Cloudflare Called 1,100 Jobs Obsolete on Its Best-Ever Revenue Quarter

Cloudflare Called 1,100 Jobs Obsolete on Its Best-Ever Revenue Quarter
Cloudflare grew 34% and still cut 20%. Every finance team in mid-cap tech now has a new opening line for the marketing budget meeting.

Cloudflare cut 1,100 jobs on May 7, 2026, the same day it reported a record $639.8 million quarter with revenue up 34% year over year. CEO Matthew Prince framed the cuts as a structural shift to an "agentic AI-first operating model," not a cost-saving exercise. Restructuring charges run $140 to $150 million, with the majority booked in Q2.

The number that breaks the usual layoff defense

For roughly two years the standard CFO conversation about layoffs has been: revenue softened, so headcount has to follow. Cloudflare just removed that script. Q1 came in at $639.8 million versus $479.1 million a year earlier, the highest quarter the company has ever booked. Full-year guidance landed at $2.805 to $2.813 billion, narrowly above consensus, per the company’s own Q1 2026 release. Twenty percent of the workforce is still gone, and the stock dropped 24% on the day, per CNBC. The market did not love the framing either.

That’s the line marketers should sit with. The cuts didn’t follow the revenue. Prince and COO Michelle Zatlyn said in their internal note, quoted by TechCrunch, that "today’s actions are not a cost-cutting exercise or an assessment of individuals’ performance; they are about Cloudflare defining how a world-class, high-growth company operates and creates value in the agentic AI era."

That sentence does different work than "we missed plan." It says entire role categories now sit below an internal AI productivity bar, regardless of how the P&L is doing.

What Prince actually said about role obsolescence

Prince’s blunter quote, reported by BusinessToday, is the one to keep handy for your next planning meeting: "A lot of the support people that provide support behind them, those roles aren’t going to be the roles that, you know, drive companies going forward."

He also said Cloudflare’s internal AI usage went up more than 600% in three months. That number is the math that produced the cut decision. Not "we expect to save X dollars." More like "humans doing this task look 600% more expensive than the agent doing it," which is a different category of argument and a much harder one to negotiate.

The detail that matters: sales staff with revenue quotas were exempt. Everyone behind quota-carrying roles, support, ops, internal admin, was on the table. Marketing largely sits behind quota-carrying roles in most org charts. From what I’ve seen, that fact is the one most marketing leaders quietly understand and don’t flag in budget meetings.

The CMO data was already moving here

A Spencer Stuart survey reported by Yahoo Finance found that 36% of CMOs anticipate cutting headcount in the next 12 to 24 months "by utilizing AI or eliminating redundancies." At companies generating $20 billion or more in annual revenue, the number was 47%.

Salesforce’s Marc Benioff cut roughly 4,000 customer support roles and said publicly, "I need less heads." Microsoft offered voluntary retirement to 8,750 U.S. employees this year. Meta announced 8,000 cuts effective May 20, with HR and recruiting absorbing 35% to 40% of those reductions, per Personnel Today. The trend is not Cloudflare-specific. Cloudflare just gave it the cleanest financial framing.

For marketing leaders specifically, separate Gartner data referenced in industry coverage shows marketing job postings grew roughly 6% year over year from 2024 to 2026, while total marketing output grew about 24%. That gap is the empirical version of what Prince is describing. Output is rising. Headcount is barely moving. Some companies are now resolving the gap by cutting headcount instead of hiring into the production curve, which is a more aggressive read of the same data.

Why marketing is an easier “obsolete” call than support

Marketing functions look uncomfortably similar to the roles Cloudflare just removed. A lot of the daily work in a 10 to 20 person team is asset production, copy review, campaign QA, reporting, and internal coordination. Those map almost one for one to AI-augmentable tasks. The 2 to 100x productivity multipliers Prince has been quoting internally are not hard to replicate with a stack of well-trained agents on copy variants, data pulls, and routine creative builds.

The harder roles to defend in 2026 are mid-level marketers whose output has shifted toward "review the agent’s draft, send to legal." Senior strategists generally remain fine, mostly because their work involves cross-functional judgment that doesn’t live cleanly inside a single agent context. Junior generalists are the soft target. There’s already early movement here: 23% of agencies cut junior copywriting headcount in 2025 and 31% planned further cuts this year, per industry survey data referenced across recent CMO coverage. The economics of replacing a $90K junior copywriter with a $600 agent stack are not subtle, and CFOs already see them on a slide.

This is the part of the picture I think most marketing teams will continue to underrate, mostly because they don’t have a CFO sitting across from them quoting Matthew Prince. Yet. The same conversation Anthropic and OpenAI are forcing on the consulting world (see our piece on how Anthropic and OpenAI both skipped the consulting layer on May 4) is the one mid-cap tech CFOs are going to start having about marketing operations.

The defense to write down before the next budget meeting

This isn’t an argument against AI productivity gains. It’s an argument for getting your team’s defense on paper while the framing is still under your control.

Three things worth doing this week, with rough benchmarks:

  1. Run the marginal-output audit. For every role on your team, write down what they produced last quarter that an AI agent demonstrably could not. Specific campaigns, specific judgment calls, specific qualitative work. Aim for at least three concrete artifacts per role. Anything that comes back blank is one Prince-style framing meeting from being on the table.
  2. Convert review tasks into decisions. If a role is mostly "review the agent’s output," document the decision criteria they’re applying. That is the actual non-replaceable work. Write it as a decision policy with named authority. A line item with a documented decision policy is meaningfully harder to call obsolete than a generic "QA" role on a chart.
  3. Reprice the alignment tax. Marketing teams already lose roughly 20% to 30% of cycle time to internal alignment, brief revisions, and stakeholder handoffs, per industry surveys. That work is exactly what an agentic stack does badly today. Document where your team is shouldering that tax and tie it to revenue moments it protected. Not glamorous, but it is the part of the job that survives the productivity argument.

A closing note that’s less clean than I want it to be

I genuinely don’t know if Cloudflare will actually be more productive in 12 months. It could turn out the support roles being cut were doing more sense-making than the official framing suggests, and the productivity numbers smooth out into the kind of mid-2020s "we found efficiencies" story that gets revisited in 2028 with a quieter rehire cycle.

But the framing isn’t going to wait for the post-mortem. CFOs already have the new sentence. "Cloudflare grew 34% and still cut 20%" is the version every finance team in mid-cap tech will quote at marketing leaders for the rest of 2026. Whether the outcome holds up doesn’t matter much for your next planning conversation.

So get the defense on paper before the meeting, not during.

Notice Me Senpai Editorial