The 2026 layoff wave does not fit the usual recession script. Job cuts across the technology sector have already passed 100,000 this year, by the count of trackers such as TrueUp, with some projections pointing toward 370,000 before the year is out. What unsettles people is not the scale alone but the context. Many of the firms doing the cutting are profitable and growing. They are not shrinking because they have to. They are shrinking because they have decided to spend the money elsewhere.
Payroll out, compute in
Meta is the clearest example. The company is preparing to spend more than 100 billion dollars in 2026 on AI data centres and related hardware while laying off around 8,000 workers and redirecting roughly 7,000 staff toward AI focused roles. The pattern repeats across the sector. Cisco's chief executive Chuck Robbins described the company as investing in AI infrastructure to avoid being left behind, language that has become a near universal justification for trimming traditional teams. The capital that used to fund headcount growth is being routed into chips, power and floor space.
How much of this is really AI?
Not all of it, and the industry knows it. There is mounting evidence of what critics call AI washing, where firms attribute job cuts to artificial intelligence when the real drivers are overhiring, slowing revenue or pressure from investors. OpenAI's Sam Altman has acknowledged both sides at once, conceding there is some AI washing where people blame AI for layoffs they would have made anyway, while insisting that genuine displacement is also under way. Both things can be true, and that ambiguity is convenient for executives who would rather cite a futuristic cause than a mundane one.
Where the work is still landing
The destruction is uneven. Roles in machine learning infrastructure, AI safety, applied research, healthcare and skilled trades remain in strong demand even as programming, customer service, data entry and routine content work contract. That split explains why the same companies can announce thousands of redundancies and thousands of new requisitions in the same quarter. The headcount is not so much falling as being rebuilt around a different centre of gravity.
The risk in the trade
Trading people for compute is a bet that the productivity unlocked by AI will more than replace the work that walked out the door. For the firms with genuine demand for their models, it may pay off handsomely. For those simply following the herd and dressing up ordinary cost cutting in the language of transformation, the danger is that they have thinned their organisations faster than the technology can actually backfill. The next few quarters will start to separate the two.
