NET · Q2 2025 Earnings

Bullish

Cloudflare

Reported July 31, 2025

30-second summary

Cloudflare printed $512.3M in Q2 revenue, up 28% YoY, with current RPO accelerating to +33% and total RPO to +39% — a clear reacceleration narrative backed by the largest customers spending at "the highest levels since 2022." Management raised the FY2025 revenue midpoint to $2.114B and EPS to $0.855, and Matthew Prince used the call to reframe Cloudflare as positioned to power the agentic web (Act 4) rather than just an infrastructure vendor. The bull case is now leaning heavily on Act 4 (agentic web) monetization that Prince himself says is too early to model.

Headline numbers

EPS

Q2 FY2025

$0.21

Revenue

Q2 FY2025

$0.51B

+28.0% YoY

Gross margin

Q2 FY2025

74.9%

Free cash flow

Q2 FY2025

$0.03B

Operating margin

Q2 FY2025

-13.1%

Key financials

Q2 FY2025

Metric	Q2 FY2025	YoY
Revenue	$0.51B	+28.0%
EPS	$0.21	—
Gross margin	74.9%	—
Operating margin	-13.1%	—
Free cash flow	$0.03B	—

Guidance

Prior quarter data unavailable — comparison not possible.

Platform metrics

Q2 FY2025

Segment	Q2 FY2025
Annualized Revenue Run Rate	$2.0+ billion
Current RPO YoY Growth	33%
RPO YoY Growth	39%
Global Website Coverage	20%+ of all websites

Profitability

Q2 FY2025

Segment	Q2 FY2025
Non-GAAP Gross Margin	76.3%
Non-GAAP Operating Margin	14.1%
Free Cash Flow Margin	6.0%
Operating Cash Flow	$99.8 million

Management tone

This is first coverage, so multi-quarter arc analysis is unavailable. The standout shift versus Cloudflare's historical positioning is the elevation of the agentic web (Act 4) from a product theme to the strategic centerpiece of the call.

Prince framed Act 4 as a fundamental opportunity, not a feature. His framing: "Cloudflare is increasingly the platform the most innovative companies are choosing to power the future of AI." The signal is that management believes the next leg of growth comes from sitting between AI companies, publishers, and agents — not just selling more Act 1 and Act 2 products.

The infrastructure narrative also flipped. Where distributed networks were historically discussed as a capacity and capex challenge, Prince now frames the architecture — every server runs every service — as the source of cost advantage during AI inference and DDoS spikes. Crucially, because Cloudflare's caching proxy means outbound traffic exceeds inbound under normal conditions, major DDoS attacks don't drive incremental bandwidth cost. The narrative has moved from "infrastructure is what we build" to "infrastructure is why we win economics."

Management also pulled forward AI monetization confidence by pointing to concrete bookings, most notably a one-year $15M pool-of-funds Workers AI deal with a rapidly growing AI customer that consolidated all inference workloads from a hyperscaler onto Cloudflare. That is a sharper claim than the typical hyperscaler-adjacent AI commentary that hedges on timing.

The overall tone is bullish and notably more assertive than Cloudflare's historically measured engineering-led posture. Hedging language is largely confined to standard guidance qualifiers and Prince's repeated insistence that Act 4 revenue is "way too early" to model.

Recurring themes management leaned on this quarter:

Agentic AI-first operating modelCustomer adoption of AI workloads acceleratingPlatform consolidation driving customer spendOperating leverage and margin expansionCompetitive differentiation through AI capabilitiesInternational growth expansion

Risks management surfaced:

Macroeconomic uncertainty and customer spending patternsCompetitive intensity in AI and security marketsExecution risk on AI product roadmapCustomer concentration riskRegulatory and compliance challenges

Q&A highlights

Keith Weiss · Morgan Stanley

Can you provide more color and visibility on the agentic web business model - what business models are you enabling for customers and how will Cloudflare monetize this?

Matthew Prince acknowledged the business model is too early to determine exactly, with multiple models likely to emerge and consolidate over time. He referenced the iTunes 99-cent song model as an analogy and suggested models could range from fractions of a penny per transaction to variable pricing. Cloudflare's primary focus is adoption and positioning itself as the universal translator regardless of protocol, given its position with much of the internet behind it.

Multiple different business models expected to emerge and consolidateAnalogized to Apple's iTunes model evolution to Spotify's $10/monthCould range from fraction of a penny per transaction to variable pricing by siteToo early to model revenue impact

Andy Nowinski · Wells Fargo

Can you discuss the Act 1 segment (WAF, DNS, DDoS) which appears to be inflecting alongside Act 2, 3, and 4 products? What's the competitive advantage?

