Cloudflare Fired 1,100 People and Called It the Agentic Era

There is a particular kind of insult built into laying people off right after a strong quarter and then describing the move as strategic evolution. It is not the oldest move in tech, but it might be the cleanest one. If a struggling company cuts jobs, at least the story is honest about panic. If a growing company cuts more than a thousand people while telling the market this is what the future requires, the story changes. It stops being about survival and starts being about permission.

Cloudflare made that shift explicit. In the company’s own May 7 post, co-founders Matthew Prince and Michelle Zatlyn said Cloudflare was reducing its workforce by more than 1,100 employees globally. They also said the company’s usage of AI had increased by more than 600% in the previous three months and that teams across engineering, HR, finance, and marketing were now running thousands of AI agent sessions each day. The key line was the one they probably hoped would sound clarifying: these actions were “not a cost-cutting exercise” but part of defining how a world-class, high-growth company creates value in the agentic AI era.

That sentence should bother more people than it probably will. Cloudflare’s Q1 2026 results show revenue of $639.8 million, up 34% year over year. TechCrunch framed the contradiction exactly right: record-high revenue, 1,100 jobs gone anyway, AI supplied as the rationale. This is the part of the AI transition that gets dressed up as inevitability because inevitability sounds cleaner than choice.

There is a version of the management argument that is worth taking seriously. If internal tooling really has changed how teams work this quickly, leadership would be irresponsible not to rethink org design. Processes that required larger teams six months ago may genuinely require fewer people now. Some of those savings may be real. Cloudflare also appears to be handling severance more generously than many companies do, promising base pay through the end of 2026 and extended healthcare support in the U.S. That matters. How you treat people on the way out is not trivial.

But generosity during layoffs does not make the underlying logic any less revealing. What Cloudflare is really saying is that productivity gains no longer need to show up first as shorter cycles, better products, lower prices, or saner workloads. They can show up as fewer humans. AI becomes the acceptable language for that move because it sounds modern, inevitable, and efficiency-coded. It is easier to tell the market you are architecting for the agentic era than to say directly that once software can plausibly do more work, labor loses bargaining power almost immediately.

This is why the old argument that AI will mostly just help workers feels thinner every month. Sometimes it will. Sometimes it already does. But as I wrote in AI Isn’t Cheaper Than Hiring People, the economics were never the only story. Cost is one part of it. Control is the other. Tools that make output more measurable, workflows more standardized, and staff easier to rank against machine-augmented benchmarks also make headcount easier to cut once leadership decides the culture can absorb the hit.

Cloudflare is not unusual because it laid people off. Tech has been doing that for years. Cloudflare is useful because it said the new part out loud. The company is growing. The company is using AI heavily. The company says the organization now needs to be architected differently. That is the sentence structure we are going to hear over and over from companies that are doing fine financially and still want fewer people in the loop. AI, in that framing, is not merely a productivity multiplier. It is a permission structure for shrinking the human side of the business while calling it innovation.

Maybe some of that is economically rational. It may even be durable. But nobody should pretend the social meaning is ambiguous. When strong quarters and mass cuts can comfortably coexist under the banner of the future, the future being described is not one where AI simply helps people work better. It is one where the burden of proving your continued necessity gets much heavier, much faster, and much more public.

That is why stories like this land harder than the usual quarterly layoff churn. They function as previews. If a company with good growth and strong branding can reposition more than a thousand eliminations as evidence of seriousness about the future, plenty of executives elsewhere are going to decide they can do the same. The danger is not that Cloudflare is uniquely cruel. The danger is that Cloudflare is intelligible. The memo is neat. The language is polished. The rationale sounds modern. Which means a lot of people in boardrooms probably read it and saw a template.

Sources: Cloudflare blog memo – Cloudflare Q1 2026 results – TechCrunch