Twilio rallies 30% on an AI story, but four points of its 20% growth are pass-through and acquisitions

---

On April 30, 2026, Twilio reported a quarter that almost everyone agreed was good. Revenue of $1.407 billion grew 20% year over year, the highest reported growth rate the company had printed in more than three years. Earnings per share of $1.50 on a non-GAAP basis blew past the $1.27 consensus. Management raised full-year guidance. And the market did what markets do with a turnaround narrative wrapped in the letters A and I: it bought first and parsed later, sending the shares up roughly 30% in the immediate aftermath and leaving the stock up about 53% over the trailing year. The story sold itself — the commodity-messaging pipe that growth investors had left for dead in 2023 had reinvented itself as an AI-native customer-engagement platform.

The thesis here is not that Twilio is a fraud, or that the quarter was bad, or that the business is shrinking. None of those things is true. The thesis is narrower and, for anyone paying 33 times forward earnings, more uncomfortable: the gap between the number the company markets and the number the company keeps has widened at exactly the moment the multiple expanded to price the headline. Twilio is being valued as a re-accelerating platform. A material slice of the re-acceleration is pass-through fees it remits to carriers and revenue it bought. The platform story may yet be right. But the price now requires it to be right, and the evidence for it is thinner than the rally implies.

The four points that aren't really yours

Start with the single most important sentence in the entire earnings release, and the one least likely to survive into a CNBC chyron: reported revenue grew 20%, but organic revenue grew 16%. That four-point wedge is not a rounding artifact. It is, by Twilio's own explanation, the combined contribution of acquisitions and incremental application-to-person — A2P — carrier fees.

A2P fees deserve a moment, because they are the cleanest example of the denominator illusion in modern software accounting. When a business sends an SMS through Twilio to a consumer in the United States, the wireless carriers — the AT&Ts and Verizons of the world — levy a per-message fee. Twilio collects that fee from its customer and remits it onward. Under gross revenue recognition, the fee inflates the top line on its way through, even though Twilio keeps essentially none of it. When carriers raise those fees — and they have, repeatedly — Twilio's reported revenue rises with zero corresponding improvement in the underlying business. Management even disclosed the other side of this coin: non-GAAP operating margin absorbed roughly a 70 basis point headwind from incremental U.S. carrier fees in the quarter. Read that twice. The same fees that pad the revenue line are a drag on the margin line, because Twilio earns nothing on them but reports them as if it had.

So when an investor sees "20% growth" and mentally pays for 20% growth, the investor is paying, in part, for AT&T's pricing decisions and for businesses Twilio acquired with shareholder money. The organic, keep-the-cash number is 16%. That is still a genuine acceleration from the 12% organic growth Twilio posted in the first quarter of 2025 — a real improvement, and one the bears must concede. But it is not 20%, and the four-point difference is precisely the kind of low-quality growth that compounds into a disappointing multiple compression when it inevitably laps out of the comparison.

A multiple priced for the headline, not the organic

At a roughly $21 billion market capitalization and a forward price-to-earnings ratio in the low-to-mid thirties — independent trackers put it around 33 — Twilio now trades like a durable mid-teens grower with expanding margins and optionality on artificial intelligence. That is a defensible valuation for a 20% grower. It is a demanding valuation for a 16% organic grower whose own full-year guidance calls for organic growth of just 9.5% to 10.5%.

Hold those two numbers next to each other, because management put them in the same press release. First-quarter organic growth: 16%. Full-year 2026 organic guidance: 9.5% to 10.5%. The company's own forecast, in other words, embeds a sharp deceleration in the organic business over the back half of the year — roughly a halving of the organic growth rate from the quarter that just triggered a 30% rally. Reported full-year guidance of 14% to 15% papers over this with the same acquisition and fee tailwinds that inflated the quarter. The market re-rated on the 20% and the 16%. The company is guiding to the 10%. Priced-for-perfection asymmetry does not get much more legible than a stock that jumps 30% on a quarter while the issuer quietly forecasts the underlying growth engine to slow by nearly half.

Cyclical demand wearing a secular costume

The most seductive line in the quarter was the attribution of growth to AI. Voice revenue up 20%, self-service voice up 45%, messaging up 25%, "driven by AI use cases." It is a wonderful narrative because it converts a usage-based, consumption-priced business — historically valued like a utility because customers pay per message and per minute — into a platform business that deserves a software multiple.

