🇺🇸US Financial and Crypto Licenses: Three Axes of Choice

Concept

"Getting a US license" means nothing until three coordinates are fixed. Most people hold one axis in mind — federal versus state. In practice a regime sits at the intersection of three. First, who grants the permission: a federal agency (OCC, FDIC, the Federal Reserve, FinCEN, SEC, CFTC) or a state authority, of which there are more than fifty once territories are counted. Second, deposit status: may the institution take deposits, and are they FDIC-insured. Third, access to Federal Reserve settlement: does it hold an account at a Reserve Bank, or do its dollars travel through a correspondent.

The third axis long looked like plumbing. By 2026 it is the single biggest driver of a product's economics. An institution with a Fed account settles in central bank money and does not depend on whether a sponsor's compliance department still likes it. An institution without one rents access, pays the spread and lives with the risk of an abrupt exit. This axis separates a real license from a nameplate: a charter that does not open the settlement door delivers a sign above the shop, not a business model.

The map answers two questions: what actually backs the money on an account — FDIC insurance, a full reserve, segregation of client funds, or nothing beyond an intermediary's balance sheet — and which door to file at, since in the United States picking the wrong one costs years and a redrawn structure. Non-US groups are not exempt: serving US customers pulls a foreign operator into state money transmission, and from 2028 into the stablecoin perimeter.

Section map: which question to start from, and where the answer is.

The Regime Table

The first table is each regime's perimeter: who grants it, what it permits and who it suits.

The second is the price of entry for the same regimes: capital and timeline.

Three narrow charters complete the list. The federal savings association is a working but unpopular track, because the qualified thrift lender test in 12 U.S.C. § 1467a(m) ties the balance sheet to mortgage and consumer assets. The CEBA credit card bank keeps a commercial parent outside the Bank Holding Company Act only while it meets all five conditions of 12 U.S.C. § 1841(c)(2)(F), which confine it to card issuance funded by large time deposits. Georgia's merchant acquirer limited purpose bank, created by a 2012 state act, gives a processor direct Visa and Mastercard membership without a sponsor bank; Fiserv was the first holder, chartered on 27 September 2024 and permitted to begin business on 11 April 2025. How these tracks compare with an ILC and a national trust charter on capital, timing and supervision is set out in the piece on industrial banks.

The Federal Banking Charters: OCC, FDIC and the ILC Queue

For the first time in fifteen years the federal charter track is genuinely open. On the OCC's own numbers as of 11 August 2026, the agency received 40 de novo applications in 18 months, where the 2011–2014 average had been fewer than four charter applications a year. Many decisions land within 120 days of a complete filing, and for the first time in five years a full-service national bank received final approval and opened.

The OCC Trust Charters

The structural shift is the trust charter without deposits. On 12 December 2025 the OCC conditionally approved five applications at once: First National Digital Currency Bank (Circle), Ripple National Trust Bank, BitGo Bank & Trust, Fidelity Digital Assets and Paxos Trust Company, joining roughly 60 existing national trust banks. Permitted activities span custody, settlement, staking and stablecoin issuance. The OCC release itself states no capital figures: individual conditions are issued as separate decision documents, and on law firm analysis of those decisions Tier 1 capital ranged from $6.05m to $25m.

Earlier, on 15 October 2025, Erebor Bank took a preliminary conditional approval. Among 2026 decisions Coinbase's is primary-source confirmed: Corporate Decision No. 1370 permits institutional custody only, no retail deposits and no lending.

One caveat: the full 2026 approval list diverges across sources, with trade summaries and law firm trackers classifying Bridge, Laser Digital and Morgan Stanley Digital Trust as either conditionally approved or still pending. Any name is checked against the OCC Weekly Bulletin and Corporate Applications Search; only press-release and corporate-decision confirmations are stated as fact here.

Conditional approval has stopped being a nod. The OCC moved the operational examination of a bank's architecture to the stage after it: organisers have 18 months to reach "no supervisory objection", and the capital fixed at approval can be revised upward, particularly for digital-asset banks where off-balance-sheet assets make conventional ratios uninformative. Hence the gap in the statistics — approvals are plentiful, yet for a long stretch the only fully operational digital-asset national trust bank was Anchorage Digital.

Denials have become specific. On 21 July 2026 the OCC rejected Wise's application for Wise National Trust on four grounds, none about the product:

1. a July 2025 Multistate Consent Order for BSA/AML failings;
2. organisers with insufficient knowledge of banking law;
3. a board and management short on fiduciary competence;
4. no experience under 12 CFR 9.

The OCC screens on people and compliance history, not business model; the denial does not bar a fresh application.

