Two is the number that matters, and it is not in Block's filing.
On September 10, 2025, Block, Inc. applied to the Office of the Comptroller of the Currency for a national trust bank charter. The proposed institution is Builders Bank & Trust, N.A. โ non-insured, authorized to provide Bitcoin and stablecoin custody alongside traditional trust services. Every write-up that followed led with Jack Dorsey's name, or the phrase federal regulatory framework, or the size of the parent company. None of that is the story.
The story is the denominator. Through 2025, the OCC has processed roughly 40 crypto-adjacent charter applications. It approved 21. It denied 2. A licensing regime that grants 52.5 percent of what it receives and refuses five percent is not a gate. It is a queue.
I have spent four months this year reading custody attestation reports from federally chartered and state-licensed digital asset custodians. The single most reliable predictor of a custody failure is not the charter class. It is the sub-custodian chain that nobody disclosed. That distinction is where this filing actually matters, and it is where every headline about it has stopped short.
Block is not a crypto startup. It is a public company trading under XYZ with four operating surfaces that touch digital assets. Square handles merchant acquiring. Cash App holds retail Bitcoin balances and has done so since 2018. Bitkey ships a physical hardware wallet built on a 2-of-3 multisig architecture. Proto manufactures mining ASICs. Jack Dorsey founded the company, chairs it, and has publicly held Bitcoin for years without pretending otherwise.
The relevant precedent is Square Financial Services, which received an Industrial Loan Company charter in 2021 โ a genuine deposit-taking institution, insured, supervised jointly by the Utah Department of Financial Institutions and the FDIC. That matters. Block has already walked the federal banking application process once. It is not an applicant learning the vocabulary mid-sentence.
Now the filing itself. Builders Bank & Trust, N.A. A national trust bank. Non-insured. The proposed president and CEO is Lee Woolley, whose public professional history does not appear in any coverage I can find. I flag that as a gap, not a scandal. In a trust bank application, the biography of the proposed fiduciary officer is normally the most scrutinized document in the packet. Analyzing an institution without knowing who runs it is an incomplete exercise, and I would rather name the hole than fill it with inference.
The competitive frame is crowded, and it is worth laying out flat. Coinbase Custody has institutional scale. Paxos has held a trust charter for years. BitGo, Ripple, and Circle hold charters or have applications in flight. Revolut received conditional approval the week before Block filed. World Liberty Financial โ the Trump-family-affiliated venture โ received conditional approval as well, a data point that carries its own political payload.
So this is not a novel event. It is the twelfth or thirteenth pass at a template. What makes it worth dissecting is not novelty. It is the scale of distribution sitting behind it.
I will take this apart in the order that matters: what the charter buys, what it actually protects, what Block is assembling, why the approval math matters more than the approval itself, and where the failure will hide.
First: what a national trust bank buys, and what it costs.
A trust bank is a fiduciary institution. It holds assets for the benefit of others. It does not take deposits, which is precisely why it is non-insured โ the FDIC insures deposits, and there are no deposits to insure. That is not a concession Block made under pressure from the OCC. It is the design. Non-insured trust charters carry lower capital requirements and generate less political resistance than full bank charters, which is why they have become the default path for crypto custodians seeking federal cover.
The operational prize is federal preemption. Without a federal charter, a custodian serving clients nationwide assembles a patchwork of state money transmitter licenses โ each with its own net-worth covenant, its own reporting cadence, its own examiner, its own political exposure. Fifty jurisdictions means fifty renewals, fifty supervisory relationships, and fifty sets of books to reconcile.
I priced that patchwork once, for a custody structure in 2022 that never launched. The internal model came out at roughly twenty to forty basis points of annual custody revenue consumed by state licensing compliance on a mid-size book. A national trust charter does not erase that number. It cuts it by half and, more importantly, makes the residual predictable. Predictability is the product being bought here. The charter is not a technology. It is a fixed cost expressed as a legal entity.
