Cloudflare
This is the archived full research. Read the story
A San Francisco company whose network sits in front of about a quarter of all websites, blocking attacks, speeding pages up and running code for developers.
The scores
Growth4/4 ↑
Revenue grew 29.8% in 2025 and 35.9% in Q2 2026, contracts signed ahead rose 38% and the 2026 forecast was raised twice.
Business quality3/4 →
In front of most identifiable proxied sites with no customer above 10%, but gross margin is falling and the clouds sell the same.
Financial position3/4 →
About $869M of net cash and free cash flow positive, but $4.2B of convertible notes, much of it in the money.
Execution record3/4 →
Growth near 30% for five years and forecasts raised, but its own errors caused outages and a 20% staff cut this year.
Per share capture2/4 →
Stock pay of 20.2% of revenue exceeds free cash flow, and diluted shares rose about 3% a year since 2021.
Potential3/4 ↑
US revenue is about 15% of a large-firm market built from units, with agent traffic a new use not yet in the accounts.
Risk profile3/4 →
No customer above 10% and net cash, but outages it caused, larger cloud rivals and founder voting control.
Valuation1/4 ↓
About 34.8 times sales, 720% of a three-stage value near $37.32 at an 8% return: priced for perfection, beyond the 20% margin.
Evidence quality3/4 →
Filed figures carry growth and cash, but agent revenue is unreported and the market share comes from an outside web survey.
If it works 1.0× · if it fails you keep about 90%.
The Verdict
We rate Cloudflare Unattractive. Its network sits in front of about a quarter of all websites and revenue grew 36% in the second quarter, so the business is solid. But share pay exceeds its spare cash, its bonds turn into shares below today's price and the gross margin keeps falling. At about 51 times sales, the price needs revenue to grow about 28% a year for ten years, and sits far above our value and our best case. As a bet on tools for AI agents it is defensive: even a big success leaves the company worth about today's price.
The long caseIf agents start paying sites for what they read, someone must vouch for each agent and collect a tiny fee per request, and Cloudflare already guards much of the web.
How we weighed it
The price against our value
Against it
Price at 720% of our value
The price sits far above our value and above the bull case.
The business
In its favour
Solid
Its network sits in front of a large share of all websites.
Growth against what the price assumes
In its favour
Growth sped up beyond the line we set.
What could go wrong
Against it
Share pay is heavy, its bonds convert into new shares and the gross margin keeps falling.
Against competitors
Against it
Far dearer than its two filed rivals, a gap that charges for all its growth and more.
How its group is behaving
Cuts both ways
Security names move together on breach and spending news.
The latest news and report
In its favour
The last report went against our view on growth.
The evidence under this - 5 claims
- A sale at 15 times revenue instead of 10 would cut the needed growth to about 23% a year.
- Shares issued for the 2026 notes in August are not yet on a filed cover, so the value is slightly understated.
- He called the figures unaudited and directional, and gave the Visa figure from memory.
- Cloudflare blocks much AI crawler traffic for its customers and does not charge for blocking it, so a request seen is not a request paid.
The company
Cloudflare sells one network that sits between a website and its visitors. Traffic passes through Cloudflare first, which blocks attacks, keeps copies of pages close to visitors so they load fast, and lets developers run their code on the same machines.
About 332,000 companies pay for it, and no customer brings in a tenth of revenue. The alternative is buying separate boxes and services from several suppliers, or using the tools inside a big cloud. It spends about 15% of revenue a year building out the network.
The evidence under this - 4 claims
How it makes money
Customers pay a subscription or by usage, and most sign one- to three-year contracts. Large customers paying over $100,000 a year are 4,698 and bring in 73% of revenue. Once a company routes its traffic through Cloudflare it tends to add products: existing customers spent 20% more than a year earlier. Contracts signed but not yet earned reached $2.73B, up 38%, faster than revenue.
The evidence under this - 4 claims
- 2026-11-05editedQ3 2026 10-Q: 5,006 customers over $100,000, up 24.9% (was 4,698, up 26.6%)
- 2026-11-05editedQ3 2026 10-Q and release: remaining performance obligations $2,876.8M (was $2,732.0M)
- The rate fell to 111% at the end of 2024 before recovering, so it moves with spending cycles.
- It nets customers who leave against those who expand; Cloudflare does not report gross retention.
