Cloudflare Cuts One-Fifth of Its Workforce (Over 1,000 People): The CEO's Theory on Who Risks with AI
Cloudflare has laid off over 1,100 employees in May, about 20% of its workforce. In the following weeks, its engineering team grew by 45%, increasing from 1,308 to 1,894 units, according to an analysis by BNP Paribas based on LinkedIn profiles. This data was confirmed by CEO Matthew Prince, who used it as proof of a precise thesis: AI does not indiscriminately thin out staff, but rather targets a well-defined category of work.
Prince divides the company into three groups, reviving a distinction made by Peter Drucker in "The Practice of Management" from 1954: those who build the product (builders), those who generate revenue (sellers), and those who measure, track, coordinate, and supervise the work of the other two (measurers). The roles that were cut mostly belong to the last category: middle managers, operational staff, financial analysts, and marketing coordinators.
"The vast majority of those we laid off last week were measurers," he stated.
The logic, according to Prince, is that AI excels precisely in the work of measurers: "If you think about what AI is most effective at, it’s looking at sets of data and summarizing them." Hence the prediction, expressed in a piece on the Wall Street Journal picked up by Fortune, that many support roles "will not be the roles that drive companies in the future." He notes that sellers with revenue shares were excluded from the cuts, as he is committed to hiring more engineers if AI makes them more productive, not less.
A Record Quarter and a Falling Stock
The financial scenario complicates the linear narrative. In the first quarter of 2026, Cloudflare reported revenues of $639.8 million, a 34% year-over-year increase and the highest in its history, but closed with an operating loss of $62 million. The day after the results were published, the stock fell by about 24%. On the internal front, the use of AI within the company grew by over 600% in the three months preceding the announcement, and all code produced with AI is now reviewed by autonomous agents.
A Pattern Prince Expects to See Again
Prince anticipates that the same mechanism will repeat across the sector, and some indications point in that direction. In May, GitLab cut 7% of its staff, eliminating up to three levels of management and reorganizing into 60 autonomous teams—a move that CEO Bill Staples described as preparation for the agentic era. Meanwhile, the monitoring platform TrueUp reports that tech job openings have grown by 14% compared to the previous year, with hardware engineering up 52%, while operations, HR, and general management have decreased. Cloudflare, for its part, reports a record number of open positions and an expected total workforce higher in 2027 than at any point in 2026.
However, the AI-centric reading remains contested. The operating loss for the quarter and the decline in margins leave room for a more ordinary explanation: a restructuring of costs after years of aggressive hiring, making the cuts to measurers also a more presentable way to narrate it. There is also a limit to the numbers themselves: the 45% increase in engineering relies on an analysis of LinkedIn profiles and is not independently verifiable.