Cloudflare, Inc. · NYSE:NET
Cloudflare — Investor Day (June 2026)
193 slides from Cloudflare's investor day presentation, Jun 9, 2026
Slide 1: Investor Day
Investor Day June 9, 2026 | New York Stock Exchange
Slide 2: Safe Harbor and Other Information
Safe Harbor and Other Information This presentation has been prepared by Cloudflare, Inc. (“Cloudflareˮ) for informational purposes only and not for any other purpose. Nothing contained in this presentation is, or should be construed as, a recommendation, promise, or representation by the presenter
Slide 3: Phil Winslow
Phil Winslow Vice President of Strategic Finance, Treasury, Deal Desk, & Investor Relations
Slide 4: Agenda and speakers
Phil Winslow 100 pm - 105 pm Opening Remarks Vice President of Strategic Finance, Treasury, Deal Desk, & Investor Relations Matthew Prince 105 pm - 130 pm Vision Co-founder & Chief Executive Officer Rita Kozlov 130 pm - 150 pm AI Vice President of Product Sam Rhea 150 pm - 205 pm AICloudfla
Slide 5: Matthew Prince
Matthew Prince Co-founder & Chief Executive Officer
Slide 6: iia ae a
iia ae a =—scINN P a ns) CLOUDFLARE -—
Slide 7: Rita Kozlov
Rita Kozlov Vice President of Product for Cloudflare Developer Platform
Slide 8: Cloudflare's vision for developers
Cloudflareʼs vision for developers To help build a better Internet with a modern cloud platform that accelerates developer velocity and drives better end user experiences. Remove friction Provide the best Deliver enterprise Enable and improve modern primitives grade solutions new AI developer veloci
Slide 9: Bold, non-consensus decision: isolates
A bold, non-consensus decision we made when building the foundation of our developer platform: isolates 0ms Cold starts (no container runtime) for auto-scaling 10x Higher price-performance vs. other cloud platforms Generation One Generation Two Generation Three Generation Four One box; No metal box
Slide 10: The greenfield developer platform opportunity is bigger than ever
The “greenfieldˮ developer platform opportunity is bigger than ever Cheaper Code Explosive Volume Generative AI has driven the cost of writing code More applications are being built today than at any to near zero. What once took weeks now takes point in history. The greenfield market is hours. expan
Slide 11: Last year…
® crouvrtare Last year...
Slide 12: Last year… next phase of AI: automation
Last year… We believed the next phase of AI was going to focus on automation, or enabling autonomous agents to own end-to-end tasks. 12
Slide 13: OpenClaw fastest-growing repository in GitHub history
OpenClaw is the fastest-growing repository in Github history Cumulative Github Stars What React achieved in 13 years, OpenClaw surpassed in 3 months. 13 Source: Cumulative GitHub stars. OpenClaw from the GitHub GraphQL API (full stargazer history); React and Linux from GH Archive star events, inde
Slide 14: Last year… more code in five years than all of history
Last year… We said more code was going to be written in the next five years than in all of history. 14
Slide 15: Explosion in coding productivity
This year has seen an explosion in coding productivity Growth in various indicators of coding output: New Websites - Y/Y Growth New iOS Apps - Y/Y Growth Github PRs - Y/Y Growth Wrangler Total Downloads 34% 68% 153% 942% Y/Y Y/Y Y/Y Y/Y 2022 2024 2026 2022 2024 2026 2022 2024 2026 2022 2024 2026
Slide 16: Last year… best platform for building and scaling agents
Last year… We believed Cloudflare would be increasingly recognized as the best platform for building and scaling agents. 16
Slide 17: Developers are getting it
Developers are “getting itˮ March 2026 5.5+ million 5.5 December 2025 4.5+ million December 2024 3+ million April 2024 2+ million June 2024 2.4+ million MILLION November 2022 1+ million developers on the November 2021 350,000 Cloudflare platform May 2022 450,000 2019 2020 2021 2022 2023 202
Slide 18: Vite ecosystem weekly downloads are skyrocketing
Vite ecosystem weekly downloads are skyrocketing npm Package Weekly Downloads — vite and @cloudflare/vite-plugin Vite is the #1 JavaScript front-end build tool, powering virtually every major framework with the sole exception of Next.js. When AI coding agents generate modern web applications, they i
Slide 19: What's next?
Whatʼs next? 19
Slide 20: Still early stages of the rise of agents
We are still in early stages of the rise of agents Agents are a function of two Train Infer Automate Agents primary components: intelligence / Intelligence reasoning Execution / reasoning and execution. 20
Slide 21: How do agentic AI systems work?
How do agentic AI systems work? AI Workflows + APIs = Agents User Feedback Manager Agent No Analyst Agent Checker Planner Agent External Yes. Agent Systems User Satisfaction User Interpretation Workflow Workflow Iterative Output/ Instructions & Reasoning Generation Execution Improvement Outcome 21
Slide 22: AI + Workflows + APIs = Agents: intelligence / reasoning
AI + Workflows + APIs = Agents — intelligence / reasoning User Feedback Manager Agent No Analyst Agent Checker Planner Agent External Yes. Agent Systems User Satisfaction User Interpretation Workflow Workflow Iterative Output/ Instructions & Reasoning Generation Execution Improvement Outcome 22
Slide 23: Models have significantly improved
AI INTELLIGENCE / REASONING In the past 6 months, models have significantly improved Trend in Best Software Engineering Benchmark Verified Result by Model 23 Source: SWE-bench verified leaderboard archive (github.com/SWE-bench/experiments), as of June 2, 2026. Each line = one lab's rolling-best su
Slide 24: Flexibility in model choice and availability
AI INTELLIGENCE / REASONING Cloudflare offers flexibility in model choice and availability AI Gateway Workers AI What you pay for on a hyperscaler Savings What you pay for on Cloudflare Choose from 150+ models, Best provider for both select private and without vendor lock-in open-source models (e.
