Cloudflare at $330: Expensive or Inevitable?

There is a version of this letter that simply says: Cloudflare crushed Q2, raised guidance, the stock is up 16%, and you missed it. That version is useless. The more interesting question is what Cloudflare is actually worth right now, whether the market has it right, and where a disciplined investor sits on a day like this. That is what we are going to work through.

What Happened on Earnings Day

Cloudflare reported Q2 2026 results on August 7. Revenue came in at $696.1 million, up 36% year-over-year, about 4.6% ahead of the roughly $665 million consensus. The company had guided conservatively after a rough Q1 restructuring announcement, and it beat that guidance convincingly. Non-GAAP earnings per share landed at $0.29 versus $0.27 expected. Non-GAAP operating income reached $96.1 million, a 13.8% margin and a 33% increase year-over-year. Free cash flow was $56.4 million, or 8% of revenue, up 69% from the same period a year ago.

Guidance went up across the board. The full-year revenue outlook moved to $2.864 to $2.870 billion, from $2.805 to $2.813 billion. Full-year non-GAAP EPS guidance rose to $1.25 to $1.26, from $1.19 to $1.20. Q3 revenue guidance came in at $736 to $737 million. The stock was trading around $330 in premarket, a gain of roughly 16% from the prior close.

None of that is the interesting part.

The Number That Changes the Framing

The figure worth sitting with is this: AI agent requests on Cloudflare’s global network grew more than 1,700% year-over-year. Alongside that, the share of non-human traffic on Cloudflare’s network has crossed 50% for the first time. The company’s own CEO said that threshold arrived roughly 18 months earlier than he had anticipated.

This matters because Cloudflare’s business model was built for a human internet. Subscriptions priced per seat, per domain, per tier. That model still works. But the network it runs on is rapidly becoming something different: an infrastructure layer for automated agents, crawlers, and software systems that retrieve, process, and act on information without a person in the loop.

Cloudflare sits between virtually all of that traffic and the internet it is trying to reach. What the company does with that position is the central question for the next two to three years. And as of Q2 2026, it has started answering it with products rather than promises.

What the Business Is Becoming

Cloudflare’s legacy description is a content delivery network with security baked in. That framing no longer captures what the company actually does, or where its growth is coming from.

The Workers developer platform is the clearest sign of the shift. Workers is Cloudflare’s serverless compute layer, meaning developers can run code directly on Cloudflare’s global network without provisioning servers or managing infrastructure. It now has more than 7.4 million registered developers, with nearly 2 million added in Q2 alone. That single quarter exceeded the total developer additions across all of 2025. When developers build on Workers, their workloads run on Cloudflare’s network, and usage-based billing means revenue scales with the volume of compute events, not just the number of seats sold.

Beyond Workers, Cloudflare has launched a suite of products aimed directly at the economics of machine traffic. New AI crawler controls let customers classify automated requests into three buckets: Search, Agent, and Training. Each can be allowed or blocked independently. Starting September 15, 2026, all newly onboarded domains will have updated defaults that block Training and Agent crawlers on ad-supported pages while continuing to pass Search traffic. That is not a minor product tweak. It is Cloudflare deciding, at the infrastructure level, who gets access to the web and under what conditions.

The Attribution Business Insights dashboard adds another layer. It shows publishers exactly how much of their content AI companies are consuming, and how much human traffic those AI companies actually return in exchange. For the first time, content owners have the data to negotiate. That kind of product creates dependency. Once a publisher is making access decisions based on Cloudflare’s dashboards, removing Cloudflare becomes a significant operational disruption.

The Numbers Behind the Story

• Revenue growth: 36% year-over-year in Q2, up from 34% in both Q1 2026 and Q4 2025
• Large customers ($100K+ annually): 4,698, up 27% year-over-year; 986 net additions over the trailing 12 months, a company record
• Dollar-based net retention: 120%, up 2 points sequentially and 6 points year-over-year
• Workers developers: more than 7.4 million, with nearly 2 million added in Q2 alone
• Current remaining performance obligations: grew 35% year-over-year, providing contracted near-term revenue visibility
• Non-GAAP gross margin: 73.1%, up 30 basis points sequentially; down 320 basis points year-over-year from 76.3%
• GAAP gross margin: 71.8%, compared with 74.9% a year earlier
• Free cash flow: $56.4 million in Q2, or 8% of revenue; Q1 2026 was $84.1 million at 13% of revenue
• Cash and investments: $4.16 billion as of June 30, 2026
• GAAP net loss: $170 million in Q2
• Restructuring charges: $150.7 million recorded in Q2, the primary driver of the GAAP operating loss
• Network capex: 7% of revenue in Q2; full-year 2026 guidance of 14% to 15% of revenue
• GAAP operating loss: $205.7 million in Q2

Is It Cheap? An Honest Assessment.