The legal frame moved too. The final chartering rule, adopted 27 February 2026 and effective 1 April, replaced "fiduciary activities" in 12 CFR 5.20 with "the operations of a trust company and activities related thereto", aligning with 12 U.S.C. § 27(a). The OCC maintains powers are neither expanded nor narrowed; the CSBS calls it a back door to full banking without insurance, the ABA wants new trust charters barred from the word "bank", and the BPI weighed litigation. Whether payments and money transmission sit inside "the operations of a trust company" is exactly what the rule does not answer, and exactly why many applicants are there: the regulatory risk falls after the charter, not before. The regime in full is covered separately.

The FDIC and the ILC Queue

The deposit side was rebuilt too. On 10 August 2026 the FDIC introduced a two-phase process: contingent authorization within 120 days of filing, then full approval over the following 12 months, giving founders certainty before they raise capital and hire. In parallel the ILC came back to life: a Utah or Nevada charter plus FDIC insurance delivers full banking functionality without bank holding company status for the parent — which is what lets a commercial company own a bank.

Four ILCs were approved in 2026, and the toll is regulatory capital two to three times the normal level.

All carry minimum 15% tier 1 leverage and dividends requiring FDIC consent for three years. The Nissan, PayPal and Affirm applications are pending, opposed by the BPI and ICBA. The queue is mapped in the piece on industrial banks.

The Third Axis: Access to Fed Settlement

The rules come from the Federal Reserve's Guidelines of 15 August 2022: three tiers of review — Tier 1 for federally insured institutions, Tier 2 for uninsured institutions under federal prudential supervision, Tier 3 for everyone else. The outcome is documented: per CRS, between December 2022 and May 2026 only three Tier 3 applicants obtained a master account, while roughly half of the tier's applications were withdrawn or denied.

Custodia and Kraken: One Charter, Two Outcomes

Two cases show the fork does not run along charter type. Custodia, a Wyoming SPDI, was denied Fed membership by the Board on 27 January 2023 and refused a master account by the Kansas City Fed the same day, on the reasoning that its model was novel and crypto-focused; a district court ruled against it in March 2024, rejecting the argument that the Monetary Control Act of 1980 guarantees access, and on 31 October 2025 the Tenth Circuit affirmed (No. 24-8024, 2–1), holding that Reserve Banks have discretion over master accounts; rehearing en banc was denied in April 2026.

Three years later the same Reserve Bank granted a master account to Kraken Financial, a Wyoming SPDI with a comparable model: 4 March 2026 was the first direct access by a digital bank to Fed payment rails. Conditions are strict — full reserve, at least 100% liquid assets against client fiat deposits, a phased institutional launch. The BPI notes the account is capped at one year, that Vice Chair Bowman called it a pilot, and that the decision "front-runs the Board's public comment process". What separates Custodia from Kraken is timing and negotiating position — the charter type is the same; both are examined separately.

The Payment Account: Stripped-Down Access

The door is now being institutionalised. On 20 May 2026 the Fed formally proposed a "payment account": a stripped-down account reaching Fedwire Funds, FedNow, NSS and Fedwire Securities on a transfer-only basis, with no FedACH, checks, FedCash or correspondent relationships, and no interest, intraday credit or discount window. Balance ceilings are individual under a hard $1bn cap; review runs 45 days for Tier 1 and 90 for Tiers 2 and 3; Reserve Banks were advised to pause Tier 3 decisions until 31 December 2026.

Companion changes to the PSR Policy, Regulation D and Regulation A appeared on 26 May, comments closing 27 July 2026. A day earlier an executive order of 19 May gave the Board 120 days to assess its authority to extend payment access to uninsured depository institutions and nonbank financial companies.

What matters most is the omission: the pool of legally eligible applicants does not widen, and remains depository institutions. A fintech without a depository charter still cannot open the door; the real beneficiaries are state SPDIs, uninsured national trust banks and bank stablecoin issuers. The absence of FedACH hollows the product out for payroll and B2B flows — a point pressed by crypto firms demanding broader scope and by community banks objecting to the concept entirely. Parameters and positions are in the piece on payment accounts.

Stablecoins: the GENIUS Act and Three Routes

The GENIUS Act, Public Law 119-27, was signed on 18 July 2025 after votes of 68–30 in the Senate and 308–122 in the House. It creates the status of permitted payment stablecoin issuer and three routes into it: a subsidiary of an insured depository institution; a federal qualified nonbank issuer approved by the OCC; and a state qualified issuer under state law, available only while outstanding issuance stays below $10bn.

Everything turns on 18 January 2027. The Act commences on the earlier of two triggers: 18 months from enactment, or 120 days after final rules. Not one final rule exists out of roughly twenty-five required, and every agency missed the statutory deadline of 18 July 2026 — so the hard calendar date governs. The paradox is that delay pulled the deadline closer rather than pushing it away: the later the final texts arrive, the less time remains to comply, and issuers will build against unadopted drafts.

