Open in interactive viewer → charts, metric popovers & call review

DLocal Prints a Record Quarter, Then Hands Back Six Points of It to FX

77% TPV growth, a record $0.23 EPS and the best operating margin since 2023 — undercut by a Q4 currency swing the CFO said in August wouldn't come, and a Mexico fix that keeps slipping
DLO · Earnings Call · 2026-11-12

DLocal Prints a Record Quarter, Then Hands Back Six Points of It to FX

On the numbers alone, this was the quarter DLocal bulls have been waiting two years for. Total payment volume hit $18.3 billion, up 77% year-over-year — the “eighth consecutive quarter above 50% growth and the fourth one at or above 70%” — Pedro Arnt, Chief Executive Officer · 2026-11-12, as Pedro Arnt put it. Gross profit rose 28% to $132 million. Operating profit rose 37% to $76 million, and at 58% of gross profit it marked the best conversion since the fourth quarter of 2023. Net income of $68 million and diluted EPS of $0.23 were both company records. Headcount fell 3% sequentially, and gross profit per employee crossed $100,000 for the first time.

Then management spent the back half of the call telling investors not to annualize any of it.

The Guide Is a Cut Wearing a Reiteration's Clothes

The headline was "we reiterate our guidance ranges." The detail was messier. TPV is tracking toward the upper end of the 60%–70% range — fine. But full-year gross profit growth is now expected "around the midpoint" of 25%–30%, and operating profit growth has moderated toward the lower end of 27.5%–32.5%. The culprit is explicit: “previously unanticipated FX headwinds in Brazil, Mexico and Argentina during Q4, which we estimate as of today as having six percentage points of swing” — Pedro Arnt, Chief Executive Officer · 2026-11-12 on year-over-year gross profit growth. Add softer trading in the smaller, high-take-rate markets of Africa and Asia.

At the same time, we now expect previously unanticipated FX headwinds in Brazil, Mexico and Argentina during Q4, which we estimate as of today as having six percentage points of swing on our year-on-year gross profit growth for the quarter.

Pedro Arnt, Chief Executive Officer · 2026-11-12

Here is the tension. In August, Arnt was asked whether the currency volatility that dogged the first half would persist. His answer: “unless something else happens in terms of quick dislocations of currency values, I don't think you'll have this level of volatility in future quarters” — Pedro Arnt, Chief Executive Officer · 2026-08-13. Something else happened. That is not a crime — DLocal operates in the most FX-fragile corners of the world and says so every quarter — but it does mean a record print lands alongside a de facto trim, and the market has to decide which one to weight. The stock, at roughly $14.95 in mid-September and up only about 10% over the prior year against 70%-plus volume growth, suggests the market has already been discounting the growth story heavily.

Mexico: The Third Verse of the Same Song

Mexico is now a recurring line item rather than an event. In August, Arnt described the problem as a cost issue — pricing power holding, processing costs outrunning scale — and said “there's work to be done in Mexico, but it's more cost management, which I feel relatively confident we will deliver on” — Pedro Arnt, Chief Executive Officer · 2026-08-13. Three months later the diagnosis is identical and the tone is different: “I wish it would have happened sooner. We're very focused on this, but this is not something that we think has to take multiple quarters” — Pedro Arnt, Chief Executive Officer · 2026-11-12. The plan has two prongs — leverage record TPV into a lower cost base, and climb out of the lowest pricing tiers that a handful of concentrated large merchants have pushed them into. The resolution date is now "next year or maybe even as we exit this year." A fix that has been one quarter away for two consecutive quarters is worth flagging, even if the mechanism is credible.

Asia Stops Being a Slide and Becomes a Number

The genuinely new thing on this call was geography. Asia has been an aspiration for several quarters — in May, the segment's strength was still driven by Africa, with Asia in "initial steps." This quarter Arnt is flying to the region the day after the call, a local CEO has been hired, and management finally put a figure on it: a mid- to high single-digit contribution to top-line revenue in year one, i.e. 2027. The ambiguity was pronounced enough that a Truist analyst asked him to re-clarify, and he trimmed his own framing to “mid-single digits to potentially high single digits depending on success” — Pedro Arnt, Chief Executive Officer · 2026-11-12. Anchor merchants are named in type if not in name — one of the world's largest music streaming companies is already processing, and there are conversations with an AI lab.

Note the shape of the language: Asia Pacific has entered the company's own keyword register for the first time this quarter, at the same moment as the broader growth vectors framing that Arnt uses to organize existing merchants, new merchants, new products and new geographies into one investment case. The four-vector language is more than a structural detail. It is the argument for why operating leverage gets reinvested rather than paid out.

What Management Stopped Saying — and What It Started

The most telling change is an absence. For over a year the take-rate debate dominated this call: net take rate was among the highest-momentum themes in the August 2026 quarter and ranked first for the company as far back as 2024. On the Q3 2026 call the topic barely surfaces, and the keyword has fallen out of the top of the register entirely. In its place: marketing spend timing, cost discipline, the credit facility signed in August, and the buyback program that repurchased roughly $34 million in the quarter and about $120 million year-to-date. Arnt's concession on take rate was almost throwaway — declines are slowing sequentially, but he "wouldn't say we've hit the bottom."

That pivot is deliberate, and the CFO policed it hard. Guillermo Perez flagged that the 11% sequential OpEx decline was part structural (automation, lower third-party spend, the first sequential drop in payroll this year) and part timing (marketing was front-loaded). His warning: “I wouldn't think about this quarter's result as a margin target, and I would be careful about extrapolating from a single quarter to future quarters” — Guillermo Perez, Chief Financial Officer · 2026-11-12. Compare that to Jeff Brown on the August 2025 call, when the message was the opposite: “we continue to expect to increase OpEx over the next few quarters” — Jeffrey K. Brown, Chief Financial Officer · 2025-08-13. The investment cycle genuinely has rolled off. What has not been proven is that 58% is a floor rathe than a peak.

Two emerging themes got very different treatment. Agentic protocol commerce arrived in the company's vocabulary this quarter and Arnt was scrupulously honest about it — most agent traffic arrives disguised as a human, protocol-driven volume is "still quite low," and the whole thing is "more our expectation than anything we're actually seeing in the numbers." That is a rare and useful piece of candor, and it lines up with the market-wide personal AI agents narrative running through this reporting season without yet showing up in payment volumes anywhere.

Stablecoin is the opposite case, and represents a real upgrade in disclosure. In March, Arnt said flatly: “we are not seeing significant volumes or pick up in stablecoin at checkout” — Pedro Arnt, Chief Executive Officer · 2026-03-18. On Thursday: “The digital asset and stablecoin vertical is one that's actually delivering quite solid growth” — Pedro Arnt, Chief Executive Officer · 2026-11-12 — with stablecoins appearing in remittance and repatriation flows, and a growing roster of merchants accepting them at checkout. DLocal positions itself on the fiat-to-stable on-ramps and liquidity where the FX margin sits, which is precisely where stablecoin adoption actually monetizes rather than where the narrative lives. That is the more interesting thread to watch than either Mexico or agentic commerce.

The Verdict

DLocal is a $4 billion company compounding volume at 77% while its stock does roughly nothing, and this quarter explains part of why: the operating model is finally deleveraging its cost base, but the revenue line is hostage to three currencies and one stubborn market. The existing merchant base — 173% TPV retention, 140% net revenue retention — is the engine and it is not slowing. The question for the next two quarters is narrower and duller than the growth story: does 58% operating margin on gross profit hold through a Q4 with six points of FX drag, and does Mexico finally stop being a topic?