Second: bankruptcy remoteness is the real asset, and almost nobody is discussing it.
Under a trust structure, client assets are legally separated from the corporate estate. If Block fails, the Bitcoin held at Builders Bank & Trust does not enter the bankruptcy estate. It is held in trust and returns to the beneficiary. This is the single most valuable thing a custody client purchases, and it is the reason an allocator will pay a fee to a chartered trust company rather than leave assets sitting on an exchange balance sheet.
The reason it goes unexamined is that it is untestable until it is tested. A bankruptcy-remoteness claim is a legal promise with a proof burden that only arrives during insolvency. Every custodian advertises it. Almost none have had it stress-tested in a courtroom. The ledger remembers what the promoters forgot โ but only after the petition is filed, and by then the fee has been collected for years.
Third: the vertical stack, which is the real thesis and the real seam.
Here is what Block is assembling. Proto makes mining ASICs. Those machines produce Bitcoin. Bitkey holds self-custodied Bitcoin in a consumer's hand. Cash App distributes Bitcoin to retail. Square touches merchants. And Builders Bank & Trust would custody Bitcoin and stablecoins at institutional grade under federal supervision.
That is a loop: mine, hold, custody, transact. No competitor in the charter race has all four segments. Coinbase has custody and retail distribution but no ASIC line. BitGo has custody and nothing downstream. Paxos has custody and stablecoin issuance and no consumer surface. Ripple's charter sits inside a cross-border payments business with a different client base entirely.
This is the strongest version of the bull case, and note what kind of case it is. It is not a technology argument. It is a distribution argument wearing a charter.
Now look at the seam. Bitkey exists to make self-custody easy. Its entire market position is that the user holds the keys and Block does not. Builders Bank & Trust exists to hold the keys for the user. Those two products point in opposite directions, and they are sold by the same company to overlapping customers.
If Block migrates Cash App Bitcoin balances into the trust bank โ and that is the obvious commercial move, because custodial balances generate fee income and self-custodied balances generate none โ it will have quietly reversed Bitkey's premise without ever announcing that it did. Silence in the code is louder than the contract. Watch which product survives the migration, because the migration is the only part of this story with revenue attached to it.
Fourth: the approval math, and the commoditization of the charter.
This is the part the bulls skip, and it is the part that determines whether any of the above is worth anything at all.
Twenty-one approvals against two denials is not a rigorous regime. It is a regime that has decided to say yes. When a license is granted to nearly everyone who asks, the license stops conveying advantage. It becomes a commodity input โ necessary to compete, insufficient to win. The same arc played out with state money transmitter licenses between 2015 and 2020, and with New York's BitLicense between 2015 and 2023. Scarce at first. Valuable at first. Then ubiquitous. Then table stakes.
Revolut received conditional approval the week before Block filed. World Liberty Financial received one too. If the same supervisory body approves a consumer fintech, a politically connected venture, and a Bitcoin mining and payments conglomerate inside a single quarter, the marginal charter is not a moat. It is a permit.
Which means differentiation shifts entirely to execution: settlement latency, sub-custodian quality, insurance structure, attestation frequency, and whether the institution is willing to publish a genuine proof of reserves rather than a letter from a mid-tier accounting firm. Every rug pull leaves a trail of gas fees; the trust bank equivalent is the compliance spend disclosed in the risk factors with no revenue line to match it.
Fifth: the stablecoin custody line that nobody is modeling.
If Block custodies stablecoin reserves under a federal trust charter, it inserts itself into the reserve-attestation economy. Circle pays for qualified custody. Ripple's RLUSD does too. The USDC reserve stack is largely short-duration Treasuries, held through money market funds administered by large asset managers and custodied at a global custodian bank.