- 2026-11-05editedQ3 2026 10-Q: net retention 122% (was 120% in Q2)
Where the money comes from, stream by stream
| Stream | Share of sales (Q2 2026) | How it earns it |
|---|---|---|
| Customers paying over $100,000 a year | 73% | Enterprise contracts of one to three years, increasingly pool-of-funds deals customers draw down across products, for security, delivery, office access and Workers. |
| Other paying customers | 27% | Smaller enterprise contracts and self-serve Pro and Business plans paid monthly or yearly by card, plus usage fees from developers. |
| Revenue by product group | not disclosed | Cloudflare reports no split between security, office access, the developer platform and AI; management says Workers grows fastest. |
Who runs it
Matthew Prince, a co-founder, has run Cloudflare since 2009, with co-founder Michelle Zatlyn as president and Thomas Seifert as finance chief since 2017. Insiders hold about half the votes through ten-vote shares. There is no cash bonus; the founders' share awards vest only if the share price reaches set goals, and nothing pays on cash flow or official profit. Insiders sold and bought nothing over the year, almost all under pre-set plans.
The evidence under this - 4 claims
Why it wins, and who could beat it
Why customers come back
The network gets better as it grows. Seeing traffic for a quarter of all websites lets Cloudflare spot an attack on one site and block it for every other, which a smaller rival cannot match. An outside survey finds it in front of 85.5% of the websites whose shield it can identify, and its share of websites using it to deliver pages rose in a year.
Leaving is costly, too. A US data breach costs a company about $10.22M on average, so few want to switch the gatekeeper that works.
The evidence under this - 4 claims
- The rate fell to 111% at the end of 2024 before recovering, so it moves with spending cycles.
- It nets customers who leave against those who expand; Cloudflare does not report gross retention.
- 2026-11-05editedQ3 2026 10-Q: net retention 122% (was 120% in Q2)
The threat
Akamai and Fastly sell the same security and delivery services to the same buyers. The bigger danger is Amazon, Google and Microsoft, whose clouds offer similar tools inside services customers already use. Zscaler competes in one product group, safe office access. Cloudflare's own annual report names no rival and warns that large cloud and telecom firms could move in.
The evidence under this - 4 claims
The rivals: AKAM, FSLY, AMZN, GOOGL, MSFT, ZS, V, MA.
Who else does this, rival by rival
| Ticker | Rival | Where it competes | How far along | Priced against it |
|---|---|---|---|---|
| AKAM | AkamaiPriced against it: Yes | Cloudflare's single reportable segment: network services sold by subscription and by usage | On sale | Yes |
| FSLY | FastlyPriced against it: Yes | Cloudflare's single reportable segment: network services sold by subscription and by usage | On sale | Yes |
| AMZN · GOOGL · MSFT | Amazon Web Services, Google Cloud and Microsoft AzurePriced against it: One segment only | Cloudflare's single reportable segment: network services sold by subscription and by usage | On sale | One segment only |
| ZS | ZscalerPriced against it: One segment only | Cloudflare's single reportable segment: network services sold by subscription and by usage | On sale | One segment only |
| V · MA | Visa and Mastercard's own agent protocolsPriced against it: One segment only | Agent identity and payments | On sale | One segment only |
How big it could get
How big this could get
The United States has about 21,000 firms with 500 or more staff. If each paid what Cloudflare's average large customer pays, about $433,000 a year, that market would be about $9.1B.
Cloudflare already bills US customers about $1.34B a year, roughly 15% of it. That leaves room, and the estimate leaves out every other country, smaller firms, developers and any fees from AI agents.
The evidence under this - 5 claims
The average hides a wide spread: some large customers pay over $5M a year and many just over $100,000.
- Large customers are defined by spend, not staff, so many are smaller firms and some large firms will never buy.
- It leaves out every other country, smaller customers, developers and agent fees.
US revenue includes small customers and developers, so it overstates the share of large firms' spending.
Where the growth is coming from, business by business
Where the growth is coming from
| Business | Share of revenue | Growth: last quarter | Last half-year | Last year | Share in 3 years at its half-year rate |
|---|---|---|---|---|---|
| Reportable Segment | 100% | +39.2% | +37.6% | - | 100% |
| Whole company | 100% | +39.2% | +37.6% | - | 100% |
On the filed split through 2026-09-30, if each business keeps its last half-year's growth, the company grows about 38% next year and 38% in year 3, against 33.5% over the last twelve months. A scenario, not a forecast: no business holds its rate.
The results
The numbers
Sales grew 36% over the twelve months to Sep 2026, faster than the 30% in the year to Dec 2025.
Operating margin is −14%, about where it was (−14%) in the year to Dec 2023.
Running costs take 86¢ of each sales dollar, down from 97¢ in the year to Dec 2021.
Shareholders' money lost 17% a year, up from −32% in the year to Dec 2021.
It turned 13% of sales into free cash, $359M over the last twelve months.
It holds $5.13B of cash against $4.44B of debt, $687M more cash than it owes.
The share count rose 14% in 4.8 years, and pay in shares costs 21% of sales.
