Stop Pretending This Is Surprising
Everyone wants to act shocked. Another tech company, another round of layoffs dressed up in AI language, another LinkedIn post about “difficult decisions.” But here’s what the outrage crowd keeps missing: Cloudflare didn’t fail its way to 1,100 layoffs. It succeeded. Record revenue of $639.8 million, a 34% year-over-year jump, and a workforce that is now, by the company’s own account, doing more with less. If you’re going to be angry about this, at least be angry about the right thing.
What Actually Happened
Cloudflare cut roughly 1,100 employees — about 20% of its global workforce — citing AI-driven productivity gains across engineering, HR, finance, and marketing. This wasn’t a company bleeding out and blaming AI to save face. This was a company posting its highest-ever quarterly revenue and simultaneously deciding it no longer needed a fifth of its headcount to sustain that output.
That distinction matters enormously. We’ve seen plenty of tech layoffs where AI was used as a convenient excuse to cover up mismanagement, over-hiring during the zero-interest-rate party years, or just plain bad strategy. Cloudflare’s situation reads differently. The revenue growth is real. The productivity claims, while unverified in granular detail, are at least internally consistent with what AI tooling has demonstrably done inside engineering and operations teams at companies this size.
The Part Nobody Wants to Say Out Loud
AI didn’t sneak up on these 1,100 people. The tools that replaced or displaced their roles — in HR workflows, financial reporting, marketing copy, code review — have been publicly available and widely discussed for the better part of three years. The uncomfortable truth is that many companies, Cloudflare included, spent that time quietly figuring out which roles could be absorbed by AI-assisted processes, and which couldn’t.
What Cloudflare did is simply make the internal math public. Most companies are running the same calculation right now and just haven’t announced the results yet.
Why the “Record Revenue” Framing Is Actually the Story
The juxtaposition of record revenue and mass layoffs is being framed as hypocrisy. I’d argue it’s the most honest version of what AI adoption looks like in practice. Companies don’t use AI to grow headcount. They use it to grow output per employee. When that ratio shifts far enough, headcount drops. The revenue line going up while the payroll line goes down is not a contradiction — it’s the entire business case for AI adoption, stated plainly.
If Cloudflare had posted flat revenue and cut 1,100 jobs, that would be a crisis. What they posted is a case study. An uncomfortable one, but a case study nonetheless.
What This Means for Anyone Reviewing AI Tools
At agnthq.com, we spend a lot of time stress-testing AI agents and tools against real workflows. And the Cloudflare story is a useful gut-check for how to think about what these tools actually do at scale.
- AI doesn’t eliminate departments — it compresses them. Engineering, HR, finance, and marketing all still exist at Cloudflare. They’re just running leaner. The tools didn’t replace functions; they reduced the human hours required to execute them.
- Productivity gains are real, but they’re not evenly distributed. The employees who stayed are presumably the ones who either work alongside AI effectively or do work that AI can’t yet replicate. That gap is narrowing, not widening.
- The “AI took my job” narrative is too simple. In most cases, AI changed the scope of a job until the job no longer needed a dedicated person. That’s a meaningful difference when you’re thinking about which skills to build or which roles to hire for.
The Honest Takeaway
Cloudflare is not a villain in this story, and it’s not a hero either. It’s a preview. A well-run company with solid revenue growth made a calculated decision that AI tooling had matured enough to absorb a significant portion of its operational workload. The 1,100 people who lost their jobs are real, and the disruption to their lives is real. That deserves acknowledgment without pretending the underlying business logic is somehow illegitimate.
The companies worth watching now are the ones that haven’t made this announcement yet. Because the math doesn’t change just because you haven’t published the results.
🕒 Published: