Cloudflare fired 1,100 people during its best quarter ever
The "agentic AI era" has its first honest layoff, and Wall Street didn't like the honesty.
Cloudflare cut 20% of its workforce last Wednesday, roughly 1,100 people, the same day it reported $640 million in quarterly revenue. That’s a 34% jump year over year. The best single quarter in the company’s 16 year history.
CEO Matthew Prince and co-founder Michelle Zatlyn published a blog post titled “Building for the Future” alongside the earnings call. The core claim: AI agent usage inside Cloudflare grew 600% in 3 months. Employees across engineering, HR, finance, and marketing now run thousands of AI agent sessions daily. The roles being cut, Prince wrote, were roles the company no longer needs in an “agentic AI-first operating model.”
Two things make this different from the usual layoff-with-a-press-release. First, Prince said the quiet part loud. This wasn’t framed as “restructuring” or “performance management” or the classic “difficult macroeconomic environment.” He pointed directly at AI and said these jobs are gone because the work is being done by agents now. Second, Cloudflare is profitable and growing fast. This is the first major tech company to fire a fifth of its staff during a record quarter and explicitly credit AI for making it possible.
Wall Street panicked. The stock dropped 24% the next day, even though earnings beat expectations (25 cents EPS vs. 23 cents expected). The sell-off wasn’t about the numbers. The numbers were fine. Investors got spooked by the signal: if a healthy, growing company is cutting this aggressively based on AI adoption, what does the rest of the sector look like in 12 months?
Here’s what I think. Prince is probably right that AI agents are handling work that humans used to do at Cloudflare. I’ve seen this pattern inside large enterprises. Support tickets get routed and resolved by agents. Internal docs get drafted, reviewed, and summarized without a person touching them. HR intake, finance reconciliation, marketing copy, all of it gets faster when you plug agents into the workflow. That part is real.
But 600% growth in AI usage over 3 months is a vanity metric without context. 600% of what baseline? How are they measuring a “session”? Is this actual productivity gain, or is it employees using ChatGPT to rewrite their emails? A Harvard Business Review piece from January 2026 nailed this: companies are laying off based on what AI might do, before they’ve proven it actually works. The cost of being wrong here is high. You can’t rehire institutional knowledge in 6 months when the agents hallucinate your compliance workflows.
The generous severance (full base pay through end of 2026, equity vesting through August) tells me Cloudflare knows this is a bet. You don’t pay that much to leave if you’re confident the math already works. I’d bet Cloudflare quietly rehires for at least some of these functions within 18 months, probably under different titles with “AI” in them.
The broader picture is hard to ignore. Over 113,000 tech workers have been laid off in 2026 so far. Amazon, Microsoft, Alphabet, and Meta plan to spend $725 billion on AI infrastructure this year. The capital is flowing from headcount to compute. Cloudflare just said it plainly.
---
Watch Cloudflare’s Q2 earnings in August. If revenue growth holds above 30% with 20% fewer people, every mid-size SaaS company will have the same conversation at their next board meeting.
---