All the numbers
Filed figures from the company's 10-K and 10-Q, read 2026-11-10. Fiscal years end in December.
| Dec 2021 | Dec 2022 | Dec 2023 | Dec 2024 | Dec 2025 | 12 mo to Sep 2026 | |
|---|---|---|---|---|---|---|
| Size and growth | ||||||
| Sales | $656M | $975M | $1.30B | $1.67B | $2.17B | $2.73B |
| Sales growth | +52% | +49% | +33% | +29% | +30% | +36% |
| Profit | ||||||
| Gross margin | 78% | 76% | 76% | 77% | 75% | 72% |
| Operating margin | −19% | −21% | −14% | −9.3% | −9.6% | −14% |
| Net margin | −40% | −20% | −14% | −4.7% | −4.7% | −8.7% |
| What growth costs, per sales dollar | ||||||
| Research and development | 29% | 31% | 28% | 25% | 24% | 23% |
| Sales and marketing | 50% | 48% | 46% | 45% | 42% | 42% |
| General and admin | 18% | 18% | 17% | 17% | 18% | 16% |
| Pay in shares | 14% | 21% | 21% | 20% | 21% | 21% |
| Capital spending | 14% | 15% | 8.8% | 11% | 15% | 12% |
| Cash | ||||||
| Cash from operations | $64.6M | $124M | $254M | $380M | $603M | $682M |
| Free cash flow | −$28.3M | −$20.0M | $140M | $195M | $287M | $359M |
| Balance sheet | ||||||
| Cash and short-term investments | $1.82B | $1.65B | $1.67B | $1.86B | $4.10B | $5.13B |
| Debt | $1.16B | $1.44B | $1.28B | $1.29B | $3.27B | $4.44B |
| Net cash | $663M | $214M | $390M | $569M | $836M | $687M |
| Returns and shares | ||||||
| Shares (diluted) | 312M | 326M | 334M | 341M | 348M | 357M |
What growth costs
Growth has become cheaper. From 2021 to 2025 selling and research costs fell as a share of revenue, so the official loss narrowed from about a fifth of revenue to about a tenth. The cost the adjusted figures leave out is pay in shares, which rose to about a fifth of revenue.
The evidence under this - 3 claims
The four quarters include $27.9M of stock pay from the restructuring, which will not recur.
The last report (5 Nov)
Quarter ended September 2026, filed figures.
Revenue grew 39% to $782.5M, well above the $737M top of the forecast, and the forecast for the year rose by about $97M to $2,963M. Existing customers spent 22% more than a year earlier, and customers paying over $100,000 a year grew 25% to 5,006.
Profit measures moved less well. Gross margin fell to 71.2% from 74.0%, the loss under standard accounting widened to $67.9M, and free cash flow was 10% of revenue against 13% a year earlier.
What it did to the thesisIt went against our view on growth: faster than we had set as the line. The price side of our case stands, since pay in shares still exceeds spare cash and the shares cost far more than our cheaper route. Unattractive holds.
The quarter in detail
Sales growth from the previous quarter, filed figures.
Earnings per share $0.38 against $0.34 expected.
The evidence under this - 4 claims
- The rate fell to 111% at the end of 2024 before recovering, so it moves with spending cycles.
- It nets customers who leave against those who expand; Cloudflare does not report gross retention.
- 2026-11-05editedQ3 2026 10-Q: net retention 122% (was 120% in Q2)
- Q2 revenue of $696.1M beat the $664M to $665M guide by about $31M, so the guide has sat well below results.
- The finance chief says pool-of-funds and usage deals make quarters harder to forecast, so a miss is possible even in a strong year.
- 2026-11-05editedQ3 2026 release: full-year guide raised by about $97M to $2,963M to $2,964M; Q3 beat the top of its guide by $45.5M
- The explanation is management's; AI inference and computing products may carry lower margins than security, and no product margins are disclosed.
- The adjusted gross margin rose 0.3 points from Q1, the first rise in eight quarters.
- 2026-11-05editedQ3 2026 release: GAAP gross margin 71.2% (was 71.8% in Q2), still falling
What management did with the money
Cloudflare pays no dividend and buys back no shares. It borrows through bonds that pay no interest and can turn into shares: in August it sold $2.175B due 2031 to repay $1.29B that fell due. It keeps buying small companies to strengthen its developer tools, most recently Deno, on undisclosed terms.
Insiders in 2026Q3: 28 purchases and 332 sales, from their Form 4 filings.
The evidence under this - 3 claims
- 2026-11-05editedQ3 2026 release and 10-Q: cash and securities $5,128.7M after the August refinancing; the 2026 notes no longer outstanding
The price, and where we disagree
In short. Priced for perfection. The price is about 720% of what we think the business is worth, at an 8% yearly return.
The shares are near their high for the year. They more than doubled from April, mostly because buyers now pay far more for each dollar of sales, not because sales grew that much. Analysts have followed, raising ratings and targets.