Slide 25: AI + Workflows + APIs = Agents: execution
AI + Workflows + APIs = Agents — execution User Feedback Manager Agent No Analyst Agent Checker Planner Agent External Yes. Agent Systems User Satisfaction User Interpretation Workflow Workflow Iterative Output/ Instructions & Reasoning Generation Execution Improvement Outcome 25
Slide 26: Why agents must be able to write code
Why agents MUST be able to write code: 1. Context window limitations 2. Lack of accuracy 3. LLMs are bad at tool calling 26 26
Slide 27: Context window limitations
WORKFLOWS APIs EXECUTION — WHY AGENTS MUST BE ABLE TO WRITE CODE CONTEXT WINDOW LIMITATIONS Why agents MUST be able to write code: Context window limitations Actual Context Tool Definitions System Prompt Task Window 1M 2.5M 2K Conversation 8K When AI agents encounter complex tasks ○ That means
Slide 28: Lack of accuracy
WORKFLOWS APIs EXECUTION — WHY AGENTS MUST BE ABLE TO WRITE CODE LACK OF ACCURACY Why agents MUST be able to write code: Lack of accuracy LLMs lack situational context like date and time. To obtain this information, you need to make additional tool calls, which costs more tokens, or create work
Slide 29: LLMs are bad at tool calling
WORKFLOWS APIs EXECUTION — WHY AGENTS MUST BE ABLE TO WRITE CODE LLMs ARE BAD AT TOOL CALLING Why agents MUST be able to write code: LLMs are bad at tool calling https://www.anthropic.com/engineering/code-execution-with-mcp ✖ “Tool callingˮ is an artificial injection that happens post-training.
Slide 30: LLMs were never trained for tool calling
WORKFLOWS APIs EXECUTION — WHY AGENTS MUST BE ABLE TO WRITE CODE LLMs ARE BAD AT TOOL CALLING LLMs were never trained for tool calling… Metaphor: Making an LLM perform tasks with tool calling is like putting Shakespeare through a month-long class in Mandarin and then asking him to write a play
Slide 31: LLMs were trained on writing lots of code
WORKFLOWS APIs EXECUTION — WHY AGENTS MUST BE ABLE TO WRITE CODE LLMs ARE BAD AT TOOL CALLING …but LLMs were trained on writing lots and lots of code. async () => { const today = new Date(); const eventDate = new Date(today.getFullYear(), today.getMonth(), today.getDate(), 18, 0, 0); const isoD
Slide 32: Need a place to securely execute code at scale
Now, we need a place to securely execute all of this code at scale. So, why canʼt agents just scale on the hyperscalers? 32
Slide 33: How hyperscalers have scaled
How hyperscalers have scaled… Microservices were all about vertical scaling, or creating many copies to run the same thing. From To Monolith Microservices* * same application, scaled to millions of users 33
Slide 34: …won't work for agents
…wonʼt work for agents Microservices rely on being able to run many instances of one application, whereas agents generate new instances of applications on each run. From To To Monolith Microservices* Agents** * same application, ** infinite ephemeral applications, scaled to millions of users many pe
Slide 35: Microservices and hyperscalers: one menu
METAPHOR INDUSTRIAL KITCHEN vs. PERSONAL CHEF Microservices and hyperscalers: One menu, serving millions per day 35
Slide 36: How hyperscalers scale microservices
METAPHOR INDUSTRIAL KITCHEN vs. PERSONAL CHEF How hyperscalers scale microservices: spin up new franchise kitchens, operationalized around the same menu 36
Slide 37: The problem for agents with hyperscaler scaling
METAPHOR INDUSTRIAL KITCHEN vs. PERSONAL CHEF The problem for agents with how hyperscalers scale: all compute until today has been designed to execute the same code 37
Slide 38: What agents need to scale
METAPHOR INDUSTRIAL KITCHEN vs. PERSONAL CHEF What agents need to scale: ephemeral pop-ups, specific to their needs, that can be instantly spun up and down for a task 38
Slide 39: Why the current cloud approach won't scale for agents
Why the current cloud approach wonʼt scale for agents United States 100M Knowledge Workers 1 Agent per worker 10 Agents per CPU 10M CPUs Global 1B Knowledge Workers 10 Agents per worker 10 Agents per CPU 1B CPUs Current global server CPU production: 20x current global server 35 45M per year CP
Slide 40: The next bottleneck to AI adoption: CPU for execution
Whatʼs the next bottleneck to AI adoption? Itʼs the CPU for execution. — So, how do we scale to billions of agents? Unpredictable CPU workloads Automate driven by Train Infer Agents agents is the next Consistent GPU workloads Unpredictable GPU workloads bottleneck to Intelligence / Reasoning Execu
Slide 41: The cloud that got us here won't get us where agents go
The cloud that got us here wonʼt get us to where agents need to go 41
Slide 42: Isolates were built for this moment
Remember that bold, non-consensus decision we made back in 2017? Isolates were built for this moment! 42
Slide 43: Isolates can run up to 100x more efficiently
Because isolates can reshare the runtime against many runs, they can run up to 100x more efficiently 43
Slide 44: Cloudflare lowers TCO by 50% for agents
What does this mean for customers? Cloudflare lowers TCO by 50% for agents and the code they write Serve a single-region web app Deploy 10K agent-generated apps Run 1M agent sessions 1B requests per month Every day, for 30 days Per month 15.2% lower TCO on AWS 63.6% lower TCO on Cloudflare 74.7% lo
Slide 45: Andy Jassy was right about primitives
Andy Jassy was right about primitives… “The best way we know how to do this is by building primitive services. … Primitives are the raw parts or the most foundational-level building blocks for software developers. … And, over the last 20 years, primitives have been at the heart of how weʼve innovate
Slide 46: The agentic cloud needs new primitives
…but the agentic cloud needs new primitives 2006 2026 S3 🡒 Durable Objects & Artifacts EC2 → Isolates Workers) SQS → Workflows 46
Slide 47: Customer quote: Discord
We know what the hyperscalers will look like in 10 years…the same as they do now. Weʼre looking to Cloudflare for what the next generation cloud will look like.ˮ – Mark Smith, VP of Core Tech at Discord 47
Slide 48: Internet traffic is changing as a result of agents
Internet traffic is changing as a result of agents, driving demand for agentic cloud primitives to scale Desktop and WHERE One Room → On-premises → Everywhere → Anywhere HOW WHEN Batch Deliberate Ubiquitous Continuous 48 Note: Framework inspired by Stratechery.
