Skip to content

Cloudflare Beat Q1 Estimates Thursday. Hours Later, It Cut 1,100 Jobs for the "Agentic AI Era."

DS
LDS Team
Let's Data Science
9 min
Cloudflare reported $639.8 million in revenue, beat Wall Street on every line, and raised full-year guidance. Then CEO Matthew Prince posted a memo announcing a 20% workforce reduction. The stock fell 24.31% on Friday.

At 4:05 PM Eastern on Thursday, May 7, Cloudflare's investor relations team pushed out a press release that read like a victory lap. The numbers all printed green:

Q1 2026 metricResult
Revenue$639.8 million, up 34% year over year
EPS$0.25, two cents above the $0.23 consensus
Full-year revenue guidanceRaised to $2.805 billion to $2.813 billion
Non-GAAP gross margin73%
Large customer countRecord high

In any normal quarter, that print sends the stock higher.

Minutes later, Cloudflare published a second post on its corporate blog. The title was "Building for the future." The authors were CEO Matthew Prince and President Michelle Zatlyn. The body opened with a sentence that erased the press release: "We are writing to let you know directly that we've made the decision to reduce Cloudflare's workforce by more than 1,100 employees globally."

Friday's close put the stock at $194.36, down 24.31% in a single session. That move erased roughly 22 billion dollars in market value. The same set of numbers Wall Street had cheered at 4:05 PM was being punished by 9:30 AM the next day.

The bridge between those two reactions was a single phrase Prince repeated five times in the memo: the agentic AI era.

The Beat That Did Not Matter

Cloudflare's Q1 was, by every conventional measure, excellent. Revenue grew 34% year over year. Non-GAAP gross margin held at 73%. The company added a record number of large customers, which it defines as those paying more than $100,000 annually. It raised both revenue and EPS guidance for the full year.

What it did not do was issue Q2 guidance that calmed analysts. Cloudflare's projected second-quarter revenue range topped out at $665 million.

The FactSet consensus estimate was $666.1 million. The miss was small. The reaction was not.

The reason traders sold was not the Q2 number. It was the message embedded in the layoff memo: a software company that just posted 34% growth and beat every Wall Street estimate has decided that 20% of its people are not needed for the next era. If that is true at Cloudflare, traders reasoned, it is true at every infrastructure software vendor on the NYSE.

What the Memo Actually Said

Prince and Zatlyn did not soften the announcement with the standard layoff vocabulary. The memo is unusual in tech for how directly it ties the cuts to AI consumption inside the company.

"The way we work at Cloudflare has fundamentally changed. We don't just build and sell AI tools and platforms. We are our own most demanding customer. Cloudflare's usage of AI has increased by more than 600% in the last three months alone. Employees across the company from engineering to HR to finance to marketing run thousands of AI agent sessions each day to get their work done."

Matthew Prince and Michelle Zatlyn, "Building for the future" (Cloudflare Blog, May 7, 2026)

The 600% number is the load-bearing claim. Prince framed the layoffs as a logical consequence of the productivity that AI agents have already delivered inside Cloudflare. Roles that supported customer-facing engineering or code creation, he wrote, "are not going to be the roles that drive companies going forward."

That language is what spooked the market. Cloudflare is not a company that uses AI sparingly. It sells a public agentic AI platform, runs the largest deployment of Workers AI on the global edge, and ships a remote MCP server for autonomous agents. If a company that builds AI infrastructure has concluded its own back office can shrink by a fifth, every other software CFO will be asked the same question on the next earnings call.

How Friday Unfolded

The reaction to the memo played out on a compressed schedule.

THU MAY 7, 4:05 PM ET
Q1 2026 earnings beat published
Revenue $639.8M, up 34% YoY. EPS $0.25 vs $0.23 expected. Full-year guidance raised. Stock initially trades up 3.3% in after-hours.
THU MAY 7, AFTER MARKET CLOSE
"Building for the future" memo posted
Prince and Zatlyn announce reduction of more than 1,100 employees, roughly 20% of the workforce. Memo cites 600% rise in internal AI usage over the prior three months.
THU MAY 7, 5:00 PM ET
Earnings call. Q2 guide lands soft
Q2 revenue range of $664M to $665M comes in just below the $666.1M FactSet consensus. After-hours stock reverses, falling sharply.
THU MAY 7, EVENING
SEC 8-K filed
Cloudflare discloses restructuring charges of $140 million to $150 million, with $105 million to $110 million in cash costs and $35 million to $40 million in accelerated equity vesting.
FRI MAY 8, 4:00 PM ET
Stock closes at $194.36, down 24.31%
Worst single-day drop in Cloudflare's history as a public company. Roughly $22 billion in market capitalization erased.

The 600% Number Is Doing a Lot of Work

The memo's central claim is that internal AI usage at Cloudflare grew more than 600% in the three months preceding the announcement. Prince did not define what unit was growing 600%, and Cloudflare has not published a methodology. The memo refers to "thousands of AI agent sessions each day" across engineering, HR, finance, and marketing.

That number is what justifies the size of the cut. It also justifies its specific shape. Affected roles, according to multiple internal communications reported by The Register and Personnel Today, are concentrated in support functions: tier-one customer support, sales operations, internal IT, recruiting coordination, and back-office finance.