Cloudflare is building tools that let websites charge AI agents and agents pay, but no revenue is attached yet. The next test is third-quarter results on 29 October, against a forecast of $736M to $737M.
The evidence under this - 5 claims
- The business did speed up: growth rose to 36%, net retention to 120% and the forecast twice (C05, C04, C06), which justifies some re-rating.
- In April the shares sold for about 27 times sales, below the 28 times six-year normal, so part of the rise repaired a fall.
- Q2 revenue of $696.1M beat the $664M to $665M guide by about $31M, so the guide has sat well below results.
- The finance chief says pool-of-funds and usage deals make quarters harder to forecast, so a miss is possible even in a strong year.
- 2026-11-05editedQ3 2026 release: full-year guide raised by about $97M to $2,963M to $2,964M; Q3 beat the top of its guide by $45.5M
- The chief executive says the payment system must handle about 10 million transactions a second, far beyond any payment network today, which is unproven.
- The products are announced, not reported, so their uptake cannot be checked.
How we value it
- Method
- Cash the business will produce, with any shortfall raised in new shares
- Required return
- 8%
- Long-run growth
- 4%
- Moat
- An average moat
| Bear | Base | Bull | |
|---|---|---|---|
| Sales growth, years 1 to 3 | 13.4%, 14%, 11.3% | 26.7%, 27.9%, 22.6% | 33.4%, 34.9%, 28.3% |
| Normal operating margin | 14.4% | 18% | 19.8% |
| Return on new capital after year ten | 8% | 12% | 15% |
| Value a share | $14.28 | $37.32 | $57.04 |
- Price
- $269
- Priced on
- 6 Nov
- Price as a share of the base value
- 720%
- Yearly return the price offers on the base case
- 4.6%
- Risk check: valued at 12%, the rate for a business not yet making a profit
- $15.86
Where each figure comes from
| Figure | Source | Where exactly |
|---|---|---|
| Sales, the last twelve months | Filed report | Quarterly report (10-Q) for the period to 30 Sep 2026, filed 5 Nov 2026; the four quarters to that date. |
| Operating margin today | Filed report | Quarterly report (10-Q) for the period to 30 Sep 2026, filed 5 Nov 2026; the four quarters to that date. |
| Tax rate | Our method's rule | The normal rate where no rate is filed to read: 25% for a US company, 21% elsewhere. |
| Sales per dollar of new capital | Our method's rule | Our default of 2, since the filings give no operating capital to read. |
| Shares | Filed report | Quarterly report (10-Q) for the period to 30 Sep 2026, filed 5 Nov 2026; the diluted share count. |
| Cash and investments | Filed report | Quarterly report (10-Q) for the period to 30 Sep 2026, filed 5 Nov 2026; the balance sheet on that date. |
| Debt | Filed report | Quarterly report (10-Q) for the period to 30 Sep 2026, filed 5 Nov 2026; the balance sheet on that date. |
| Sales growth, years 1 to 3 | Analysts' forecasts | FMP's consensus for the years to 31 Dec 2026, 31 Dec 2027 and 31 Dec 2028, as it stood when the value was struck on 10 Nov 2026: the average of 12 to 26 analysts' revenue estimates, 26.7%, 27.9%, 22.6%. FMP publishes how many analysts, not who they are. |
| Normal operating margin | Analysts' forecasts | FMP's consensus operating margin for the year to 31 Dec 2028, as it stood when the value was struck on 10 Nov 2026, the average of 25 analysts' estimates, less stock pay and less the 3% the analysts add back on their adjusted basis: 18% in the base case. FMP publishes how many analysts, not who they are. |
| Return on new capital after year ten | Our method's rule | Set by our reading of the moat: the 8% rate with none, 12% with an average one, 15% with a strong one; the bear case earns the rate. |
| The bear and bull cases | Our method's rule | Our method's spread: the bear case takes half the growth and a fifth off the margin, the bull a quarter more growth and a tenth more margin. |
| Price | Market price | FMP's closing price on 6 Nov 2026. |
What each case rests on
Bear caseGrowth runs at half the analysts' path, 13%, 14% and 11%, the margin settles a fifth lower at 14.4%, and new capital earns only the 8% rate: the downturn test.
Base caseRevenue grows 27%, 28% and 23% over three years, the analysts' path (C50), then fades toward 4%, closing 30% of the gap each year; the margin moves to 18.0%, the analysts' year-three figure after stock pay and on the filed basis (3.0% less), with stock pay, today 20.7% of sales, at a mature 6%.
Bull caseGrowth runs a quarter above the analysts' path, 33%, 35% and 28%, the margin reaches 19.8%, and new capital earns 15%.
