Everything About AI
NextFin News - Cloudflare raised its full-year adjusted-profit forecast to $1.35 a share, the third raise this year, as artificial-intelligence traffic pushed third-quarter revenue up 39% to $782.5 million and flipped the composition of the internet itself. The company now expects non-GAAP net income per share of $1.35 for fiscal 2026, up from a previous range of $1.25 to $1.26 and above the $1.27 Wall Street had penciled in. The lift comes as machine-generated requests overtook human browsing on Cloudflare's network for the first time in the second quarter - a milestone the chief executive says could widen to non-human traffic as much as 1,000 times human traffic within five years if current trends hold. The question for investors is no longer whether AI is boosting demand for the network. It is whether that demand converts into durable profit faster than the AI spending cycle cools.
The Quarter: Growth Accelerates, Guidance Goes Up - Again
Cloudflare's third quarter, ended September 30, 2026, was its strongest on record by the company's own measure.
"The third quarter was Cloudflare's best quarter ever," said Matthew Prince, co-founder and chief executive officer. "We crossed $3 billion in annualized revenue with revenue of $782 million, up 39% year over year. We accelerated growth, delivering our fastest revenue growth in nearly four years, while expanding profitability."
Non-GAAP net income came in at $143.7 million, or $0.38 per diluted share, against $102.6 million and $0.27 in the year-ago period. On a GAAP basis the company still lost money - a net loss of $34.1 million, or $0.10 per share - but operating discipline improved: non-GAAP income from operations was $137.1 million, a 17.5% margin, up from 15.3% a year earlier. Free cash flow was $82.1 million, or 10% of revenue, and cash, cash equivalents and available-for-sale securities totaled $5,128.7 million at quarter-end.
The guidance raise is the story within the story. Cloudflare has now lifted its profit outlook three times in 2026. The new full-year adjusted-earnings guide of $1.35 a share sits above the midpoint of the prior $1.25-to-$1.26 range and clears the $1.27 consensus. For the fourth quarter, management expects revenue of $845.0 million to $846.0 million and non-GAAP earnings per share of $0.43. Full-year revenue is now projected at $2,963.0 million to $2,964.0 million, implying roughly 32% growth for the fiscal year. The leading indicator behind that guide - current remaining performance obligations, a measure of contracted future revenue - grew 37% year over year.
The market's first read was mixed. Cloudflare shares fell 2.79% to $244.51 during the regular session on November 5, part of a broader pullback across AI-infrastructure names amid election-driven concerns over data-center scrutiny. After the results landed, the stock reversed in extended trading, reaching $282.99, or 6.16% above the prior close of $266.58. The intraday split - red during the day, green after the numbers - captures the tension investors are wrestling with: strong execution, but a valuation that already assumes AI keeps compounding.
Why AI Traffic Hits Cloudflare's Revenue Line
The mechanism is straightforward once you trace how an AI answer engine consumes the web. A human shopper might visit five websites before buying a product. An AI agent researching the same purchase can make thousands of requests across those same sites - fetching product pages, comparing prices, reading reviews, and checking inventory, all without a person clicking anything. Each of those requests, in Cloudflare's telling, has to pass through the same network edge that serves human traffic, and each can be sold the same suite of security, performance, and developer services.
That is why the traffic mix matters more than the traffic total. In the second quarter, for the first time, more than half of the requests flowing across Cloudflare's network were not human. Prince said on the earnings call that he had been "quite surprised" when his team told him in May that non-human traffic had passed human traffic online, and he offered a projection with a large caveat: if current trends continue, non-human traffic could reach as much as 1,000 times human traffic within five years. He has called AI "the biggest tailwind we've ever seen in Cloudflare's history."
The shift is not theoretical for Cloudflare's customers. As search moves from ten blue links to AI-generated answers, the websites that want to be cited - and that want to defend themselves against AI scrapers, credential-stuffing bots, and denial-of-service attacks riding the same wave - need more edge capacity, not less. Cloudflare sells that capacity. The result is visible in the quarter's leading indicator: current remaining performance obligations grew 37% year over year.
There is a second channel, less visible but arguably more durable: Cloudflare is eating its own dog food. In May 2026, when the company announced it would cut roughly 1,100 roles - about a fifth of its workforce - Prince wrote in a letter to employees that Cloudflare's usage of AI had increased by more than 600% in the three months prior, with staff across engineering, HR, finance and marketing running thousands of AI-driven workflows each day. That internal shift does not show up directly as revenue, but it matters for the profit guide: the same tools that lift Cloudflare's own productivity are the ones it is productizing and selling to enterprises that want to deploy agents without exposing their networks.
Cyclical Wave or Structural Shift? It Is Both - and That Is the Trap
Here is the call investors need to make, and it is easy to get wrong. Is the AI demand hitting Cloudflare a cyclical surge - a capital-spending boom that will mean-revert when AI labs slow their buildout - or a structural regime change in how the internet works?
The answer is both, and they point in opposite directions over different horizons. The cyclical leg is real and it is dangerous. The AI buildout is a capital-spending wave, and capital-spending waves crest. If the large AI labs trim inference spend, slow model training, or consolidate their vendor stack, the growth rate of requests on Cloudflare's network can decelerate quickly. Cloudflare's own gross margin already shows the cost of the wave: GAAP gross margin was 71.2% in the third quarter, down from 74.0% a year earlier, as the traffic mix shifted toward compute-heavy AI workloads that are more expensive to serve.
