Fintech Highlights — Tokenization Wants the Credit Stack
Securitize is buying its way from tokenized assets into asset-backed lending. Plus payment infrastructure for agents and Revolut's global-bank expansion.
Here's what we've been reading this week…
This week, tokenization moved closer to the credit stack, payment infrastructure started opening directly to AI agents, and one of fintech’s most valuable companies showed that licenses—not just software—still determine how far a financial product can travel.
Pinned to the Top
Tokenization wants the credit stack
Securitize agreed to acquire River One, an institutional asset-backed financing platform that connects warehouse credit facilities with stablecoin capital.
Securitize already handles the issuance and servicing of tokenized assets. River One would add more of the financing lifecycle: origination, structuring, borrowing-base reporting, collateral monitoring and funding operations. The transaction is expected to close in the fourth quarter, subject to customary conditions.
That is the meaningful shift. Putting an asset onchain changes the wrapper. Connecting the asset to underwriting, collateral controls and a source of financing changes what the platform can do.
Tokenization has spent years focused on issuance. The harder—and potentially more valuable—work is building the credit infrastructure around the asset after it exists.
The BFD
The financial product is escaping the app
For years, the app was the product. It controlled the interface, the customer relationship and the actions a user could take.
That boundary is starting to loosen.
Grasshopper Bank lets authenticated business customers query live banking data from Claude and ChatGPT through read-only connectors. Pliant is opening commercial-payment infrastructure to outside agents through Model Context Protocol connections. Chime says agents could make it easier for consumers to compare and switch financial products—and is planning secure access for customer-chosen agents.
These are not equivalent products, and most transaction authority remains limited or prospective. But they point toward a different distribution model: the institution remains the regulated system of record while the customer chooses the interface that calls it.
If that continues, proprietary screens become less defensible. Product economics, data quality, permissions, reliability and trust matter more.
The financial product does not disappear. It becomes a permissioned service another interface can use.
Read below for more.... 👇
Payments become callable infrastructure
Pliant raised an $80 million Series C and says it has passed $100 million in annual recurring revenue. The more important product move is opening its commercial-card, account, foreign-exchange and spend-management infrastructure to AI agents.
Outside agents can connect through MCP, while permissions, controls and auditability remain in the payment layer.
That is a practical answer to the agentic-payments problem. Intelligence can sit in the agent, but authority should remain attached to regulated infrastructure that can define limits, record activity and stop the transaction.
The funding is useful. A scaled payments platform making its functions callable is the stronger signal.
The bank account gets a different interface
Grasshopper’s experimental connectors give business customers read-only access to balances, transactions and other banking data from Claude and ChatGPT.
The controls are intentionally narrow. The AI cannot initiate a transaction or change the account. Customers grant access explicitly, administrators can manage or revoke it, and the connection must be re-authenticated within the AI platform every seven days.
That makes this less dramatic than an autonomous financial agent—and more useful as a product model. Start with data access. Keep authority out of scope. Learn how customers use the interface before adding higher-risk actions.
Agents could weaken distribution inertia
In its third-quarter presentation, Chime argued that AI agents will make it easier for consumers to compare financial products and switch providers. It also described plans to scale its own personal-finance agent and create secure access for customer-selected outside agents.
Those are strategic plans, not proof of a broadly deployed product. But the distribution argument matters.
Large banks have historically benefited from customer inertia. If an agent can compare fees, features and trust—and eventually help move the relationship—the institution with the best app does not necessarily win. The institution with the best underlying product might.
From the Stash
Revolut’s valuation follows the licenses
Revolut completed a secondary share transaction establishing a $115 billion valuation.
The number is eye-catching. The operating footprint explains more. Revolut says it now serves more than 80 million retail customers and, during 2026, launched banks in the UK, Mexico, Australia and New Zealand; secured banking licenses in France, Colombia and Argentina; and received conditional approval for a U.S. national bank charter.
Revolut is not simply distributing one fintech product across borders. It is assembling locally regulated banking infrastructure market by market.
For product teams, that is the real moat behind the valuation: licenses, operations and balance-sheet capabilities that let the same customer proposition become a bank in more places.
That’s it for this week. Thanks for reading. Please pass along 😄