Slide 49: Agent-native businesses are powered by Cloudflare
Mobile-native businesses Agent-native businesses were powered by the cloud are powered by Cloudflare ⟶ ⟶ ⟶ ⟶ ⟶ ⟶ ⟶ 49
Slide 50: In the past month alone…
Figma @figma ic) vwdym of course there's an agent right on the design canvas that’s fluent In Figma and native to the way your team works Lovable @Lovable o Your Lovable apps are now built to be found. New apps are now server side rendered and discoverable through search ‘engines like Google and Al
Slide 51: Fastest, most secure, most cost-effective place to build agents
Cloudflare is the fastest, most secure, and most cost-effective place to build, deploy, and scale… The code Agents & they write 51
Slide 52: One cloud to power them all
8 billion Multiple agents Infinite people → per person → lines of code One cloud to power them all 52
Slide 53: Sam Rhea
Sam Rhea Chief Information Officer
Slide 54: We don't just build and sell AI tools
We donʼt just build and sell AI tools and platforms… 54
Slide 55: We are our own most demanding AI customer
…we are our own most demanding AI customer 55 55
Slide 56: What is Cloudflare OS?
What is Cloudflare OS? 56
Slide 57: Equip and enable every team member to do more
Equip and enable every team member to do more One-Click AI Agent Centrally Managed Secure System of Model Optionality by Workspace Context Layer Record Access Use Case and Role Full agent harness running Curate the jobs-to-be- Provide access to critical Default to the right model in a Cloudflare con
Slide 58: Cloudflare's agentic workspace
Cloudflareʼs agentic workspace 58
Slide 59: Framework for where and how team members build
A new framework for where and how team members build 59
Slide 60: Scaling productivity without scaling cost
Scaling productivity without scaling cost Cost-efficient Al usage for every employee: ✓ Model optionality Control model usage without lock-in risk ✓ Dynamic routing Map employee roles and skill files to specific models ✓ Infire Faster, more cost-efficient inference ✓ Caching Save token spend on at-s
Slide 61: Investing in a central context layer
Investing in a central context layer Defining the jobs-to-be-done as centrally managed skill files: ● Tap experts to document workflows and 1 create skill files. ● Create a “magic inboxˮ email alias where 2 everyone else can send work. ● Staff the inbox with Al agent operators 3 who run it as a serv
Slide 62: Cloudflare OS in Action
Cloudflare OS in Action AL A CLOUDFLARE
Slide 63: A comprehensive skill library
A comprehensive skill library Every team member 63
Slide 64: Builders / Sellers
® crouvrtare Builders Sellers 64
Slide 65: Code review lets us ship even faster, safely
Code review lets that us lets ship us ship eveneven faster, faster, safely safely Builders 65 65
Slide 66: Agents fix problems so engineers can focus on building
Agents fix problems so engineers can focus on building Builders 66
Slide 67: BDRs research prospects at scale
BDRs research prospects at scale for outbound work Sellers 67
Slide 68: Every seller starts with a prescriptive plan
Every seller starts their day with a prescriptive plan Sellers 68
Slide 69: Generate custom slide decks for each conversation
Generate custom slide decks for each conversation Sellers 69
Slide 70: Spend event planning time with customers instead
Spend event planning time with customers instead Sellers 70
Slide 71: Sellers create solution architecture proposals
Sellers create their own solution architecture proposals Sellers 71 71
Slide 72: Sellers position Cloudflare by use case and region
Sellers position Cloudflare by use case and region Sellers 72 72
Slide 73: And, we're just getting started
And, we're just on started | — =——
Slide 74: Stephanie Cohen
Stephanie Cohen Chief Strategy Officer
Slide 75: Cloudflare was built for this moment
Cloudflare was built for this moment The business model Cloudflare is the only cloud Every agentic interaction of the Internet is changing built for the Agentic Internet on the Internet is a chance for commerce …and weʼre already seeing results 75
Slide 76: Platform of choice for the agentic web
Platform of choice for the agentic web 76
Slide 77: We are already seeing results: media sector proof points
We are already seeing results Media sector proof points 37%+ revenue growth 1.3x higher revenue per customer in Q1ʼ26 vs. Q1ʼ25 than company average t 60%+ expansion rate 117% DNR for retained customers in Q1ʼ26 r in Q1ʼ26 e n Increase Win new d Grow existing customers Win New Customers customers
Slide 78: The business model of the Internet is changing
The business model of the Internet is changing 78
Slide 79: AI is a platform shift with unprecedented adoption
AI is a platform shift, and the adoption is unprecedented Years to reach 1 billion users globally Internet 14 years Social Media 8.5 years 2x+ faster Smartphone 6 years adoption of Gen AI than any other preceding platform shift Gen AI 2.5 years Source: Our World in Data / ITU (internet users, 1B
Slide 80: AI adoption is transforming where we spend our time
AI adoption is quickly transforming where we spend our time 2000s Internet Open Web: 100% 2010s Mobile Open Web: 100% 2015 22 Social Open Web: 55% Social:45% Social: 45% 2024 Open Web: 25% Social Social and AI75% and AI 75% AI ERA Share of online attention Note: Illustrative / directional frami
Slide 81: AI is breaking the business model of the Internet