What moves the base value
| Lower | Higher | |
|---|---|---|
| Growth in years 1 to 3, a quarter lower or higher | $29.64 | $46.60 |
| The normal margin, 15% lower or higher | $31.26 | $43.42 |
| Capital per dollar of growth, a quarter more or less | $36.29 | $38.01 |
| Return on new capital after year ten, at the rate or one step up | $28.22 | $40.97 |
| The rate, a point higher or lower | $28.62 | $51.98 |
| Long-run growth a point lower | $33.12 | - |
Value a share as growth and margin change
| Normal margin ↓ · growth → | 50% | 75% | 100% | 125% | 150% |
|---|---|---|---|---|---|
| 14.4% | $18.42 | $23.29 | $29.30 | $36.63 | $45.40 |
| 16.2% | $20.88 | $26.47 | $33.27 | $41.60 | $51.64 |
| 18% | $23.40 | $29.64 | $37.32 | $46.60 | $57.91 |
| 19.8% | $25.98 | $32.82 | $41.39 | $51.74 | $64.19 |
| 21.6% | $28.57 | $36.01 | $45.46 | $56.88 | $70.61 |
Growth is shown as a share of the base case's. Green: the price is at or under 80% of that value. Red: at or over 120%.
How we read the multiple
Investors pay about $127.6B for the company. What has to be true:
1. Revenue over the last year was $2.51B, so the price is about 51 times sales. 2. A buyer wanting 9% a year, a common yardstick, needs revenue near $30.2B in ten years. 3. That is growth of about 28% a year, every year, for a decade.
Revenue is growing about 32% this year. Few companies keep that up for ten years, and the spare cash today is smaller than the pay in shares.
Against its own history. The price per dollar of sales is well above its own recent normal, so buyers are paying for a faster future, not the business as it has been.
The evidence under this - 5 claims
- A sale at 15 times revenue instead of 10 would cut the needed growth to about 23% a year.
- Shares issued for the 2026 notes in August are not yet on a filed cover, so the value is slightly understated.
- Q2 free cash flow carried $99M of severance paid in the quarter, so the run rate is understated by a one-off.
- Network spending ran at 7% of revenue in Q2 against a 14% to 15% full-year guide, so the second half will spend more.
The four quarters include $27.9M of stock pay from the restructuring, which will not recur.
Priced against its rivals
The competitors its own annual report names, each confirmed against the rival's filing. Measured on company value for each $1 of sales, cheapest first:
| Ticker | Company | company value for each $1 of sales |
|---|---|---|
| AKAM | Akamai Technologies | 5.4× |
| FSLY | Fastly | 7.1× |
| NET | Cloudflare | 51.7× |
The middle of the other 2 is 6.2. Cloudflare, Inc. is dearest of 3.
Rivals' figures come from a data provider, so read the spread of the group rather than any single pair.
The gap to its two filed rivals is Cloudflare's own: it grows far faster than both, and the price charges for all of that and more.
Private or foreign rivals that cannot be priced: Vercel.
The bet: what it pays if it works
The bet is on Cloudflare's tools for AI agents: proof of who an agent is, which Visa and Mastercard use, and a way for sites to charge agents a tiny fee per visit. None reports revenue, and charging is in a closed test.
Counting the rest of Cloudflare at a fair price leaves a slice of today's price for these tools. Even if agents paid on a tenth of all requests, as the boss hopes, that is worth about the slice: an expensive bet, little lost if it fails.
| Agent identity and payments revenue in two to three years | Slices of 1% of the network's requests needed | At the stated multiple |
|---|---|---|
| $150.00M | about 3 at 100% kept | 0.9× |
| $470.00M (the big win) | about 10 at 100% kept | 1.0× |
| $1.40B | about 30 at 100% kept | 1.2× |
To pay 2× (a Fair bet) it would need about $5.72B of Agent identity and payments revenue (about 121 slices of 1% of the network's requests); to pay 5× (a Cheap bet), about $21.32B of Agent identity and payments revenue (about 451 slices of 1% of the network's requests), at the same multiple.
Each row is Agent identity and payments alone, plus the rest of the company (Network services (security, delivery, office access, developer platform)) counted at about $117.00B.
The steps it has to take
- Done Visa's Trusted Agent Protocol built on Web Bot Auth; Mastercard Agent Pay adopts it
- Done Monetization Gateway, wallets and cloudflare.pay announced; pay per crawl in closed beta
- Not started Cloudflare discloses revenue or paid volume from agent identity or payments due 2027-02-28
- Not started Pay per crawl opens to all customers due 2027-06-30
- Not started Agents pay for about 10% of the network's requests due 2028-12-31
The detail behind the label
The evidence under this - 7 claims
- Mastercard's 10-K describes Agent Pay without naming Cloudflare, and Web Bot Auth is an open protocol others can use.
- No revenue from this is disclosed.
- The chief executive says the payment system must handle about 10 million transactions a second, far beyond any payment network today, which is unproven.
