Two companies dominated the alerts this week: Checkout.com and Mambu. Both are in the database, both had interesting news, and together they tell a coherent story about where payment and banking infrastructure is heading. That’s this week’s issue.

Featured company: Checkout.com

Checkout.com doesn’t get the consumer press coverage of Klarna or Revolut, which is partly by design — it’s a B2B payments infrastructure company, and its customers are companies like Agoda, Wise, and Binance rather than individual shoppers. But the news coming out of the company this week is worth paying attention to.

The headline partnership is with Agoda, the digital travel platform. Agoda is integrating Checkout.com’s infrastructure to handle AI-powered payment performance at scale — optimising authorisation rates across markets, currencies, and payment methods for one of the world’s largest travel booking platforms. It’s a significant enterprise win and a signal of where Checkout.com is positioning itself: as the payments layer that handles complexity at scale, not just transaction processing.

The more interesting story though is a piece of research Checkout.com published in June. The company is forecasting that agentic commerce — AI agents making purchases autonomously on behalf of humans — could reach 1.3 billion users globally by 2031. That’s not a small number. The implication is that the next wave of payment volume won’t come from humans tapping their cards or phones. It will come from AI systems making purchasing decisions, triggering transactions, and managing payment flows without a human in the loop for each transaction.

This matters for payment infrastructure companies because agentic commerce has different requirements than human-initiated payments. An AI agent booking flights, buying software subscriptions, and managing a procurement workflow needs payment rails that can handle programmatic access, machine-readable responses, and authorisation flows that don’t rely on a human completing a CAPTCHA. The companies building for this now are making bets that will compound significantly if the forecast is anywhere close to right.

Checkout.com also added an Equiti Group partnership this week, expanding its worldwide payments reach through Equiti’s brokerage and financial services network. A quieter piece of news, but consistent with the pattern of steady enterprise expansion.

→ View the Checkout.com profile: fintechdatabase.eu/company/checkout-com

Also in the alerts: Mambu

Mambu, the cloud-native core banking platform headquartered in Amsterdam, appeared in two places this week. Analytics Insight included it in their top embedded finance companies to watch in 2026 — alongside the usual suspects — describing it as providing “cloud-native core banking technology supporting embedded finance, digital banking transformation, lending operations.”

The more interesting piece was from Fintech Finance News on what they’re calling the Digital Sidecar Migration Model. The concept is worth understanding: rather than attempting a full rip-and-replace of a legacy core banking system (which is expensive, risky, and often fails), the sidecar approach runs a modern cloud-native core alongside the legacy system, gradually migrating products and customers across. The legacy system keeps running until it’s no longer needed. The new system grows until it handles everything.

It’s a pragmatic solution to a problem that has cost several major banks hundreds of millions of pounds in failed core banking programmes. And it’s the kind of approach that benefits platforms like Mambu, which are designed to run alongside existing infrastructure rather than demanding a full commitment from day one.

→ View the Mambu profile: fintechdatabase.eu/company/mambu

One insight

The agentic commerce forecast is the kind of number that sounds speculative until it doesn’t. 1.3 billion users in five years is a large claim. But consider: AI assistants that book restaurants, manage calendars, and send emails on behalf of users are already mainstream. The jump from “AI that schedules a meeting” to “AI that purchases the coffee for the meeting” is not technically large. It’s a permissions and payments infrastructure question.

What’s not built yet is the payment layer that makes agentic commerce frictionless. Human payment flows are designed around friction — 3DS authentication, one-time passwords, CAPTCHA verification — because fraud prevention has historically required a human to prove they’re human. An AI agent can’t complete a CAPTCHA. The companies solving this — through tokenised payment credentials, agent identity frameworks, and programmatic authorisation flows — are working on infrastructure that most people haven’t heard of yet.

Whether Checkout.com’s 1.3 billion number is right or wrong, the direction is clear. The companies building payment infrastructure today are betting that a significant proportion of future transaction volume will be machine-initiated. That’s a different product requirement than what the last twenty years of consumer payments have needed.

One question for you: which European fintech do you think is most quietly important — the one doing genuinely significant work that gets almost no coverage?

Hit reply. I read everything.

— The European Fintech Digest