But the structural leg is the stronger one, and it is why the profit guide keeps going up. A cyclical fluctuation changes how much traffic flows. A structural shift changes what the traffic is and who pays for it. The internet is being rewritten for machine-to-machine communication - answer engines instead of browsers, agents instead of sessions, API calls instead of page views. That rewrite does not reverse when capex slows, because the applications being built on top of it - autonomous research, agent-driven commerce, enterprise workflows - do not go back to being ten-blue-link search. Cloudflare sits at the junction where that traffic enters and leaves every enterprise network, which is a structural position, not a cyclical one.
The evidence floor for the structural call is met: this is a technology-and-architecture regime change - the protocol-level shift from human browsing to agent traffic - the historical baseline no longer applies, because non-human traffic has never been the majority before, and the driver does not self-correct, because an agent that replaces a browser session does not turn back into one. The evidence floor for the cyclical warning is also met: there is a short-term capital-spending driver, and margin compression is already showing up in the reported numbers.
So the investment question is not "AI or not AI." It is a timing and mix question: does the structural growth of machine traffic outrun the cyclical risk of an AI capex slowdown, and does pricing power keep ahead of the rising cost of serving that traffic?
The Second-Order Trade Everyone Is Missing
The first-order read of this quarter is easy: AI traffic is up, revenue is up, guidance is up. The market has priced that. What the market has not fully priced is the second-order consequence of non-human traffic becoming the majority: Cloudflare's customer base is about to bifurcate, and the company becomes both a beneficiary and a gatekeeper in the same transaction.
Think through the chain. Event: non-human traffic crosses 50% of the network. First-order effect: more requests, more revenue, higher growth - this is the consensus trade. Second-order effect: the entities generating that traffic split into two camps with opposite economics. On one side are the AI labs and answer engines, which want to crawl the web as cheaply as possible and will resist paying for access. On the other side are the publishers, merchants, and enterprises whose content is being consumed, and who will pay Cloudflare to control, meter, secure, and monetize that consumption. Cloudflare's revenue exposure is weighted toward the second group - the defenders, not the crawlers - which is why its growth accelerates even as the crawlers complain about cost.
Third-order effect: this puts Cloudflare in the position of infrastructure for the AI economy's settlement layer. If agents begin completing transactions - not just browsing but buying - the need for identity, fraud detection, bot differentiation, and rate limiting at the edge becomes non-optional. That is a larger market than the one priced into a 39% revenue-growth quarter. It is also a market that hyperscale cloud providers are already competing for, which is why the valuation premium cannot be taken for granted.
The Bear Case, Stated at Full Strength
The strongest argument against the bull case does not attack the growth - the growth is real. It attacks the durability of the profit and the price being paid for it. There are three pillars.
First, margin compression. GAAP gross margin fell 280 basis points year over year to 71.2%, and free cash flow as a percentage of revenue fell to 10% from 13%. AI traffic is heavier and more expensive to serve than human browsing. If the mix keeps shifting toward AI without commensurate pricing, the profit guide rises on revenue scale while the margin structure quietly deteriorates. A company can grow into a profit target and still destroy value if the capital intensity of that growth keeps climbing.
Second, valuation. Cloudflare trades at a premium multiple that embeds years of uninterrupted AI-driven compounding. Any deceleration in the growth rate - even to a still-strong 25% or 30% - would compress that multiple faster than earnings could catch up. The stock's intraday reversal on November 5, falling in the regular session before rallying after-hours, is exactly what a crowded, high-expectations trade looks like when the bar is set high.
Third, competition and concentration. The same hyperscalers that sell GPU capacity to AI labs are building edge and security offerings of their own. If AI spending slows and customers consolidate vendors, Cloudflare could face price pressure from providers that bundle edge services with compute. And while Cloudflare serves a broad customer base, a meaningful share of the AI traffic growth traces to a small number of large AI companies - concentration that the top-line diversification does not fully neutralize.
The bull answer to all three is that Cloudflare's network scale and developer platform create switching costs that pricing pressure cannot easily penetrate, and that the company has raised its profit guide three times in a row while margins on a non-GAAP operating basis actually expanded to 17.5%. That is a fair answer. It is not a complete one.
What Would Prove the Bull Case Wrong
The falsifying signal is specific and observable. If Cloudflare's non-human traffic growth decelerates for two consecutive quarters while GAAP gross margin compresses by more than 200 basis points year over year, the structural-shift thesis is damaged - it would mean AI demand is cyclical after all, and that the cost of serving it is structural. A second signal: if current remaining performance obligations growth - 37% this quarter - falls below 20% while revenue growth stays above 30%, it would indicate the company is pulling revenue forward rather than building a durable backlog.
What Comes Next
The near-term path runs through the fourth-quarter guide. Revenue of $845 million to $846 million implies continued acceleration, and non-GAAP EPS of $0.43 implies the profit raise has legs. The base case is that Cloudflare delivers the quarter and the stock grinds higher as the AI-infrastructure selloff fades. The upside case is that agent-driven commerce shows up in the metrics - higher monetization per non-human request, expanding large-customer counts, and gross margin stabilizing above 72% - which would re-rate the stock toward its growth ceiling. The downside case is an AI capex slowdown that shows up first in traffic growth and then in the RPO line, which would hit both the multiple and the guide.
By time horizon: in the short term, sentiment around AI capex and the election-driven regulatory overhang will dominate the tape. In the medium term, the next two quarters of gross margin and RPO growth will determine whether the profit raises are sustainable. In the long term, the question is whether machine-to-machine traffic becomes the internet's default mode - and whether Cloudflare remains the gate through which that traffic is secured and monetized.
Cloudflare's quarter proves AI is reshaping the internet's plumbing. The next act will prove whether the company that owns the plumbing can charge tolls faster than the cost of maintaining it rises.
Explore more exclusive insights at nextfin.ai.