- The products are announced, not reported, so their uptake cannot be checked.
- He called the figures unaudited and directional, and gave the Visa figure from memory.
- Cloudflare blocks much AI crawler traffic for its customers and does not charge for blocking it, so a request seen is not a request paid.
- No fee for agent payments is published; Cloudflare's cut of what a site charges could be far below or above the Workers price.
- The request count is the chief executive's unaudited estimate.
- A sale at 15 times revenue instead of 10 would cut the needed growth to about 23% a year.
- Shares issued for the 2026 notes in August are not yet on a filed cover, so the value is slightly understated.
Where we disagree with the market
Cloudflare is the network AI agents will run on, and the jump in its shares is the start of a decade of fast growth.
Cloudflare is a strong business growing about 30%, priced as if agent revenue had already arrived and lasted ten years.
The market has moved ahead of the numbers. Agent traffic is real, but Cloudflare reports no revenue from it, and pay in shares still costs more than the spare cash.
The evidence against us is strong: growth is speeding up, existing customers are spending more, and the forecast has been raised twice. A 2027 forecast of 35% growth or more, with more spare cash, would show the price is earned.
What analysts say, and the evidence
The evidence under this - 9 claims
- None of the three figures is in a filing; the chief executive said he had called the crossing date wrong at every point.
- Much non-human traffic is blocked, and Cloudflare says it does not charge customers more for blocking it.
- The chief executive says the payment system must handle about 10 million transactions a second, far beyond any payment network today, which is unproven.
- The products are announced, not reported, so their uptake cannot be checked.
Asked directly for the size of the Workers business on the Q2 call, management declined to give it.
- Q2 revenue of $696.1M beat the $664M to $665M guide by about $31M, so the guide has sat well below results.
- The finance chief says pool-of-funds and usage deals make quarters harder to forecast, so a miss is possible even in a strong year.
- 2026-11-05editedQ3 2026 release: full-year guide raised by about $97M to $2,963M to $2,964M; Q3 beat the top of its guide by $45.5M
- The rate fell to 111% at the end of 2024 before recovering, so it moves with spending cycles.
- It nets customers who leave against those who expand; Cloudflare does not report gross retention.
- 2026-11-05editedQ3 2026 10-Q: net retention 122% (was 120% in Q2)
- Q2 free cash flow carried $99M of severance paid in the quarter, so the run rate is understated by a one-off.
- Network spending ran at 7% of revenue in Q2 against a 14% to 15% full-year guide, so the second half will spend more.
The four quarters include $27.9M of stock pay from the restructuring, which will not recur.
What could go wrong, and how we would know
The risks
- Heavy pay in shares. Pay in shares was $507.7M over the last year, about a fifth of revenue and more than the spare cash the business produced.
- New shares from its bonds. Its 2030 bonds turn into shares at about $248, below today's price, so owners will be diluted if the price holds.
- Gross margin keeps falling. Gross margin fell from 77.3% in 2024 to 71.8% in Q2 2026 as network costs rose.
- The big clouds compete. Amazon, Google and Microsoft sell similar tools inside services customers already pay for, and are far larger.
- Its own outages. Cloudflare says its own errors have caused network outages and data leaks before, and each one dents trust in the gatekeeper.
- A price built on hope. The shares more than doubled in six months, mostly on a higher price per sales dollar, so a disappointment could cut them sharply.
The evidence under this - 8 claims
The four quarters include $27.9M of stock pay from the restructuring, which will not recur.
- 2026-11-05editedQ3 2026 10-Q cover page: 358.4M shares outstanding at 22 October (was 356.1M at 21 July)
- The explanation is management's; AI inference and computing products may carry lower margins than security, and no product margins are disclosed.
- The adjusted gross margin rose 0.3 points from Q1, the first rise in eight quarters.
- 2026-11-05editedQ3 2026 release: GAAP gross margin 71.2% (was 71.8% in Q2), still falling
What we still do not know
- How much Workers and the AI agent tools earn; management calls Workers meaningful but will not give the figure.
- Whether websites will actually pay to charge agents, and at what price per request.
- Whether gross margin steadies near 72%, as management expects, or keeps falling as paid traffic grows.
The evidence under this - 3 claims
Asked directly for the size of the Workers business on the Q2 call, management declined to give it.
- The chief executive says the payment system must handle about 10 million transactions a second, far beyond any payment network today, which is unproven.
- The products are announced, not reported, so their uptake cannot be checked.
- The explanation is management's; AI inference and computing products may carry lower margins than security, and no product margins are disclosed.
- The adjusted gross margin rose 0.3 points from Q1, the first rise in eight quarters.
