Q1 FY2026 Earnings Recap

Cloudflare, Inc. (NET)

Reported May 7th 2026

News Summary

  • Cloudflare reported a 600% increase in internal AI tool usage over three months and stated that 100% of production code is now reviewed by autonomous agents; one large AI studio scaled from zero to over 1 million Dynamic Workers on its platform in just 15 days, signaling that agentic workloads are hitting real adoption inflection points that will drive edge inference demand.
  • GPU utilization on Cloudflare's fleet is approaching 70–80%, levels comparable to CPU utilization, which management contrasted against hyperscaler GPU utilization often in the single digits—a direct competitive framing relevant to inference infrastructure economics and a data point supporting the case for purpose-built AI inference hardware at the edge.
  • Cloudflare added 1 million developers to its Workers platform in a single quarter (vs. 1.5 million in all of 2025, now 5.5 million total) and reported hundreds of billions of agentic requests per month already transiting its network; management projects nonhuman agentic traffic will surpass human traffic around 2027, a macro sizing signal for AI inference and edge compute demand.
  • Q1 revenue was $639.8M (+34% YoY, 3% beat); non-GAAP EPS $0.25 beat by $0.02; billings grew 37.8% YoY to $709.4M; free cash flow was $84.1M (13% of revenue); gross margin compressed 130 bps YoY to 72.8% as paid AI traffic shifted costs into COGS; FY2026 revenue guidance raised to $2.809B midpoint (+30% YoY) with non-GAAP EPS guidance of $1.19–$1.20, above prior consensus of $1.13.
  • NET fell roughly 20% post-earnings despite the beat, driven by Q2 revenue guidance of $664.5M midpoint that missed consensus by ~$1.5M, the first quarter without revenue growth acceleration in four quarters, and a 20% workforce reduction (1,100 roles) tied to the agentic AI transition; Morgan Stanley framed the layoffs as signaling agentic AI productivity leadership, while Wolfe Research noted results were ahead of their upside model and that the RIF should expand margins going forward.

Financial Highlights

MetricQ2 '24Q3 '24Q4 '24Q1 '25Q2 '25Q3 '25Q4 '25Q1 '26
Revenue$0.4B$0.4B$0.5B$0.5B$0.5B$0.6B$0.6B$0.6B
Rev Growth (YoY)30.0%28.2%26.9%26.5%27.8%30.7%33.6% 33.5%
Gross Margin77.8%77.7%76.3%75.9%74.9%74.0%74.2% 71.2%
Operating Margin-8.6%-7.1%-7.4%-11.1%-13.1%-6.7%-7.4% -9.7%
EV/Sales74.5x67.3x83.4x84.0x136.6x135.1x117.1x117.8x

Earnings Call Highlights

Cloudflare delivered a standout Q1 with 34% revenue growth and simultaneously announced a ~20% workforce reduction framed explicitly as an 'agentic AI-first' operating model transformation, not a cost-cutting measure—a posture that was bold and largely well-received by analysts who probed but did not push back hard. Management was unusually specific and confident about AI-driven internal productivity gains, citing 600% internal AI usage growth in three months and 100% of production code now reviewed by autonomous agents, signaling that the company views itself as a live proof-of-concept for the agentic enterprise. The dominant theme was that Cloudflare's network and Workers platform are structurally positioned to capture the infrastructure layer of the agentic internet, with hundreds of billions of agentic requests per month already flowing through the network. Analyst questions were constructive and forward-looking, focused on GPU/CPU fleet economics, gross margin trajectory, and Act 4 monetization, suggesting the Street is buying the AI infrastructure narrative. The tone from Prince and Seifert was one of deliberate, confident offense—acting from strength rather than defensiveness.

  • Cloudflare disclosed that one large AI studio went from essentially zero to over 1 million Dynamic Workers running on its platform in just 15 days, illustrating explosive early adoption of its agentic execution environment as a direct alternative to containers for AI workloads.
  • Prince stated GPU utilization on Cloudflare's fleet is approaching CPU utilization levels of 70–80%, contrasting sharply with hyperscaler GPU utilization often in the single digits—a meaningful unit-economic advantage for inference workloads that Cerebras should note as a competitive framing in the market.
  • The company is routing internal AI coding and agentic tasks across its own GPU fleet via AI Gateway, achieving high utilization and near-zero marginal cost for simpler tasks, and is already demoing this 'Cloudflare OS' harness to CIOs who consistently respond that they want it too—signaling a potential new commercial product.
  • Cloudflare reported 5.5 million developers on its Workers platform, adding 1 million in a single quarter versus 1.5 million in all of 2025, with developer platform products growing significantly despite carrying below-average gross margins—underscoring accelerating AI developer infrastructure adoption.
  • Management explicitly called out that nonhuman agentic traffic is projected to surpass human traffic around 2027, and that agents visiting thousands of sites versus a human's five will be the primary driver of Act 1 network revenue growth—a macro datapoint directly relevant to AI inference and edge compute demand sizing.