AI is breaking the business model of the Internet Publication Peak Performance Jan 2026 Traffic Decline Mar 2024 8,530,891 264,861 -97% Feb 2024 7,610,480 768,792 -90% Feb 2024 5,322,037 790,002 -85% Feb 2024 1,974,331 293,898 -85% Jul 2024 15,577,298 4,045,783 -74% Nov 2024 7,754,067 2,976,994 -62%
Slide 82: It's not just the media industry
Itʼs not “justˮ the media industry 82
Slide 83: Multiple industries impacted
We are seeing multiple industries impacted Absolute human traffic by Industry June 2025 vs. April 2026 comparison) 35 – 40 % DECLINES Jun 2025 Apr 2026 Jun 2025 Apr 2026 Jun 2025 Apr 2026 Jun 2025 Apr 2026 Financial Retail Computer Software IT & Services Services Jun 2025 Apr 2026 83 Source: Cloudf
Slide 84: The business of the Internet needs saving
And, the business of the Internet needs saving 6+ billion $16+ trillion people connect to the value of the digital economy, Internet, a number that's 15% of global GDP doubled in 11 years 59% 61% of people use the Internet to of users cite information stay connected with friends access as the top re
Slide 85: Mission: Help build a better Internet
Mission: Help build a better Internet 85
Slide 86: The only cloud built for the Agentic Internet
Cloudflare is the only cloud built for the Agentic Internet 86
Slide 87: Supply, technical capabilities and demand
Supply Technical Capabilities Demand Transparency • Control • Optimize • Charge Optimize • Behave • Access • Discover Reliable scarcity Bot Management Infrastructure Filter and manage automated traffic Analytics AI Gateway, Workers, & Workers AI b Value discovery Web Bot Auth u Developer Tools Authe
Slide 88: Supply
® crouvrtare Supply 88
Slide 89: The Internet runs on Cloudflare
The Internet runs on Cloudflare Breadth Depth Importance Uniqueness The worldʼs largest 20% of the Internet 36% of the top 10,000 42% of the developer communities runs on Cloudflare sites run on Cloudflare Fortune 500 run on Cloudflare 89
Slide 90: The economy runs on Cloudflare
The economy runs on Cloudflare MEDIA SOFTWARE & SAAS RETAIL & ECOMMERCE FINANCE & FINTECH GAMING & GAMBLING HEALTHCARE 90
Slide 91: Demand
® crouvrtare Demand 91
Slide 92: Leading AI companies already run on Cloudflare
Leading AI companies already run on Cloudflare too Top 50 generative AI Web products, by unique monthly visits 78 % run on Cloudflare 92 Source: Top 50 Generative AI Web Products based on a16z Consumer report as of March 2026. This includes generative AI Web products only and does not include mobile
Slide 93: Developers increasingly build agents on Cloudflare
Developers increasingly build their agents on Cloudflare Agents SDK Downloads Agent SDK Downloads May 2026 42x 3.2M monthly downloads growth in monthly downloads August 2025 1M+ total 14M downloads downloads since March 2025 Launched launch 93 Source: npm public download statistics for the agents p
Slide 94: Technical Capabilities
® crouvrtare Technical Capabilities 94
Slide 95: The foundational beliefs that built our platform
The foundational beliefs that built our platform Secure by default Fast by default Scalable by default A property of where your code A consequence of the platform Handled by the platform, not runs, not a product you attach to it being everywhere, not a region you capacity you provision and manage pi
Slide 96: Technical capabilities required for a functioning market
Technical capabilities required for a functioning market A payment Reliable system scarcity Neutrality / Crisis Response Analytics a level playing field Streamlined Value workflows discovery 96
Slide 97: Best at identifying, classifying and controlling bots
#1 FUNCTIONING MARKETS REQUIRE RELIABLE SCARCITY We are the best at identifying, classifying, and controlling bots Our competitive advantage is visibility. The more of the Internet we see, the better we identify, classify, and stop unwanted or malicious bots — before they ever reach our customersʼ
Slide 98: Transformation of data to insights
#2 FUNCTIONING MARKETS REQUIRE ANALYTICS Transformation of data to insights is key for business development Bot Management began as a threat intelligence product for security but is now also being used for market intelligence to drive marketing and partnerships decisions. 98
Slide 99: First-party signals identify highest-quality content
#3 FUNCTIONING MARKETS REQUIRE VALUE DISCOVERY First-party signals identify highest-quality content Cloudflareʼs network processes trillions of requests daily, generating unique signals that help partners identify and surface the most relevant information. Demand Traffic Quality Real-time query vol
Slide 100: Easy Button for well-behaving crawlers and agents
#4 FUNCTIONING MARKETS REQUIRE STREAMLINED WORKFLOWS “Easy Buttonˮ for well-behaving crawlers & agents More developer integrations Building agents Creator revenue on Cloudflare /Crawl explosion means your code runs milliseconds enables agents to crawl in a way that aligns with site ownersʼ preferen
Slide 101: Cloudflare does not compete with publishers or AI labs
#5 FUNCTIONING MARKETS REQUIRE NEUTRALITY / A LEVEL PLAYING FIELD Cloudflare does not compete with publishers or AI labs As a neutral third-party, Cloudflare can serve both sides… Unique page views by AI crawlers Neutrality transforms Cloudflare from Multipliers show less access vs. #1 crawler an i
Slide 102: Largest payment network comparison
#6 FUNCTIONING MARKETS REQUIRE A PAYMENT SYSTEM LARGEST PAYMENT NETWORK 900 million 30+ trillion Average transactions per day Average transactions per day 100,000 transactions per second 500+ million transactions per second Source: Largest payment network figures per its public filings and technol