But consumption revenue is, by construction, cyclical, not secular. Twilio gets paid when its customers send more messages and place more calls, which happens when its customers are growing, marketing aggressively, and flush with capital. A surge in messaging and voice volume in a year when venture-funded AI startups are spending freely on customer outreach, and when every consumer app is bolting on an "AI agent" that texts and calls users, is precisely the kind of demand that looks secular on the way up and reveals itself as cyclical on the way down. The 45% self-service voice growth is impressive. It is also exactly what you would expect from a wave of new AI-application customers in their land-and-spend honeymoon — the period before churn, before optimization, before the next funding winter teaches them to throttle their messaging spend. Pricing a consumption business as a platform at the top of a spending cycle is one of the oldest ways to overpay in software.

The metric that disappeared at the convenient moment

Here is the forensic detail that should make any careful reader sit up. Beginning in the first quarter of 2026 — this exact quarter, the one of the great re-rating — Twilio discontinued disclosure of Active Customer Accounts as a key metric. For years, Active Customer Accounts was one of the two headline operating metrics by which the entire investment community judged Twilio's health, alongside the dollar-based net expansion rate. It is gone.

The company noted, almost in passing, that the quarter still reflected roughly 43,000 net new Active Customer Accounts. That sounds fine. But the decision to retire a long-standing disclosure is never neutral, and the timing — the precise quarter in which the narrative pivots from "customers" to "platform" and "AI" — invites the obvious question. A management team confident that its customer-account trajectory supports the AI-platform story has every incentive to keep showing the number. A management team repositioning toward usage-based AI revenue concentrated in a smaller cohort of high-spend accounts has every incentive to stop. Investors are now asked to take the breadth of the franchise on faith, at the same moment they are asked to pay a platform multiple for it. When a company removes a disclosure precisely as it asks for a higher valuation, the burden of proof should shift toward the skeptic, not away.

Dollar-based net expansion: better, and still below the old highs

The metric Twilio kept — dollar-based net expansion rate — did improve, to 114% from 107% a year earlier. That is a real positive and another concession the bears must make: existing customers are spending more, and a 114% expansion rate is healthy. But context matters. In its 2021 peak, Twilio routinely posted net expansion rates in the 125% to 130%-plus range. A 114% reading is a recovery off a depressed base, not a return to the hypergrowth franchise the bulls remember. And a portion of that very expansion is, once again, the A2P fee pass-through and price increases flowing through existing accounts — expansion that reflects carriers charging more, not customers deriving more value. The expansion rate is genuinely better. It is not evidence of a platform moat; it is evidence of a consumption business in an up-cycle with some carrier-fee tailwind baked in.

GAAP profit is real — and still leans on the adjustment

Credit where due, and this is a large credit: Twilio printed real GAAP profit. GAAP income from operations of $108 million and GAAP net income of $90 million, or $0.57 per diluted share, is a genuine milestone for a company that spent most of its public life drowning in losses. This is not a company that only profits in adjusted fantasy land. The GAAP number is positive, and that materially changes the risk profile versus the cash-burning Twilio of 2022.

But examine the wedge that remains. Non-GAAP income from operations was a record $279 million at a 19.8% margin. GAAP income from operations was $108 million. The roughly $171 million difference is dominated by stock-based compensation. Here, too, Twilio has improved markedly — SBC fell to 9.7% of revenue, below 10% for the first time since its IPO, ahead of the company's own 2027 target. That is real progress and deserves acknowledgment. Yet 9.7% of $1.41 billion is still roughly $137 million of compensation paid in stock every quarter — compensation that is a genuine economic cost to shareholders through dilution, that the non-GAAP operating margin of 19.8% pretends away, and that the buyback exists in part to mop up. The adjusted margin the bulls cite is built by excluding a cost the company is still paying. Smaller than before, yes. Zero, no.

The buyback that runs to stand still

Twilio is repurchasing stock aggressively — $253.4 million in the first quarter, roughly $1.1 billion completed of a $2 billion authorization, with about $892 million remaining as of the end of March. Buybacks are usually shareholder-friendly, and free cash flow of $132 million in the quarter and a $1.08–1.10 billion full-year free-cash-flow guide give the company the means.

But look at what the buyback is partly doing. With roughly $137 million of stock-based compensation issued per quarter, a meaningful share of every repurchase dollar is not reducing the share count so much as offsetting the dilution the company creates by paying its workforce in equity. A buyback that runs to stand still against SBC is not the same as a buyback that compounds per-share value, even though both show up under the same friendly headline. The cash is real. The accretion is less than the gross repurchase figure implies, because a chunk of it is simply recycling the equity that walked out the door as compensation.

The commodity floor hasn't moved

Underneath the AI repositioning sits the same business that got Twilio de-rated in the first place: sending text messages and connecting phone calls over carrier networks Twilio does not own. That is a price-taker's business. The carriers set the A2P fees. Competitors — Bandwidth, Sinch, Vonage's API arm, the messaging units inside every cloud platform — sell a substantially interchangeable pipe. The reason Twilio's gross margins are structurally lower than a true software company's is that a large portion of revenue is carrier cost passed through with a thin spread. No amount of AI-agent marketing changes the physics of who owns the network and who sets the toll. The AI layer Twilio is building may add genuine value on top of the pipe. But the pipe is still a commodity, the floor is still a price-taker's floor, and a platform multiple sits uneasily on top of an infrastructure business whose largest input cost is dictated by the very carriers whose fees inflate its revenue.

