Twilio Isn't a Cheaper Palantir. It's a Re-Accelerating Toll Road.

On the evening of August 6, Twilio jumped more than 17% in after-hours trading. The trigger wasn't a giant government contract or a headline AI partnership. It was usage: a voice product Twilio launched in May was pulling customers onto the platform, and the quarterly numbers beat what Wall Street had modeled.

The stock is up roughly 60% this year, while the AI name it keeps getting compared to, Palantir, is roughly flat. And since Palantir trades near 66x sales while Twilio sits around 6x, the pitch writes itself: the cheaper AI stock that's crushing Palantir. Here's the problem with that pitch, and what it hides.

The re-acceleration is real

Start with the numbers, because this part is not in dispute. Twilio reported second-quarter revenue of $1.50 billion, up 22% reported and 17% on an organic basis — that is, excluding the revenue added by an acquisition — versus a Wall Street estimate of about $1.43 billion. Non-GAAP earnings of $1.47 beat the $1.32 consensus.

The more telling figure is retention. Twilio's dollar-based net expansion rate — how much existing customers spend on top of last year — reached 116% in the quarter, up from 108% a year earlier. A net expansion above 100% means the base is growing on its own, without new logos. Combined with revenue growth that had been stuck in single digits a year or two ago, that re-acceleration is what broke the old "value trap" story.

The engine is AI agents, and the evidence is usage rather than aspiration. Voice revenue grew more than 20%, its best pace since a 19-quarter drought, and self-serve voice — small businesses running their own AI agent — surged over 50%. Software add-ons, the higher-margin intelligence layer sitting on top of the raw connectivity, grew more than 25% after a quarter of triple-digit growth. Management says AI-native customers are consolidating traffic onto Twilio and posting net expansion above the company average.

That's the observable operating evidence. Twilio also raised its 2026 free-cash-flow guidance to $1.135–1.155 billion, which works out to a roughly 17% free-cash-flow margin on about $5.6 billion of trailing revenue.

Why "cheaper than Palantir" is the wrong lens

Here is where the marketing breaks down, and it matters for how you think about the stock. Twilio is a communications-infrastructure business. It moves text messages, phone calls, and email for other companies, charging per message and per minute — a toll road. Because it passes through carrier and telecommunications fees, its gross margin is only about 49%, nowhere near the near-80% gross margins of a pure analytics franchise like Palantir.

So the 6x-sales multiple is not Palantir's 66x-sales multiple at a discount. It's the appropriate price for a lower-margin, infrastructure business growing in the high teens. Comparing the two multiples is like comparing a toll road to a software monopoly: the cheapness is real, but the reason for it is the ~50% gross margin and the commodity nature of the transport layer, not that Wall Street has failed to notice Twilio is secretly special.

There's also a quieter detail worth flagging. Reported growth was 22%, organic growth 17% — a five-point gap that comes from the November acquisition of Stytch, an identity platform for AI agents. Organic growth is the honest engine, and the current price is set against it, so keep that 17%, not the headline 22%, in mind.

What the current price actually pays for

After the run, this is no longer the deep-value setup that existed when Twilio traded near $100. Today the stock is around $227, off about 12% from a 52-week high near $258, valuing the company at roughly $35 billion. That works out to about 6x trailing sales and about 30x the $1.14 billion of free cash flow the company itself guides to for the full year — against roughly $1.6 billion of net cash on the balance sheet.

Is that cheap? Not the way it was. Is it demanding? Not the way Palantir is. A net-cash business growing around 18%, printing a 17% free-cash-flow margin, is defensible at 30x free cash flow. The forward math doesn't need an AI-multiple premium to work — which is exactly why the "cheaper Palantir" frame is the wrong reason to own it. If you value it like Palantir, you'll be disappointed. If you value it like the cash-generative toll road it is, the current price is reasonable.

The one variable that carries the case

The re-rating already happened. What decides whether it holds is whether that ~17% organic growth is durable. Right now the voice-AI traction is concentrated in small-business self-service and contact-center co-pilots, the easy wins. Regulated industries — financial services and healthcare — are still in "heavy experimentation", which is either the next leg of growth or the stall depending on how you read it.

Two structural risks sit under the surface. Twilio doesn't disclose AI-specific revenue, so you're judging by usage proxies rather than a reported line. And because carrier fees pass through its ~50% gross margin, telecom pricing can hit profitability on short notice — the built-in fragility of a toll road. The durable part is the re-acceleration and the net-cash balance sheet; the uncertain part is where the growth rate settles.

So don't buy "Palantir, but cheap" — that's a category error with a built-in disappointment. Buy the case, if you buy it at all, as a re-accelerating, net-cash infrastructure business at ~30x the free cash flow it says it will print in 2026, and treat the durability of organic growth as the single thing that decides whether the story holds. The narrative has already been paid for; the math has to carry it from here.