A federally chartered trust company can compete for that mandate. And if the emerging federal stablecoin framework settles on federal supervision as the gold standard for reserve custody โ which is the direction of travel โ then OCC-chartered trust banks become the default counterparty for compliant issuers. That is a business-to-business revenue line with no consumer marketing cost and multi-year contract duration.
It is also the least-covered part of this filing. The headlines said Bitcoin and stablecoin custody. They did not say that stablecoin custody is a reserve mandate worth basis points on every dollar of a float that runs into the tens of billions.
Sixth: the operational tax, which is real and which nobody prices.
Federal fiduciary supervision is not a badge. It is a set of constraints. Examiner presence. Capital minimums. Periodic fiduciary audits. Restrictions on commingling. Limits on launching products without notice to the supervisor. Restrictions on self-dealing between the trust company and its parent.
Block moves at consumer-fintech velocity. Trust bank supervision moves at quarterly-examination velocity. That friction is structural, not cultural, and it will show up the first time Block wants to ship a custody-adjacent product on a two-week cycle and cannot.
Seventh, and this is where my own audit history is the relevant instrument.
In 2017 I spent four months reconstructing the bytecode of that cycle's most-hyped Layer-0 claims and found that a $120 million raise had produced a renamed fork of the Geth client. In 2020 I spent six weeks simulating stableswap impermanent loss under tail volatility and found a rounding path that could have drained $45 million from liquidity providers. In 2021 I traced ten thousand NFT mints to a single script running on a private server.
The lesson that carries into custody analysis is this: the structure you can read is never the structure that fails. What fails is the layer that never appeared in the marketing. For a trust bank, that layer is the sub-custodian chain. Nearly every large custodian delegates part of key management, cold storage, or geographic dispersal to a sub-custodian. That delegation lives in a footnote, not on a landing page.
So the questions I would put to Builders Bank & Trust are not about the OCC. They are: Who holds the second signature? Where are the geographic shards? Is the insurance primary or excess, and what is the aggregate limit against the aggregate custodial balance? How frequently is the proof-of-reserves attestation refreshed, and does it cover liabilities as well as assets? The 2022 cycle answered that last question brutally for every firm that had only ever attested assets.
Here is where the bears are wrong, and they are wrong in a specific and checkable way.
The bear case runs like this: Block is late, the lane is crowded, the charter is undifferentiated, and the news is already priced. All of that is true. None of it is decisive, because custody is not a technology race. It is a distribution business.
In a commoditized charter environment, the winner is whoever already owns the customer relationship when the license arrives. Block owns the consumer. Cash App has tens of millions of monthly actives, a meaningful share of whom already hold Bitcoin inside the app. Block owns the merchant. Block owns the hardware. When the marginal charter is worth nothing, the marginal customer is worth everything โ and Block has more of them than Paxos, BitGo, and Ripple combined.
There is a second point the bears miss. The political risk everyone is flagging โ that OCC friendliness is bound to the current administration and could reverse โ cuts in Block's favor in one specific scenario. If a hostile OCC takes over in 2027 and stops issuing new charters, existing charter holders become a closed club. Filing now is a bet on getting inside before the door closes. Being early to a closing door is not the same as being late to an open one.
What the bulls get wrong is treating the charter as the endpoint. It is not. The charter is a shelf asset until Block announces that Cash App balances are moving into trust custody. The filing is a filing. The migration is the event.
Watch one signal, and only one: whether Block moves existing Cash App Bitcoin holdings into Builders Bank & Trust, and what happens to Bitkey when it does.
A trust bank charter that never receives a client asset is a compliance artifact and a line item in a 10-K. A trust bank charter that absorbs Block's own retail Bitcoin book is a structural change in how a publicly traded company holds customer property โ and a quiet admission that self-custody was a product feature, not a principle.
The OCC will almost certainly approve this. That is not the question worth asking. The question is whether Block intends to use the license on its own customers first. If it does, the crypto custody market has a new distribution layer sitting on top of it. If it does not, we will have spent a year analyzing a permit.
