on. Then interest rates rose, the free money stopped, and investors started rewarding "efficiency" instead of growth. The extra people were the first thing to go.

Two things from that winter still shape the market. Elon Musk cut half of Twitter and the product kept running, which every executive noticed. And the stock market rewarded almost every layoff announcement, so CEOs learned that investors will applaud a cut.

2024: same headcount, different people

By 2024 the headlines shrank and the cuts got smaller, more frequent and more targeted. SAP restructured 8,000 roles, "in particular [for] business AI." Intuit cut 1,800 people and hired the same number back into AI-focused jobs. Cisco cut twice to pay for moves into AI and security, and Dropbox cut 20%, citing an "AI transition." Total headcount at these companies barely moved. Who held the jobs changed a great deal.

Most analysis missed this phase. The story stopped being "tech is shrinking" and became "tech is swapping people." The layoff and the hiring spree happen at the same company in the same quarter. That's how record layoffs and strong demand for AI, security and data skills could both be true at once.

2025: companies started saying AI out loud

In 2025 the euphemisms fell away. CrowdStrike's CEO wrote that AI "flattens our hiring curve." Salesforce cut 4,000 support roles and credited its own AI agents. Accenture told staff who couldn't be reskilled for AI that they would be "exited." Amazon framed its largest-ever layoff around getting ready for the "AI era." TCS cut 12,000, the largest layoff on record in Indian IT, aimed at skills that no longer matched AI-led delivery.

The thing that sets this wave apart is that the companies doing the cutting are healthy. Profits are at records and revenue is growing while headcount shrinks, a combination that was rare in every previous tech cycle.

That's why we read this phase as structural rather than cyclical. There's no recovery to wait for, because by the companies' own measure nothing is broken.

So which wave are you exposed to? If your risk looks like 2022 or early 2023 (an employer that over-extended, or demand settling back to normal), the old advice still holds. Keep your fundamentals sharp and move to a healthier company.

If your risk looks like 2024 or 2025, you've got a different problem. Your employer is doing fine, and your tasks are being absorbed anyway. Changing companies won't help much, since the company across the street is making the same moves. What helps is moving from doing the work AI now does to directing it, reviewing it, securing it and building on top of it.

We don't expect the correction to end so much as settle in. Flat headcount on rising revenue is turning into the standard business model, and the sooner you plan for that market, the more options you'll keep. A lot of people are still planning for the 2021 one. It isn't coming back.