No. Not at $330. Not even close by traditional measures.

At that price, Cloudflare trades at roughly 40 to 45 times trailing revenue. The broader IT sector trades closer to 2 times sales. Even high-growth software infrastructure peers command a fraction of that multiple. There is no valuation framework rooted in current fundamentals that gets you to $330 without requiring a very specific, very optimistic set of assumptions about what the next three years look like.

The gross margin trend is the most important thing to monitor. Non-GAAP gross margin peaked at 79.5% in Q1 2024. It has since compressed to 73.1%, as the company has shifted more network costs into cost of revenue and absorbed a heavier mix of compute-intensive developer and machine traffic. The Q2 result showed the first sequential improvement in eight quarters, a 30 basis point gain. Management described it as a sign of stabilization. That is a start, not a recovery. Moving the margin back toward the mid-70s range the company has historically targeted is what a genuine re-rating requires.

The GAAP net loss of $170 million looks alarming until you separate the restructuring charge of $150.7 million from the core business. Strip that out and the operating picture improves substantially. The non-GAAP operating margin of 13.8% is real and expanding. The GAAP gap is a timing artifact, not evidence of a deteriorating business.

So the honest framing is this: Cloudflare is not cheap by any measure we normally use here. What it might be is reasonably priced relative to what it could become, if the agentic monetization model delivers. That is a very different kind of valuation question, and it requires a different kind of patience.

Bull / Base / Bear

Bull: The 1,700% growth in AI agent traffic is early innings on a curve that compounds quickly. The AI crawler controls, Pay Per Use licensing, and Attribution Business Insights dashboard convert machine traffic into per-request revenue at a scale the current financials do not yet reflect. Gross margins recover toward the mid-70s as the restructured cost base settles and Workers consumption workloads, which carry better economics than raw network transit, become a larger share of revenue. The $5 billion revenue target for 2028 that management has discussed becomes credible ahead of schedule. The stock re-rates on that recognition.

Base: Revenue growth holds in the low-to-mid 30s through 2027. Large customer additions remain healthy. The usage-based model introduces some quarterly variability but does not fundamentally change the growth trajectory. Gross margins stabilize near current levels and improve slowly. Restructuring charges clear over the next two or three quarters, closing the GAAP/non-GAAP gap. Agentic monetization tools gain traction but are not yet breaking out as a separately disclosed contributor.

Bear: The monetization of machine traffic takes longer than the product roadmap implies. Gross margin erosion resumes, driven by compute-intensive agentic workloads that cost more to serve than human page requests. The roughly 20% workforce reduction announced in Q1 2026 creates friction in enterprise sales at the moment when the largest contracts matter most. Any guidance shortfall reprices the stock sharply. At 40-plus times sales, there is very little margin for error.

Cheap Investor Scorecard

• Business Quality: Strong. A global network with genuine scale advantages. The more traffic that flows through it, the harder it is to displace. Network effects are real here.
• Financial Strength: Moderate. $4.16 billion in cash is a substantial buffer. GAAP losses are restructuring-driven. Core non-GAAP unit economics are healthy and improving.
• Revenue Growth: Strong. 36% in Q2, accelerating from 34% in Q1. The Q3 guide does not imply a slowdown.
• Gross Margin Trend: Watch carefully. 73.1% non-GAAP, with the first sequential uptick in eight quarters. Still well below the 79.5% peak. This is the single most important number to track each quarter.
• Competitive Position: Strong. Cloudflare occupies a position between AI agents and the internet that is genuinely difficult to replicate at comparable scale. It is not the only player, but it is the most pervasively deployed.
• Customer Expansion: Strong. Dollar-based net retention at 120%, up from 114% a year ago. Record large-customer net additions over the trailing 12 months.
• Cash Flow: Moderate. 8% free cash flow margin in Q2, down from 13% in Q1. The restructuring is the primary cause. Confirm the rebound as charges roll off.
• Valuation: Expensive. 40 to 45 times trailing price-to-sales. There is no margin of safety at current prices relative to current earnings power. The bet is entirely on future execution.
• Catalyst Strength: Strong. September 15 default changes, Pay Per Use model in market, new agentic monetization tooling live. Concrete milestones, not roadmap promises.
• Margin of Safety: Low. At $330, one bad quarter ends the conversation. The stock has already shown it punishes deceleration without hesitation.