Slide 103: Bridge with the traditional payments ecosystem
#6 FUNCTIONING MARKETS REQUIRE A PAYMENT SYSTEM Cloudflare is the bridge with the traditional payments ecosystem Web Bot Auth enables merchants to cryptographically verify agent identity before processing transactions. PAYMENT NETWORK ADOPTERS Trusted Agent Verification Deliver secure agentic comme
Slide 104: Every agentic interaction is a chance for commerce
Every agentic interaction on the Internet is a chance for commerce 104
Slide 105: The agentic surge is real
The agentic surge is real AI Agent Requests (in billions per day) 1,700% INCREASE in daily AI agent requests on Cloudflareʼs network over one year 105
Slide 106: Machine payments will grow with machine traffic
Machine payments will grow with machine traffic Cloudflare founded the x402 Foundation with Coinbase and Stripe. 2 BILLION x402 can become the standard to connect machine buyers 402 responses served per day and sellers in the request path. on Cloudflareʼs network 402 200 OK AI agent requests Agent
Slide 107: Full stack for agentic commerce
We have the full stack for agentic commerce Access control 107
Slide 108: Every agent request can become a metered event
Every agent request can become a metered event The economic surface expands from content access to recommendations, payments, fulfillment, and audit. Build Discover Recommend Act Verify Settle deploy + configure crawler access ranking + context reserve or buy identity + policy payment + audit Acce
Slide 109: It's already happening
Itʼs already happening 109 109
Slide 110: Large media companies use Cloudflare to do deals
ITʼS ALREADY HAPPENING Large media companies use Cloudflare to do deals 110
Slide 111: People Inc. blocks AI crawlers to get content deals
People Inc. uses the technology from web infrastructure provider Cloudflare to block other AI crawlers, prompting AI players to approach it with content deals. In September 2025, People attributed its decision to leverage Cloudflareʼs technology as a way to get AI companies to the negotiating table,
Slide 112: Startup AI companies use Cloudflare to access the Internet
ITʼS ALREADY HAPPENING Startup AI companies use Cloudflare to access the Internet onact 🡒 🡒 Verify site live 🡒 🡒 ive Millions of websites Check renders R2 Daily diffs Signals API Websites to monitor Daily checks Stores raw results On changes only Sends updates Daily website verification for…
Slide 113: Crawl intelligence lets us price niche content
ITʼS ALREADY HAPPENING Proprietary crawl intelligence lets us price niche content programmatically Crawl Attempts Sample Publisher Sample Health Site AI Content Scraper 23 MM 112 K Niche Content Scraping AI Chat User 120 MM 787 MM AI crawlers are scraping specific Search Bot 24 MM 459 K content (e
Slide 114: We are just getting started
We are just getting started Media sector today Tomorrow… 37%+ revenue growth Content Marketplace in Q1ʼ26 vs. Q1ʼ25 Programmatic monetization of content at scale 60%+ expansion rate Signals Standardized, machine-readable vocabulary for retained customers in Q1ʼ26 for real-time decision-making 1.3x h
Slide 115: Act I + Act III = Act IV
Cloudflare was built for this moment Act I + Act III = Act IV Connect & Protect Build Agentic Internet 115
Slide 116: Mark Anderson
Mark Anderson President of Revenue
Slide 117: What customers are saying
What customers are saying The market has come to Cloudflare 117
Slide 118: The market has come to Cloudflare
The market has come to Cloudflare Then 2024 Now 2026 ↑2.3x 61% 269 52% 173 49% 118 Q4ʼ23 Q4ʼ24 Q4ʼ25 Q4ʼ23 Q4ʼ24 Q4ʼ25 Brand Awareness Trend $1M Large Customers Phase: Market Education Phase: Market Pull ■ “Who is Cloudflare?ˮ ■ AI modernization ■ “Why Cloudflare?ˮ ■ Security transformation ■ “
Slide 119: In the middle of a platform shift
We are in the middle of a platform shift Cloud Social Agents 2000 2010 2025 2007 2022 Mobile AI 119
Slide 120: AI and agents change everything
AI and agents change everything 120
Slide 121: AI has moved from experimentation to urgency
AI has moved from experimentation to urgency 9 months ago Next 12 months AI needs to be enabled but with Autonomous agents will automate a guardrails for businesses double-digit percentage of all work Today AI usage has gone from experimentation to urgency 121
Slide 122: The agentic surge is real
The agentic surge is real 1,700 % INCREASE in daily AI agent requests on Cloudflareʼs network from June 1, 2025 to May 31, 2026 122
Slide 123: Frontier AI models change the attack landscape
Frontier AI models change the attack landscape Average time between vulnerability disclosure confirmed exploitation is collapsing 24hrs 2027 123
Slide 124: Customers need a modern platform for an AI-driven world
Customers Everyone is need a Frontier AI models are building & modern changing running agents platform for the attack an AI-driven landscape world 124
Slide 125: IT budgets and headcounts only marginally increasing
IT budgets and headcounts are only marginally increasing… 2.8 % 1.3 % average IT budget average tech worker growth in 2026 headcount growth in 2026 125 Source: Gartner CIO Agenda, 2026.
Slide 126: CIOs are expanding AI spending significantly
…but CIOs are expanding AI spending significantly 88 % 84 % 72 % 69 % Percent of organizations increasing investment in 2026 digital priorities AI Cyber/ Cloud Application information platforms modernization security 126 Source: Gartner CIO Agenda, 2026.