The pattern is the same one that emerged in Oracle's 30,000-person reduction in March and is now playing out across the sector. Coding-adjacent roles where AI agents can produce a credible first draft go first. Pure customer-facing engineering and revenue-generating sales are protected. Whether those agents actually deliver the productivity that justifies a permanent 20% headcount reduction is the bet that public-market investors are now being asked to underwrite.

The Severance Is the Quiet Story

For a layoff of this scale, the package is unusually generous. Cloudflare disclosed in its blog post and 8-K filing that affected employees in the United States will receive:

  • Full base pay through December 31, 2026, regardless of when they are formally separated
  • Continued healthcare coverage through year-end
  • Equity vesting extended to August 15, 2026, with the standard one-year cliff waived and pro-rated vesting in its place
  • Outplacement services and immigration support for visa-dependent employees

The total bill breaks down like this:

Severance componentAmount
Cash severance, benefits, and notice pay$105 million to $110 million
Accelerated equity vesting$35 million to $40 million
Implied per-employee combined valueAbout $130,000

That implied per-person value is an order of magnitude more than the six weeks Oracle paid laid-off staff in March.

Two readings are possible. The charitable one: Prince and Zatlyn, both Cloudflare co-founders since 2009, wanted the package to reflect actual obligation to the people being cut. The cynical one: the more time and money Cloudflare spends on severance, the longer affected employees take to file for unemployment, give interviews, or join the early class action that always follows a cut of this size.

The Other Side

Not everyone reading the memo accepts the AI-productivity story.

Wall Street's reaction was the most direct rebuttal. The 24% drop on Friday, against a Q1 print that beat consensus on revenue, EPS, and full-year guidance, said in dollars what no analyst report could: investors do not believe the math behind a 20% headcount cut at a company that just grew 34%. If the back office was already redundant given AI usage, why was Cloudflare hiring through Q1? If it was not redundant, what changed in sixty days?

Skeptical labor economists have spent the past year arguing that productivity gains attributed to large language models are smaller than executive memos imply. The Register quoted multiple labor analysts on Friday warning that "AI restructuring" has become the safe corporate narrative for cuts that companies wanted to make for unrelated reasons, and that without published productivity data the claim cannot be tested.

Some Cloudflare engineers writing on Hacker News and other developer forums raised a different concern. The same memo that promised AI productivity also acknowledged that Cloudflare's customer-support escalation queue was already backlogged. Cutting tier-one support staff before the agents have proven they can fully replace those workers, several engineers wrote, is the corporate equivalent of removing the safety net before testing the trapeze.

HR practitioners are reading the severance package itself as a signal. The unusually generous terms, several pointed out, are the kind a company writes when its lawyers expect litigation around disparate impact, age discrimination, or visa-holder treatment. Cloudflare ended 2025 with 5,156 employees. 1,100 is not a rounding error.

What This Means for Practitioners

For data scientists and ML engineers, the Cloudflare announcement has three implications worth tracking.

The first is product. Cloudflare's restructuring leaves Workers AI, Vectorize, the Agents SDK, and the public agentic AI platform untouched. Roadmaps for those products were reaffirmed on the earnings call. If anything, the cuts free up internal compute and engineering bandwidth to ship them faster.

The second is procurement. Cloudflare is the second large infrastructure vendor in six weeks (after Oracle) to publicly tie a major workforce cut to internal agent usage. Vendors talking to enterprise procurement teams should expect "what is your AI-driven productivity story" to become a standard line item in 2026 RFPs. Saying "we have not measured it" is no longer a neutral answer.

The third is the labor market. Roughly 1,100 Cloudflare alumni will be on the market between now and January 2027, most with at least one year of paid runway from the severance package. That is a significant inflow of senior infrastructure talent at exactly the moment the industry is debating whether infrastructure jobs still exist.

For more on the broader pattern, see Oracle's 30,000-job cut and the AI infrastructure bet behind it. For another company tying mass layoffs to a record AI capital plan, see Meta's plan to fire 8,000 people on May 20. For context on the productivity claims behind these cuts, see the controlled study finding developers were slower, not faster, with AI coding assistants.

The Bottom Line

Cloudflare beat earnings, raised guidance, and posted record large-customer growth. Then it cut 1,100 people, lost a quarter of its market cap in a single session, and told Wall Street the reason was that internal AI usage had risen 600% in three months.

The market did not punish Cloudflare for the financials. It punished Cloudflare for the message. If a company that just delivered 34% revenue growth concludes that a fifth of its workforce is no longer needed because of agents it deployed inside the past 90 days, every other software vendor will be expected to deliver the same story by the next earnings cycle. That is the trade Prince made when he posted the memo at 4:48 PM on Thursday.

Whether the 600% number holds up is a question for the next two quarters. Whether the 1,100 people get rehired into AI-supervised roles elsewhere is a question for the next two years. The question Cloudflare answered on Thursday is the only one that matters for the rest of 2026: when the choice is between protecting the workforce and protecting the multiple, the multiple wins.

As Prince closed the memo: "Today is a hard day. We hope tomorrow will be a better one." The market answered the next morning, and it cost 22 billion dollars.

Sources

Practice interview problems based on real data

1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.

Try 250 free problems
Free Career Roadmaps8 PATHS

Step-by-step roadmaps from zero to job-ready — curated courses, salary data, and the exact learning order that gets you hired.

Explore all career paths