Action Plan

Buying a 16% gap-up on a stock trading at 40-plus times sales is not the discipline this publication is built around. The earnings were good. The company is executing. Neither of those facts makes $330 the right entry for a patient investor who wants a margin of safety.

• Existing holders: Hold. The Q2 results validate the restructuring decision. Trimming at these multiples is reasonable portfolio management, not a loss of conviction. The business is performing.
• New buyers: Wait. The $295 to $305 range was meaningful resistance before the earnings report. A pullback to that zone, which would still represent a substantial premium to a year ago, gives you a far more defensible entry. The thesis does not expire if you miss the first 10 points off the low.
• Scale-in framework: If you want exposure, consider building in three steps. First position at $295 to $305 on a pullback. Second position after Q3 confirms gross margin holds at or above 73%. Third position only if September 15 default changes generate measurable publisher adoption and Cloudflare begins disclosing early revenue signals from Pay Per Use.
• The date that matters most: September 15, 2026. New default crawler settings take effect for every domain newly onboarded to Cloudflare. It is the first real test of whether the gatekeeper position has commercial teeth or is still theoretical infrastructure waiting for a business model.

Cheap Investor Checklist

1. Gross margin direction: The single most important variable. 73.1% in Q2, up 30 basis points. Does Q3 confirm the inflection or was it noise?
2. Revenue growth pace: 36% in Q2. Watch whether Q3 guidance implies continuation, acceleration, or a quiet deceleration toward the low 30s.
3. Workers monetization: Developer count at 7.4 million is impressive. But developer counts are vanity metrics until they show up in consumption revenue. Look for concrete usage billing disclosures.
4. Dollar-based net retention: 120% in Q2, up 6 points year-over-year. This is the clearest signal that existing customers are expanding. Watch for any reversal.
5. Large customer net adds: 986 over 12 months, a record. The absolute number is what matters, not the percentage growth on an expanding base.
6. Current RPO growth: 35% year-over-year in Q2. Healthy backlog growth relative to revenue growth is a sign of demand ahead. Watch this spread each quarter.
7. Free cash flow margin: 8% in Q2 after restructuring noise. Confirm it normalizes above 10% as charges clear in Q3 and Q4.
8. Agentic revenue disclosures: Does any Q3 disclosure mention revenue contribution, attach rate, or take rate from the new crawler controls or Pay Per Use model? That is the most valuable new data point in the model.
9. Restructuring completion: $150.7 million charged in Q2. Confirm it was the peak quarter and that Q3 operating losses narrow materially toward the non-GAAP picture.
10. September 15 response: Do publishers adopt the new default controls and use them to monetize agent and training crawlers? Or do they opt out? Early signals will tell you whether the toll booth collects tolls.

Bottom Line

Cloudflare reported a quarter that answered several doubts at once. Revenue growth reaccelerated. Net retention climbed to 120%. The restructuring charge is largely behind the company. Gross margin showed its first sequential improvement in two years. The Workers platform added more developers in a single quarter than it did in all of 2025. And the company placed a concrete bet on September 15 that its network position gives it the right to control, and eventually charge for, how AI agents access the web.

None of that makes the stock cheap at $330. The multiple demands a future that has not yet arrived in the revenue line. Investors who buy here are betting on an infrastructure tolling model that is currently a product announcement, not a revenue category.

The bargain hunter’s discipline is to respect what the business is building while refusing to overpay for what it has not yet proven. A pullback to $295 to $305 changes the calculus. Three strong data points from Q3, on gross margin, on Workers consumption, and on September 15 publisher adoption, change it further. Until then, the right posture is watchful, not impatient.

Quality at a reasonable price is the standard. Cloudflare clears the quality bar without argument. The price question, at $330, does not resolve in the buyer’s favor. Not yet.

The Cheap Investor