Matthew Prince detailed Cloudflare's fundamental architectural advantage: every server in their network can run every service, unlike competitors who use dedicated scrubbing centers. This eliminates cost and latency penalties during attacks and allows scaling across the entire network. The same architecture enables success in Act 2, 3 products, and acts as a cost advantage during major DDoS attacks since bandwidth costs don't spike.

Every Cloudflare server capable of running every service (WAF, DDoS, etc.)Competitors use dedicated scrubbing centers with capacity planning constraintsNo scrubbing center architecture eliminates cost increases during attacksSame architectural advantage applies to Act 2 (Zero Trust), Act 3 (Workers), and Act 4

Matt Hedberg · RBC Capital Markets

How are go-to-market improvements and technology improvements impacting ability to land larger deals? Can you update on partner momentum?

Matthew Prince emphasized the shift from product-led growth to relationship-based sales under leadership of Mark Anderson. Larger deals require explaining total platform capability and ROI, which the upgraded go-to-market team is executing. Partners are growing faster than rest of business through a 'partner-first sales strategy,' though Cloudflare won't reach 90%+ partner sales like Cisco or Zscaler. Partners are behind many large deals.

Shifted from pure product-led growth to relationship-based sales modelUpgraded go-to-market team explaining total platform capability and ROIPool of funds deals indicate customers betting on broad Cloudflare platformPartner sales channel growing faster than rest of business

Gabriella Borges · Goldman Sachs

On paper crawl / content creator payments: What are friction points with AI companies and front-chain models? Is the decision maker for Act 4 the same as Act 1-3? How to build sponsorship across organizations?

Matthew Prince stated publishers are 100% aligned and he was surprised by positive AI company reactions, as they understand original content is essential fuel for their systems. AI companies recognize the need for a level playing field where content creators are compensated. Decision makers are limited (major AI companies) or transaction-based (for individual content access). Cloudflare has relationships with the right people. May not optimize revenue extraction from media companies if Act 4 becomes highly strategic.

Publishers 100% enthusiastic; AI companies surprisingly receptiveAI companies spending billions on GPUs and talent, but content assumed freeDifference between search engines (direct traffic) vs answer engines (capture without referral)Decision makers are AI companies or content transaction-based

Patrick Colville · Scotiabank

Foundation model vendors have called out Cloudflare as third-party sub-processor. How can Cloudflare deepen relationships with FMV and which products are relevant?

Matthew Prince stated ~80% of major AI companies are Cloudflare customers. Primary entry point is security (protecting against abuse, token theft, high-volume automated requests that cost money to process). Cloudflare Workers AI is increasingly being used for inference tasks to deliver better performance and meet regulatory requirements by running inference close to users. Not yet suitable for massive LLM training but investing in supporting larger models over time.

~80% of major AI companies are Cloudflare customersSecurity is primary entry point (abuse protection, token theft prevention)Workers AI used for inference to improve performance and regulatory complianceEach AI model inference request has real cost in fractions to pennies

What to watch into next quarter

Current RPO growth sustaining above 30%. The 33% Current RPO growth materially outpaces 28% revenue growth and is the single best leading indicator. A reversion below 30% would invalidate the reacceleration thesis.

Q3 revenue landing above the $544.5M high end. Cloudflare has a history of beat-and-raise; printing at or below the midpoint after this quarter's bullish tone would read as a deceleration signal.

Non-GAAP operating margin holding 14%+ as investment continues. If Q3/Q4 margins compress despite the rhetoric, the operating leverage story breaks.

Any disclosed revenue contribution from Workers AI or pay-per-crawl. Prince said Act 4 monetization is unmodelable today; first disclosure of a discrete AI-attributable revenue line would be a major positive catalyst. Continued silence would erode the AI-first narrative.

Customers > $100k net adds and "pool of funds" deal commentary. Management cited largest-customer investment levels at 2022 highs; watch whether the $100k+ customer cohort growth (3,712, +22% YoY this quarter) confirms this in Q3 disclosure.

Sources

Cloudflare Q2 FY2025 press release (Form 8-K Exhibit 99.1), filed July 31, 2025: https://www.sec.gov/Archives/edgar/data/1477333/000147733325000136/q225exhibit991.htm
Cloudflare Q2 FY2025 earnings call commentary (Matthew Prince, Thomas Seifert prepared remarks and Q&A).

Get the next brief, free.

We publish analyst-grade earnings briefs the same day or morning after every call — headline numbers, segment KPIs, Q&A highlights, and tone analysis. Free during beta.

This is not investment advice.