Posts  / #POST-244503
REDDIT

DLO (dLocal): how real is the in-sourcing risk?

P
Aug 3, 2026 · 13:26

Long-time holder trying to stress-test the main bear point. dLocal is a high-quality, asset-light EM payments business — one API for global merchants across \~40 emerging markets. The moat is licensing/compliance breadth, and it looks durable: even Stripe didn’t build Brazilian Pix itself, it partnered with EBANX (a dLocal competitor) in 2025.

My worry is concentration + in-sourcing. Revenue sits on \~760 merchants, top 10 ≈ 61% (and rising), with one or two individually >10%. The risk: once a merchant’s volume in one country gets big enough (\~$200–500M TPV), it can get its own local license and process directly, cutting dLocal out of that market.

It’s not hypothetical — Q3 2024 a top merchant got a Brazil license, routed cards direct, and Brazil revenue fell \~26% in a quarter. Softer “redundancy” versions hit Egypt and Mexico. All in the biggest markets. dLocal’s own response — an orchestration product at explicitly lower take rate — reads like an admission they can keep the volume but not the margin.

The counter is that expansion has crushed the leakage so far (NRR 140%+, merchants going 2→21 and 19→40 countries), and the moat still protects the \~35 long-tail markets no one will ever build. But I don’t know how sustainable that is, as those are much smaller markets, and the same logic could apply.

Any thoughts?