On April 30, 2026, the U.S. Patent and Trademark Office published a patent application from Capital One Services, LLC titled “Digital Service Authentication Using a Contactless Card” (US20260122061A1). A published application is an ~18-month-delayed window onto where a company directed its filing effort. This one surfaced in the blockchain publication column for the week, but reading it shows why it is a useful counter-signal: the blockchain is at the edge of the disclosure, not its subject.

The disclosed method is about authentication. It verifies a user's login for a set of digital services, generates a cryptogram carrying the verified credentials, and stores that cryptogram on a contactless card so the card can automatically authenticate the user later. The application states its core directly:

A method, a system, a device, and a computer program product for executing authentication.— Digital service authentication using a contactless card, US20260122061A1

The chain appears only in the supporting description, where switchboard nodes can invoke transactions on a blockchain to add operational data — an optional record-keeping layer. The two CPC classes on the published record, H04L 63/0853 and H04L 63/083, are network-authentication codes; there is no blockchain-specific class on the filing at all. The document is a card-and-credentials disclosure that mentions a ledger, rather than a ledger disclosure.

A portfolio that points elsewhere

That placement is consistent across Capital One's recent publications, and the contrast with its overall volume is the point. The company is a prolific filer — the records show well over 3,000 published applications — but its filing effort is concentrated away from the chain. Recent applications cluster around contactless cards and physical-card mechanics: distributed keys generated by contactless cards for area access (US20260162481A1), synchronizing authentication attempts across a card and a mobile device (US20260155973A1), and a chip card with sonic signaling to help a user find it when lost (US20260161904A1).

A second large strand is authentication and identity infrastructure — supplemental FIDO keys in a switching network (US20260162111A1) and universal identity verification for FIDO passkey registration (US20260162110A1) — and a third is machine learning, which dominates the company's class facets: G06N 20/00 (general machine learning) leads its published portfolio with 484 records, well ahead of any payment or ledger class. Decentralized edge-device payment processing using a machine-learning model (US20260162117A1) is a representative recent filing: the word that matters in it is the model, not the ledger.

What the thin footprint signals

Read as a body, Capital One's published applications point to a direction in which the blockchain is a peripheral component. Where the chain appears, as in the April 30 contactless-card application, it is an optional place to log operational data alongside the real subject — cards, keys, authentication, and fraud models. For a general reader, the grounded inference is straightforward: the company's recent filing effort indicates investment in physical-card security and machine-learning authentication, and treats a ledger as a supporting option rather than a product line. The absence of a dedicated on-chain cluster is itself informative against the bank and network assignees that do file chain-specific work.

The caveats are real and, here, double-edged. These are published applications, not grants; claims can change before issuance, and a filing shows intent to seek coverage, not a shipped feature. And blockchain crypto-specific filing volume is genuinely thin across the sector — the week of April 30, 2026 produced roughly 145 blockchain-tagged publications, most from unnamed assignees and many, like this one, mentioning a ledger only incidentally. That thinness is part of the finding: among recognizable financial-sector filers, Capital One's disclosed on-chain footprint is small, and its April application is a clear example of the chain sitting at the edge of the work rather than at the center.