Slide 127: IT leaders' focus: accelerate AI adoption with security
IT leadersʼ focus initiatives: accelerate AI adoption with security built-in Secure workforce use of AI Govern AI agents Discover and control how employees interact with Secure interactions between human users AI tools (e.g., ChatGPT, Claude) and AI agents Protect applications Build AI securely Defe
Slide 128: A platform that was built for this moment
Customers donʼt need a whole new paradigm to build, adopt, & secure AI. for users They need a platform that was & for agents built for this moment in the first place. 128
Slide 129: Cloudflare was built for this moment
Cloudflare was built for this moment. 129
Slide 130: Only vendor to connect, protect and build AI
Cloudflare is the only vendor to help customers connect, protect, and build AI We offer AI security and AI Category Secure workforce Protect AI-powered infrastructure Govern AI agents Build AI securely use of AI applications products — all on one platform Remote MCP servers, Workers AI, AI Search, C
Slide 131: Building a world-class go-to-market organization
Building a world-class go-to-market organization 131
Slide 132: Building World-Class GTM
Accelerate AE Add stage-appropriate momentum leadership and talent Become more Win with partners first Building Crisis Response systematic and World-Class targeted GTM Sell the March towards platform the enterprise 132
Slide 133: Stage-appropriate leadership and talent
#1 ADD MORE STAGEAPPROPRIATE LEADERSHIP & TALENT Established new GTM leadership team and upleveled sales management and the field Jeff Samuels Chief Marketing Officer Rick Congdon GEO VP, Americas Tony Van den Berge GEO VP, EMEA Goran Risticevic GEO VP, APAC of GTM of account Sayoko Matsumoto GEO
Slide 134: More systematic and targeted approach
#2 BECOME MORE SYSTEMATIC AND TARGETED Improved go-to-market approach to best address the unique needs of our diverse customer set Defining clear market segments allows Majors Developer Speedboat us to apply Enterprise Cloudflare One specific sales Named Speedboat models to Public Sector target cus
Slide 135: Enterprise momentum continues to build
#3 MARCH TOWARDS THE ENTERPRISE Enterprise momentum continues to build… $1M Customers vs. Annualized Revenue 800 700 $1M+ customers 600 represent less $1M Customers 500 than 1% of 400 Cloudflareʼs 300 paying 200 customers. 100 0 $1,000 $2,000 $3,000 $4,000 $5,000 Annualized Revenue $M 135 Sourc
Slide 136: Customers in the Top 100 across all industries
#3 MARCH TOWARDS THE ENTERPRISE …across all industries Customers in the Top 100 Number of Cloudflare Source: FactSet company screening report as of May 2026; screen uses FactSet Revere Business Industry Classification System RBICS) supplemented with Forbes Top 200 US Private Companies. 136 Top 100
Slide 137: Relentless focus on innovation
#4 SELL THE PLATFORM 2022 Relentless focus on innovation Argo for Spectrum Ethereum Gateway Custom Domain Spectrum Gateway Protection 2014 Advanced Area 1 Email Domains API Shield Security Certification Manager Access Regional Services Images DLP Calls R2 DNS Firewall Argo Workers Image Resizing Pa
Slide 138: Connect, protect, and build everywhere
#4 SELL THE PLATFORM Connect, protect, and build everywhere Connect and protect Connect and protect external-facing systems internal-facing systems Cloudflare Network Services Cloudflare Zero Trust Services Cloudflare Application Services Build the next-generation of applications Cloudflare Develop
Slide 139: The power of partnership
#5 WIN WITH PARTNERS FIRST The power of partnership Partners as a % of revenue Partners reach 1000s of customers each PARTNERS CUSTOMERS 139
Slide 140: Physics 101: momentum
#6 ACCELERATE AE MOMENTUM Physics 101 p=mv Momentum, n. Mass, n. Velocity, n. the quantity of motion in a the quantity of matter within a given a quantity that designates speed moving body, expressed as the object together with the direction of product of its mass and velocity movement 140
Slide 141: Sales productivity at new all-time highs
#6 ACCELERATE AE MOMENTUM Productivity Rolling 4Quarter ACV per Account Executive Net We are encouraged by the continued improvement in sales productivity to new all-time highs. 141
Slide 142: Increased hiring of stage-appropriate GTM talent
#6 ACCELERATE AE MOMENTUM Increased hiring of stage-appropriate GTM talent Account Executive Capacity Net Y/Y Growth Composition of New AE Hires We continue to 42% 40% 72% 73% increase hiring in sales — with a focus on enterprise 58% 60% account executives with proven track 28% 27% records. 2022
Slide 143: What's next for 2026 and beyond?
Whatʼs next for 2026 and beyond? 143
Slide 144: Our own most demanding AI customer
We are our own most demanding AI customer 144
Slide 145: AI agents are becoming full-cycle co-sellers
AI agents are becoming full-cycle co-sellers Salesforce: The State of Sales Report – “How Reps Across the industry, sales reps are being Spend Their Time During an Average Workweekˮ stretched thin by two opposing forces: growing customer demands and limited time to deliver. AI agents are increasingl
Slide 146: AI-powered tools transforming sales strategies
AI-powered tools are transforming sales strategies Examples of AI use cases from one end of the sales process to the other: 1 2 3 4 5 6 7 Next best RFP/RFQ Guided Next best Research Meeting Intelligent opportunit responde pricing/ action assistant support coaching y r quoting ■ Lead gen and ■ Recomm
Slide 147: Customer lifecycle reimagined with AI
The customer lifecycle reimagined with AI 1. Onboarding 2. Adoption 3. Renewal Setting the foundation Proving value Solidifying the relationship Standardizing the initial journey for Driving deep platform usage through Ensuring commitment and identifying rapid time-to-value and technical strategic b