- 2026-11-05editedQ3 2026 release: GAAP gross margin 71.2% (was 71.8% in Q2), still falling
What would change our view
- Unattractive→NeutralThe price falls back within 20% of fair value
It needs:
- ○The share price falls to about $44.78, back within 20% of what we think the business is worth.about ≈$43.25 today, 88% below the last close
- Unattractive→AttractiveThe price falls 20% under fair value
It needs:
- ○The share price falls to about $29.86, 80% of what we think the business is worth.about ≈$27.80 today, 92% below the last close
- Unattractive→NeutralGrowth and cash that outrun the price
It needs all of:
- ○The first forecast for 2027 calls for revenue growth of 35% or more.checked at the next results, due in Feb 2027
- ○Spare cash reaches 20% of revenue over four quarters.checked at the next results, due in Feb 2027
We would also be wrong if
- Existing customers spend less than 112% of what they spent a year earlier, in any quarter of 2027.
- Customers paying over $100,000 a year grow less than 18% in a year.
- Gross margin falls below 70% for two quarters running.
What is coming
- February 2027 Q4 2026 results and the first 2027 forecast: Whether the 2027 forecast calls for growth near 28%, as analysts expect, or 35% and more, and whether net retention holds at 120% or above.
Every dated event ahead
- February 2027 Q4 2026 results and the first 2027 forecast: Whether the 2027 forecast calls for growth near 28%, as analysts expect, or 35% and more, and whether net retention holds at 120% or above.
- May 2027 Q1 2027 results: Whether gross margin has steadied above 71% and stock pay has fallen after the job cut.
Since we wrote this
- 11 NovThesis reviewUnattractive, high conviction: the view holds even in our best caseConviction re-graded on the new anchors: how robust the view is, not whether anything is still open.
- 9 NovThesis reviewStill Unattractive: priced at 513% of a value near $52.40Re-valued in three stages at an 8% return, the method now drawn on every thesis: the price sits beyond the 20% margin of safety.
- 6 NovEarningsCloudflare's call: growth was broad, profits turn positive, and the price still rules. View heldQ3 2026 results.
- 5 NovEarningsGrowth sped up to 39% and the year's forecast roseRevenue beat the top of the guide by $45M and net retention reached 122%. The price was still 34.6 times sales against a 25 times route, so the view held.
- 11 OctThesis reviewThe agent line lifted out and priced as a betThe owner's rule of 2026-10-11: a line a Futures theory names a company for, with no revenue shown, is early and owes a bet that cannot be declined.
- 10 OctThesis writtenUnattractive, medium convictionWritten nineteen days before Q3 2026 results, after the shares doubled from April on two forecast raises and the agent story.
Also checked 2 times with nothing new to act on.
Sources 29
Sources - 29
Every claim above cites one of these, or states the reasoning that derived it. The quote is frozen at the date it was read: a link can rot, and an agent fetch returns a reading of a page rather than the page. Backs counts the whole claim ledger, so it can name a claim this page does not surface.
- S01primarycompany filing2026-02-26business description, network size, customers, concentration, competition, risk factors, net retention history, revenue by geography, stock paybacksC01C02C04C05C08C12C23C34C38C41C45
- S02primarycompany release2026-08-06Q2 results, GAAP to adjusted reconciliation, balance sheet, cash flow, free cash flow, Q3 and 2026 guidance
Total revenue of $2,864.0 to $2,870.0 million
backsC05C06C08C10C11C15 - S03primarycompany filing2026-08-06key metrics, net retention, large customers, remaining performance obligations, restructuring note, convertible notes, share count, revenue by geographybacksC03C04C07C09C12C13C14C15C16C23C45
- S04primarycompany release2026-02-102025 results, free cash flow, gross margins, the first 2026 guidance
Total revenue of $2,785.0 to $2,795.0 million
backsC06C10C11 - S05primarytranscript2026-08-06Cloudflare Q2 2026 earnings call transcriptlarge-customer revenue share, non-human traffic, developers, Workers, agent payment products, gross margin explanation, restructuring cash; questions from Barclays, RBC, Morgan Stanley, Goldman Sachs, Citi, William Blair, Stifel and Scotiabank
Revenue contribution from large customers was 73% of revenue during the quarter, up from 71% in the second quarter last year. ... For the first time in human history, in Q2, more than 50% of the traffic flowing across Cloudflare's network was not human. ... we expect network CapEx to be 14% to 15% of revenue for full year 2026.
backsC03C07C09C10C11C25C28C29C42 - S06primarycompany release2026-05-07Q1 results, the workforce plan and its first charge estimate, the May guidance
As part of the plan, we expect to reduce our current workforce by approximately 1,100 people.
