Topic: Agentic Web business model and monetization
Key points:
Multiple models may emerge (e.g., fraction of $0.01 per transaction, site-specific pricing, discounts for agents); analogy to music industry evolution ($0.99/song → $10/month Spotify).
Too early to model revenue; current focus is on adoption and being a "universal translator" for protocols (MCP, Google, Microsoft).
Cloudflare's strategic position: "so much of the Internet sits behind us," enabling fee collection from facilitating agent transactions.
Mgmt stance: Neutral — acknowledges uncertainty in revenue model, prioritizes adoption over near-term monetization.
Q2 — Andrew James Nowinski
Topic: Act 1 product architectural advantage and ARR growth path
Key points:
Every Cloudflare server runs all services (no scrubbing centers); DDoS attacks consume only "a few percentage points" of network capacity vs. ~50% for competitors.
Under normal traffic, outbound > inbound; bandwidth cost is on the greater of in/out, so attacks don't change cost structure.
Same architecture enables Zero Trust (e.g., service in Kigali) and Workers scaling.
ARR surpassed $2 billion in Q2; tracking well to FY 2028 $5 billion target.
Topic: Go-to-market improvements and partner momentum for large deals
Key points:
Product-led growth alone insufficient for deals >$1M; go-to-market team (Michelle, Mark) built relationships for $5M–$20M deals, including first $100M deal last quarter.
Pool of funds deals reflect customer bets on Cloudflare's broad product suite and execution.
Partner sales channel growing faster than rest of business; partners behind many large deals.
Mgmt stance: Bullish — go-to-market upgrades and partner-first strategy driving upmarket success; partner growth will continue to outpace.
Q4 — Gabriela Borges
Topic: Pay-per-crawl adoption, AI company reactions, and media vertical strategy
Key points:
No publisher encountered was not "100% all in" on pay-per-crawl; AI companies (e.g., OpenAI spending >$10B/year on GPUs) understand need to support content ecosystem.
Three legs for AI: GPUs, talent, content; answer engines (vs. search engines) require value back to creators.
Key goal: level playing field for all AI companies (small pay less, big pay more); Cloudflare facilitates transactions.
Media vertical historically underpenetrated; free users' content may become more valuable than enterprise users' for AI training.
Mgmt stance: Bullish — AI companies receptive; Cloudflare's role as neutral facilitator is pivotal; media vertical could shift from cost to revenue source.
Q5 — Patrick Edwin Ronald Colville
Topic: Foundation model vendor relationships and Agentic AI security
Key points:
~80% of major AI companies are Cloudflare customers; use cases: security (preventing abuse of tokens/credits as currency) and inference via Workers AI (GPUs deployed globally).
Workers AI suited for smaller models; investing to support larger models over time.
Agentic AI security: agents flow through Cloudflare (since much of Internet sits behind it); guardrails and rule enforcement products already in market.
Contrast with Palo Alto Networks: Cloudflare prioritizes internal R&D over acquisitions (high hurdle rate); customers see stitched-together platforms as "Frankenstein" with gaps.
Mgmt stance: Bullish — existing products ready for Agentic AI security; internal innovation preferred; revenue timing uncertain but technology is in place.
Q&A Batch (6-7 of 7)
Q6 — Adam Charles Borg
Topic: Cloudflare One competitive positioning and macro outlook
Key points:
Cloudflare is now "neck and neck with Zscaler" in the Gartner/Forrester SASE space; strategy is to launch products early and iterate fast.
Win-back deal mentioned: a customer lost 18 months ago was won back; Cloudflare wins when given a chance due to speed, ease of deployment, and higher ROI.
Macro environment is "disjointed" — some areas strong, others struggling; Cloudflare continues to see value recognition from customers despite uncertainty.
Mgmt stance: Bullish — confident in innovation and cost advantage as recipe for winning, and sees continued performance regardless of macro.
Q7 — James Edward Fish
Topic: Workers platform workload capture and FedRAMP progress
Key points:
Strategy: get a toehold by moving one mission-critical or latency-sensitive function from a hyperscaler/on-prem to Workers; then expand as customers see cost savings and performance.
FedRAMP: on track to meet all requirements before end of 2025; once achieved, Cloudflare will be unrestricted in federal government business.
Mgmt stance: Bullish — Workers platform is gaining traction via targeted go-to-market (speedboat team), and FedRAMP completion will unlock federal market.
Q8 — Roger Foley Boyd (UBS)
Topic: Pool of funds trends and dollar net retention improvement
Key points:
Pool of funds deals with largest customers represented "low double digit" of revenue in Q2 2025, up from "less than 3%" a year ago.
Across all pool of funds deals, tracking "on, if not slightly ahead of target"; variable revenue from consumption of existing customers was a "good contributor" to Q2 outperformance.
Dollar net retention improved by 3 points in the quarter; attributed to ramping consumption against pool of funds and harvesting tailwinds from this go-to-market instrument.
Mgmt stance: Bullish — pool of funds is becoming a net tailwind, with better visibility and execution driving D&R improvement.