Slide 148: Go-to-market has achieved escape velocity
Go-to-market at Cloudflare has achieved escape velocity 148
Slide 149: Thomas Seifert
Thomas Seifert Chief Financial Officer
Slide 150: Traffic growth is accelerating
Traffic growth is accelerating Millions of Requests per Second on Cloudflareʼs Network (increase in Q1ʼ25 vs. Q1ʼ26 1,700% + 600+ bps higher Q/Q growth INCREASE in daily AI agent requests on Cloudflareʼs network from June 1, 2025 to May 31, 2026 150
Slide 151: Cloudflare Revenue Models 101
Cloudflare Revenue Models 101 151
Slide 152: Understanding Cloudflare's revenue models
Understanding Cloudflareʼs revenue models Flat Rate Pure Usage Pool of Funds Cap Model 1 Free, Pro, and 2 3 Enterprise committed 4 Enterprise committed Pay-As-You-Go) Business plans) value contracts) capacity contracts) One fixed price for unlimited/uncapped Billed in arrears for the exact numbe
Slide 153: Cap model provides predictable revenue
CAP MODEL Monthly cap contracts provide predictable revenue for Cloudflare and cost for customers… Illustrative Example of Cap Model New contract New contract with raised cap with raised cap at 12 months at 24 months Why customers select a cap model: ✓ offers predictable budgeting, as long as Reques
Slide 154: Accelerating traffic growth and cap model upsizing
CAP MODEL …but accelerating traffic growth can result in earlier-than- expected increases to cap model contracts Illustrative Example of Cap Model with Accelerated Usage New contract New contract with raised cap with raised cap at 20 months at 9 months When actual usage outpaces a customerʼs initial
Slide 155: Leading technology company increased cap twice in 2025
A Leading Technology Company: Increased monthly cap twice in 2025 Example: 1st Cap 2nd Increase Cap Existing Billable Data Transfer & Cap Increase customer increasing Revenue monthly cap contract Note: FY2023 and FY2024 revenue reflects actual full fiscal-year revenue; Q1'25Q1'26 revenue is annuali
Slide 156: Pool of funds accelerate platform adoption
POOL OF FUNDS Pool of funds accelerate platform adoption… Pool of Funds as a % of New ACV Booked Enterprises are increasingly accustomed to purchasing cloud services through flexible consumption models. 156
Slide 157: Pool of funds impact on shape of financial reporting
POOL OF FUNDS …but can impact the “shapeˮ of financial reporting 2024 PoF Cohort: PoF Revenue 2025 PoF Cohort: PoF Revenue vs. If-Ratable Revenue vs. If-Ratable Revenue Revenue recognition is often lower in the near term but accelerates over course of the contract as customers accelerate adoption an
Slide 158: Leading AI company signed two PoF deals
A Leading AI Company: Signed two PoF deals in 2024 and 2025 Signed 2-yr 215% $85M increase PoF in ACV Example: 39% Existing Q/Q 73% customer Q/Q 27% transitioning to Q/Q pool of funds Signed 1-yr 44% Q/Q 7 order $13.5M PoF 112% forms Q/Q signed in Signed 2023 35% initial vs. ‘24 contract Run-
Slide 159: Drivers of Value Creation
Drivers of Value Creation Growth + Profitability 159
Slide 160: Growth
® crouvrtare Growth 160
Slide 161: Drivers of high, sustainable growth
Drivers of high, sustainable growth Expanding TAM Growing pipeline Improving sales productivity and growing capacity Building momentum with Increasing platform Accelerating traction in larger customers adoption key product areas 161
Slide 162: A consistently expanding TAM
#1 EXPANDING TAM A consistently expanding TAM Cloudflare $384B Developer Services Areas for Potential ● Application Platform Incremental ● Object Storage Growth $329B ● AI & Agent Builder Platforms ● Inference ● AI Observability $282B Cloudflare Network Act IV $238B Services 1 Cloudflare One ● Netw
Slide 163: Pipeline continues to grow
#2 GROWING PIPELINE Pipeline continues to grow Pipeline $M ↑40% Robust pipeline growth gives us confidence to invest in sales capacity. 163
Slide 164: Go-to-market initiatives are paying off
#3 IMPROVING SALES PRODUCTIVITY AND GROWING CAPACITY Go-to-market initiatives are paying off Rolling 4Quarter ACV per Account Account Executive Capacity Net Executive Net Y/Y Growth We continue to invest in quota-carrying sales capacity and expect that productivity to continue to grow for a ve
Slide 165: Large customers are getting larger
#4 BUILDING MOMENTUM WITH LARGE CUSTOMERS Large customers are getting larger Revenue $M) by All Customers $5M 56x 39 42% $1M$5M 55x $500K$1M 37x $100K$500K of the Fortune 500 23x of the Fortune 500 $100K 6x % 2018 2019 2020 2021 2022 2023 2024 2025 165 165 Note: Growth factors (e.g., 6x
Slide 166: Land-and-expand fuels platform adoption
#5 INCREASING PLATFORM ADOPTION Land-and-expand fuels platform adoption Annualized Recurring Revenue by Annualized Recurring Revenue by Product SKU Product Attach Rate Cohorts $M Our success in driving platform adoption is evidenced by the consistently increasing amount spent with Cloudflare in h
Slide 167: Accelerating traction in key product areas
#6 ACCELERATING TRACTION IN KEY PRODUCT AREAS Cloudflare One Developer Cloudflareʼs ↑43% ↑137% multi-act play is only just getting y/y growth in ARR in y/y growth in ARR in started. 2025 2025 167 Note: See the Appendix for how we define “Annualized Recurring Revenueˮ or “ARR.ˮ 2025 ARR calculated a
Slide 168: Profitability
Profitability 168
Slide 169: Unit economics
Unit economics Unit Economic Margin by Component What does it “cost to bookˮ $1 of Cost to Book incremental ARR? CTB Sales & CTB Marketing cost per dollar of incremental ARR How much does it “cost to serveˮ Cost to Serve the lifetime revenue? CTS = CTS Delivery, R&D, Support & G&A cost as % of
Slide 170: Gross margin reporting is influenced by multiple factors