backsC06C09 - S07primarycompany filing2026-08-13the 2031 notes' size, rate, conversion price and proceeds
equivalent to an initial conversion price of approximately $496.94 per share, which represents a conversion premium of approximately 60% to the last reported sale price of $310.59
backsC14 - S08primarycompany filing2026-06-09executive biographies, beneficial ownership and voting power, pay design, the Class C proposalbacksC35C36C37
- S09primaryindustry data2026-10-10W3Techs, Usage statistics of Cloudflare as reverse proxy service W3Techs (Q-Success)Cloudflare's share of websites, measured outside the company
Cloudflare is used by 85.5% of all the websites whose reverse proxy service we know. This is 26.8% of all websites . ... Usage of the Cloudflare CDN is now at 25%, up from 19.6% one year ago and 3.9% ten years ago.
backsC24 - S10primarygovernment data2025-04-10U.S. Census Bureau, Statistics of U.S. Businesses 2022, detailed enterprise employment size table U.S. Census Bureaufirms by size; 500 and over is the total less the under-500 line
The Number of Firms and Establishments, Employment, Annual Payroll, and Receipts by State, Industry, and Enterprise Employment Size: 2022 ... United States, Total, 01: Total, 6395635 firms, 135748407 employment ... 19: <500 employees, 6374594 firms, 62250556 employment
backsC20 - S11primarycompany filing2026-02-25a rival naming Cloudflare and the public cloudsbacksC32
- S12primarycompany filing2026-02-20Akamai's three offeringsbacksC30
- S13primarycompany filing2026-09-03Zscaler's zero trust productsbacksC33
- S14primarycompany release2025-10-14the payment networks' use of Web Bot Auth
Cloudflare and Visa have worked together to develop and introduce the Trusted Agent Protocol , a new protocol that enables payment networks to integrate Web Bot Auth ... Mastercard is incorporating Web Bot Auth into Mastercard Agent Pay
backsC27 - S15primaryindustry data2026-01-07Adobe, 'Holiday Shopping Season Drove a Record $257.8 Billion Online with Consumers Embracing Generative AI Tools' Adobe Digital Insightsgrowth of AI-referred retail traffic
This season, traffic to retail sites from generative AI tools (shoppers clicking on a link to a retail site) increased by 693.4% compared to the year prior. ... While the base of users remains modest, the uptick shows the value AI can deliver as a shopping assistant.
backsC26 - S16secondaryvendor data2026-10-10Research pack, frozen 2026-10-10 (FMP market data, FINRA short interest, filed cohort)daily closes, the price decomposition, analyst estimates and ratings, price-target drift, cohort multiples, short interestbacksC16C17C18C19C31
- S17secondarynews2026-10-10the Deno purchase and its undisclosed terms
The financial terms of the acquisition were not disclosed. Deno had raised a total of $26 million , including a Series A led by Sequoia.
backsC40 - S18primarycompany filing2026-02-11Agent Pay described from the payment network's sidebacksC27
- S19primarycompany filing2026-06-30votes for and against the Class C share proposalbacksC35
- S20primarycompany filing2026-10-10five years of filed revenue, costs, stock pay, capital spending and share countsbacksC05C13C38C42
- S21primarycompany filing2026-10-10Cloudflare insider Form 4 filings read by insider_fetch.py, 2025-10 to 2026-10760 open-market sales, 756 under 10b5-1 plans, no purchasesbacksC39
- S22primaryindustry data2025-07-30what a breach costs the buyer
The global average cost of a data breach fell to $4.44 million, the first decline in five years, while the average U.S. cost of a breach reached a record $10.22 million.
backsC43 - S23primarytranscript2026-08-06Cloudflare Q2 2026 earnings call transcript: the chief executive on agent micropaymentsthe chief executive's own sizing of paid agent requests, and the Visa comparison
we handle, let's say, about 0.5 billion requests per second through Cloudflare's network. We roughly estimate that somewhere between 1% and 10% of those, you could monetize through some sort of a micro transaction. Again, these will be tiny fractions of pennies.
backsC46C48 - S24primarycompany release2026-10-02the price Cloudflare charges developers per request served
Standard | 10 million included per month +$0.30 per additional million
backsC47C48 - S25primarycompany release2026-07-28the stage of the agent payment product
Pay per crawl is currently in closed beta. ... Cloudflare acts as the Merchant of Record for pay per crawl and also provides the underlying technical infrastructure.
backsC49 - S26primarycompany filing2026-11-05Q3 2026 revenue, margins, cash flow, cash and securities, Q4 and full-year 2026 guidancebacksC06C07C10C15
- S27primarycompany filing2026-11-05dollar-based net retention, customers over $100,000, remaining performance obligations, shares outstanding, stock paybacksC03C04C07C13C15
- S28primarycompany filing2026-11-09trailing revenue, operating profit, stock pay, tax, debt, cash and the diluted share countbacksC51
- S29secondaryvendor data2026-11-09the analysts' average revenue and operating profit for the next three fiscal years, and the closebacksC50C51