Three Takeaways from Fintech Nerdcon

#1 The Liquidity Bottleneck

Now that stablecoin applications have reached real volume, especially in its power use cases like remittances and dollar savings, local liquidity now constrains its momentum. To be clear, it’s a good (or at least better) problem to have. It succeeds more foundational challenges like jumpstarting adoption (done), securing regulatory clarity (mostly done), building issuance and custody layers (done), or gaining trust from a critical mass of early adopters (mostly done, at least in consumer). What is the new liquidity challenge? Specifically, platforms that accept move stablecoins need a reliable way to turn on-chain dollars into local currency at a fair, predictable rate i.e., stable access to FX, clear regulatory rules, and banking partners who won’t unexpectedly pause or limit settlement.

It is hard to imagine a solution to that problem that does not involve localizing. Some countries have strict capital controls. Some only allow certain entities to handle FX. Some have slow settlement windows or require manual approvals. Banks vary widely in how much risk they’re willing to take, how much float they allow partners to hold, and how quickly they’ll actually move money. In other words, just like acquiring customers, solving liquidity seems corridor-specific.

#2 Evolution of the agentic payments opportunity

The discussion around agentic payments, specifically where within agentic payments can startups play, has gotten more precise from even just a few months ago. When the idea of “agents making payments” first emerged, it was all too easy to believe new players could define the new infrastructure. That now feels naive, or at the very least reductive as (for better or worse) existing players from card networks to commerce platforms to payment processors have quickly invested in the new frontier. Google’s recent Agent Payments Protocol (AP2) is the clearest example. The protocol is a multi-party standard that answers some of the challenges like authorization and auditability; Google developed the standard in partnership with the likes of American Express, Mastercard, Adyen, PayPal, Revolut, Cloudflare, dLocal, and Coinbase.

This shifts where the opportunity for venture-backed companies sits. If incumbents end up defining the core standards that agents rely on, the open space for startups then is what breaks once they start using those standards (AP2 or other). Questions on my mind, or heard from smarter fintech thinkers: What happens when an agent misunderstands a multi-step checkout flow?; How should rewards, incentives, and loyalty schemes adapt when an agent rather than a human is making the decision?; What does credit underwriting look like if agents become actual economic actors?

#3 The uppercase Bank and lowercase bank are diverging

There’s nothing new in un-bundling and re-bundling, but this feels different. I see two overlapping definitions of a bank. For simplicity’s sake there is an uppercase Bank i.e., a chartered institution with regulatory obligations, risk management, capital requirements, and most important an actual balance sheet. Then there is a lowercase bank that is basically the customer-facing financial experience that rents an uppercase Bank’s balance sheet, license, and regulatory capacity. Some banks are both (e.g., J.P. Morgan). Others are strictly Banks (e.g., Column) or banks (e.g., Chime). The distinction seems to harden as:

• AI presents a generational opportunity to rethinking the lowercase banking interface without wanting (or needing) to be slowed down by uppercase banking concerns

• The undeniable success of the likes of chartered sponsor banks like Column demonstrate an attractive model in outright owning the uppercase Banking (rather than out of reluctance in chasing deposits)

What does that mean? Ideally, it is easier and easier to build new interfaces and rent uppercase Banking capabilities from a larger marketplace of financial institutions.