COST TO SERVE COGS Gross margin reporting is influenced by multiple factors Change in Gross Margin by Factor (since Q1ʼ24 – The percentage of paid versus free traffic on our network, which is neutral to operating margin, has grown significantly. – Workers developer platform products, which are cu
Slide 171: How revenue mix influences unit economic KPIs
How revenue mix influences unit economic KPIs Act I Act II Act III Act IV 171
Slide 172: Unit economics define the growth/margin frontier
HOW REVENUE MIX INFLUENCES UNIT ECONOMIC KPIS Unit economics define the Faster growth/margin frontier Value = Revenue - (CTS) - CTB x (Attrition + Growth) Growth Revenue Unit Economic Margin = 44% Slower Lower Higher Unit Economic Margin Note: Illustrative diagram. 172
Slide 173: Unit Economics by Act: Act I
HOW REVENUE MIX INFLUENCES UNIT ECONOMIC KPIS Unit Economics by Act Faster ACT I - COST TO BOOK : AVERAGE - COST TO SERVE : AVERAGE ACT I - ATTRITION : Higher Growth Slower Lower Higher Unit Economic Margin Note: Illustrative diagram. 173
Slide 174: Unit Economics by Act: Act II
HOW REVENUE MIX INFLUENCES UNIT ECONOMIC KPIS Unit Economics by Act Faster ACT II - COST TO BOOK : HIGHER - COST TO SERVE : LOW ACT I - ATTRITION : LOWER Growth ACT II Slower Lower Higher Unit Economic Margin Note: Illustrative diagram. 174
Slide 175: Lower gross margin doesn't mean lower unit economic margin
HOW REVENUE MIX INFLUENCES UNIT ECONOMIC KPIS Lower gross margin doesnʼt mean lower unit economic margin — especially for platform businesses at scale CASE STUDY FOR ACT III COST TO BOOK Early customer acquisition relies on efficient product-led growth PLG, self-serve onboarding, Microsoft Intell
Slide 176: Platform businesses shift growth/margin frontiers
HOW REVENUE MIX INFLUENCES UNIT ECONOMIC KPIS Platform businesses Faster ACT III possess the unique capacity to structurally shift → Growth their growth/margin → frontiers upward and → to the right as revenue scales. → Slower Lower Higher Unit Economic Margin Note: Illustrative diagram. 176
Slide 177: Unit Economics by Act: Act III
HOW REVENUE MIX INFLUENCES UNIT ECONOMIC KPIS Unit Economics by Act ACT III Faster ACT III - COST TO BOOK : LOWER - COST TO SERVE : INITIALLY HIGH, ACT I LOWER AT SCALE Growth - ATTRITION : LOW ACT II Slower Lower Higher Unit Economic Margin Note: Illustrative diagram. 177
Slide 178: Unit Economics by Act: Act IV
HOW REVENUE MIX INFLUENCES UNIT ECONOMIC KPIS Unit Economics by Act Faster ACT IV - COST TO BOOK : LOWEST - COST TO SERVE : LOWEST - ATTRITION : LOWEST Growth Act IV could have unit economics “off the chart” ACT IV Slower Lower Higher Unit Economic Margin Note: Illustrative diagram. 178
Slide 179: What are our financial targets?
What are our financial targets? 179
Slide 180: $5B annualized revenue before year-end 2028
Before year-end 2028 $5B annualized revenue Year-end 2025 $2.4B annualized revenue 180
Slide 181: Our Rule of 40 North Star
Our “Rule ofˮ North Star… 2028 2027 50% + Since IPO 50% 40% 181
Slide 182: AI is driving better cost to book
COST TO BOOK S&M AI is driving better cost to book… Illustrative Example of Ratio-Oriented GTM Model Every GTM metric across the technology Non-Quota Carrying industry has historically scaled based on a MARKETING ratio of employees who carry quota AEs to every other support function within the G
Slide 183: Reinvested in more AEs to produce higher growth
COST TO BOOK S&M …that can be reinvested in more AEs to produce higher growth Illustrative Example of Ratio-Oriented GTM Model — Status Quo vs. with AI Automation From one end of ------------------ Status Quo ------------------ ---------------- w/ AI Automation ---------------- the sales Headcount
Slide 184: Supercharging a 10x engineer into a 1000x engineer with AI
COST TO SERVE R&D Supercharging a 10x engineer into a 1000x engineer with AI ONE ENGINEER ONE WEEK TWO ENGINEERS TWO MONTHS EVERY MERGE REQUEST EVERY REPO Reimplementing Next.js Rebuilding WordPress Reviewing Code across Every Repo Steve Faulkner, Sr. Director of Cloudflare One product manag
Slide 185: AI is transforming G&A workflows
COST TO SERVE G&A AI is transforming G&A workflows Low Maturity Top Performers New Frontier G&A is being rebuilt around Al-augmented workflows 8% 6% ■ Finance & Accounting: continuous Healthcare close, automated reconciliations, Principal investors & private equity real-time variance commentary
Slide 186: Long-term operating model
Long-term operating model Prior Updated Long-Term Long-Term 2022 2023 2024 2025 Model Model Gross 78 % 78 % 79 % 76 % 75% 77% 70% 77% Margin Sales & Marketing 42 % 40 % 38 % 36 % 27% 29% 23% 25% % of revenue) Research & Development 19 % 17 % 16 % 16 % 18% 20% 15% 17% % of revenue) Gene
Slide 187: Achieve GAAP profitability by 2028
Achieve GAAP profitability by 2028 latest 187
Slide 188: Matthew Prince, Michelle Zatlyn, Thomas Seifert, Mark Anderson
Matthew Prince Michelle Zatlyn Thomas Seifert Mark Anderson Co-founder & Chief Co-founder & Chief Financial Officer President of Executive Officer President Revenue 188
Slide 189: Appendix
Appendix
Slide 190: GAAP to Non-GAAP Reconciliation (Annual)
GAAP to Non-GAAP Reconciliation Annual 190
Slide 191: GAAP to Non-GAAP Reconciliation (Quarterly)
GAAP to Non-GAAP Reconciliation Quarterly) 191
Slide 192: Definitions & Methodologies
Definitions & Methodologies Paying Customers. We believe our ability to grow the number of paying customers on our network provides a key indicator of growth of our business and our future business opportunities. We define a paying customer at the end of the quarter as a person or entity who has gen
Slide 193: Definitions & Methodologies
Definitions & Methodologies Annualized Recurring Revenue ARR. Annualized Recurring Revenue, or ARR, is defined as the annualized value of our customer subscription contracts as of the measurement date, assuming any contract that expires during the next 12 months is renewed on its existing terms. A