Cloudflare’s engineering headcount has increased by 45% after cutting 1,100 jobs, and its CEO has an outline of who will survive in AI.

The number of Cloudflare engineers increased by 45 percent in the last few weeks after the company cut 1,100 jobs in May, according to BNP Paribas data taken from LinkedIn profiles. The findings, first reported by Business Insider, show that Cloudflare’s engineering workforce grew from 1,308 to 1,894 as its headcount fell by five. CEO Matthew Prince confirmed this trend and gave a framework to understand it: every company, he said, is made up of builders, sellers, and estimators, and AI eliminates the third party.

Prince told Business Insider that the difference is straightforward. Manufacturers create the product, salespeople bring in the revenue, and appraisers track, report, and coordinate the work of the first two teams. The roles being cut at Cloudflare, and across the tech industry, fall largely in the measurement category: middle managers, operations staff, financial analysts, and marketing consultants whose AI agents can now measure.

“If you think about what AI excels at, looking at data sets and summarizing them,” said Prince. He added that if his engineers work more with AI, he will hire more of them, not fewer. The idea is that AI increases the output of people who build and sell but replaces those whose main job is to supervise and report.

The BNP Paribas analysis, which Business Insider says the Prince reviewed and confirmed, could not be independently verified without that report. LinkedIn profile data captures job title changes and may not accurately reflect internal accounting. But the direction is consistent with Prince’s stated strategy: Cloudflare has cut back and invested heavily in the work it considers most important.

The pattern is not unique to Cloudflare. TrueUp, a platform that tracks tech employment, reports that open technology roles are up 14 percent in 2026 compared to last year, and hardware engineering positions are up 52 percent. Gains have been concentrated in technology and manufacturing jobs, while job openings, human resources, and general management have declined.

Companies hire many developers and few manage developers. GitLab followed a similar playbook in May, cutting seven percent of its workforce and shedding three layers of management while reorganizing its engineering division into 60 independent teams. CEO Bill Staples called the preparations “agent period,” and leaner companies are not reducing their engineering capabilities but focusing on them.

Prince’s outline carries a consistent warning for anyone whose job description focuses on communication, reporting, or process management. He was modest about the trajectory, telling Business Insider that “most support roles will not be roles that drive companies forward.” If your work can be described as a rating produced by other people, the category you are in is the one that AI targets first.

Broader labor market data complicates the picture a bit. Tech CEOs have recently moved away from warning about AI job losses but stressing that AI will create jobs, a pivot that coincides with the impending IPOs of companies including OpenAI and Anthropic. Prince’s framework sits somewhere between these two narratives: he is not saying that AI is creating jobs across the board, but that it is creating engineering jobs directly, at the expense of everyone else.

Whether the builder-vendor-scale model holds beyond Cloudflare is an open question. Not every measurement function is viable, and not every company can pull off a 45 percent increase in engineering while cutting a fifth of its workforce. The framework also assumes that AI tools are reliable enough to replace human judgment in supervisory tasks, an assumption that is still debated even among AI researchers.

What is not disputed is the direction of employment. Cloudflare’s Q1 2026 revenue grew 34 percent year over year to $640 million, and the company added a record number of enterprise customers as it shed 1,100 roles. The restructuring was not driven by financial weakness but a bet that the work that those people do can now be done with software, and that the savings are better spent on developers who write more.

Prince’s taxonomy gives a name to a change made by a number of companies at the same time but rarely defines this clearly. The question for the tens of thousands of workers being laid off across the technology sector this year is “a measurer” is a temporary label applied to roles that will eventually return, or a permanent decision throughout the work phase.