The guidance ladder and the comparison cliff

There is a structural reason the back half of 2026 should be watched more closely than the front, and it is buried in the cadence of Twilio's own guidance. The second-quarter outlook calls for reported revenue of $1.42 to $1.43 billion, growth of 15.5% to 16.5% — already a step down from the first quarter's 20% reported figure. By the time the full year is in the books, reported growth is guided to 14% to 15%. Each rung down that ladder is the acquisition and A2P-fee tailwind thinning out as it laps the prior-year periods when those same boosts first appeared. A growth rate that is being held up by year-over-year comparisons against pre-acquisition, pre-fee-increase quarters faces a cliff the moment those easy comparisons roll off. The market extrapolated the 20%. The company's own arithmetic says the 20% was the peak of the print, not the run rate.

What makes this more than a modeling footnote is the leverage in the multiple. At roughly 33 times forward earnings, a software stock needs the growth narrative to stay intact to hold its rating; multiples in the thirties do not survive a slide toward low-double-digit reported growth and high-single-digit organic growth. The asymmetry is unkind. If Twilio merely hits its own guidance — 14% to 15% reported, 9.5% to 10.5% organic — the growth profile that justifies the current multiple erodes on schedule. To keep the rating, Twilio does not need to meet guidance; it needs to blow through it, repeatedly, on organic strength the AI story has promised but not yet durably delivered.

Repositioning is not the same as re-platforming

Management's strategic language is worth a forensic read on its own. The company describes rearchitecting its back end so that data and memory flow across communication channels, repositioning data "as central to the platform strategy rather than as a standalone focus." That is corporate prose for a hard truth: Segment, the customer-data platform Twilio bought for over $3 billion in 2020 and once positioned as the brain of the operation, did not become the growth engine the acquisition promised, and the company has been retooling around that disappointment. Repositioning an asset is what you do when the original thesis for owning it did not pan out. The AI-platform narrative is, in part, the second act of a customer-data story whose first act underwhelmed. Investors paying a platform multiple today should remember that the last time Twilio sold the market a data-platform vision at a premium, the premium did not age well — and the lesson of that episode is that integration narratives are cheap to tell and expensive to deliver.

What the bulls genuinely get right

A fair forensic case has to concede the strong parts of the bull thesis, and here they are real and specific. First, the growth acceleration is genuine: 16% organic, up from 12% a year earlier, is a true inflection, not an accounting trick, and acceleration in a business everyone had written off as ex-growth is exactly the kind of surprise that deserves a re-rating. Second, the profitability turn is authentic and GAAP-confirmed — $108 million of GAAP operating income and $90 million of GAAP net income is not adjusted vapor; it is real money, and it transforms the downside risk relative to the loss-making Twilio of three years ago. Third, the SBC discipline is exceptional: pulling stock-comp below 10% of revenue ahead of schedule is precisely the cultural change skeptics demanded, and it is happening. Fourth, the dollar-based net expansion recovery to 114% shows existing customers genuinely leaning in. Fifth, the AI use cases are not purely rhetorical — self-service voice up 45% and multi-product adoption up 29% suggest real product pull, and Twilio's installed base of developers and its data assets give it a credible right to win as conversational AI agents proliferate. A consumption business positioned underneath every AI agent that needs to text or call a human is a genuinely interesting place to sit. If the AI-agent thesis is right, Twilio is a picks-and-shovels beneficiary with an existing customer base most rivals would envy. The bull case is not stupid. It is simply priced as a certainty when it is still a probability.

The kicker

What the rally bought was a number — 20% — and what the company keeps is a smaller number — 16% organic, decelerating to 10% on its own guidance — sitting on top of an infrastructure business whose revenue is partly the carriers' fees and whose adjusted margin is partly the stock it still hands to employees. None of that makes Twilio a short at any price; the GAAP profit is real, the organic inflection is real, and the AI optionality is real. It makes Twilio a stock where the gap between the marketed growth and the retained growth has widened at exactly the moment the multiple expanded to pay for the marketed version, and where a long-trusted disclosure vanished in the same quarter the valuation asked for a leap of faith. The platform may arrive. The question the 30% pop never paused to ask is how much of the pipe you are paying for at platform prices.

The cleanest tell in the whole quarter is that Twilio raised reported guidance to 14–15% while guiding organic to 9.5–10.5%, and the market chose to celebrate the larger number it does not get to keep.