The Federal Route: the OCC's Draft Rules

The drafts are already substantive. The OCC's NPRM, published in the Federal Register on 2 March 2026, creates a new 12 CFR 15 and puts 211 questions to the market. Per Sullivan & Cromwell, capital is set case by case rather than as a single minimum; the draft's key parameters:

On these parameters the federal route can already be costed; the state route cannot yet.

The State Route and Foreign Issuers

The state route exists on paper only. Eligibility depends on the state regime being recognised as substantially similar; Treasury only proposed the assessment principles on 1 April 2026, the decision rests with the Stablecoin Certification Review Committee — Treasury Secretary, Fed Chair and FDIC Chair, unanimously — and the expedited review procedure has been overdue since 14 January 2026. No state is certified. The $10bn threshold is measured on consolidated issuance, and breaching it forces migration to federal supervision within 360 days, so "start in a state and decide later" writes a change of regulator into the business plan.

Foreign issuers have their own route: OCC registration after a Treasury comparability determination, default approval at 30 days, and a hard stop of 18 July 2028 for access to US digital asset service providers. How the legal forms of a dollar token relate is set out in the pieces on the digital dollar and stablecoins, the statute in GENIUS Act, and the first token from a US national bank in the SoFi case.

The States: Money, Crypto and Fifty Filings

The nonbank route begins with something that is not a license. FinCEN MSB registration on the RMSB form, successor to Form 107, is filed within 180 days of formation and renewed every two years; it authorises nothing and merely fixes BSA obligations — an AML program, reporting, a five-year agent list. Sanctions are out of proportion to the paperwork: the statutory $5,000 per violation applies as $10,556 after the 31 CFR 1010.821 inflation adjustment of 17 January 2025, each day counting separately, plus criminal exposure of up to five years. The regime is unpacked in the MSB piece.

Authorisation comes from the states. Money transmitter licenses are issued by 49 states, the District of Columbia and the territories through NMLS, and the economics of full coverage are sobering — on industry consultants' figures the bill looks like this:

The CSBS model law has been adopted in whole or in part in 31 states on the official CSBS count of February 2026 — about thirty-three counting the 2026 Louisiana and Oklahoma enactments — reducing divergence without removing the need for fifty filings. Practice is covered separately.

Two states layer their own crypto regimes on top, and a third offers a depository charter:

Adjacent Regimes: Securities and Derivatives

Where the product is an instrument rather than money, the perimeter changes entirely. Securities dealing and brokerage require broker-dealer status with FINRA membership and net capital under Rule 15c3-1; a secondary venue for tokenised securities requires ATS registration; registry maintenance requires transfer agent status on Form TA-1, or 12 CFR Part 341 for banks. Asset management splits by RAUM threshold.

Private funds and venture managers use the lighter exempt reporting adviser status. Derivatives and leveraged spot fall to the CFTC — DCM, FCM, DCO — and since the initiative of 4 August 2025 a listed spot market has been built on existing CEA authority, joined by perpetuals in 2026.

The legislative frame remains unfinished. The CLARITY Act cleared the House on 17 July 2025 by 294–134 and was approved by Senate Banking on 14 May 2026, but reached no floor vote before the August recess: Republicans lack the votes for cloture, and secondary estimates of passage in 2026 have fallen to roughly one in three. If enacted it takes effect after 360 days. That logjam is why the SEC took the rulemaking route: after cancelling its open meeting of 14 August 2026, the Commission proposed Regulation Crypto Assets on 19 August — a startup exemption of up to $5m over four years and a fundraising exemption of up to $75m per 12 months, with comments open to 20 October 2026. The price of speed is reversibility: what one administration installs by rule, the next removes by rule.

Choosing a Regime by Product

• Insured deposits and lending at a financial company. A national bank charter: OCC, FDIC insurance, holding company status with the Fed. A decision is possible in 120 days, but 18 months in organisation and a likely upward capital revision follow.
• The same, but the parent is non-financial. An ILC in Utah or Nevada plus FDIC insurance. The door has been open since January 2026; the price is 15% tier 1 leverage and capital from $150m.
• Fiat deposits without insurance or lending, crypto business. A Wyoming SPDI at 100% reserve — the only track by which a nonbank crypto operator has actually obtained a master account, and then only as a one-year pilot.
• Custody, fiduciary services and digital asset settlement without deposits. An OCC national trust bank, where capital is set case by case ($6–25m across the December 2025 decisions, on law firm analysis), or a New York limited purpose trust company if one state is enough — it also removes the New York MTL.
• A stablecoin above $10bn. Federal only: the OCC as federal qualified nonbank issuer, or an insured bank subsidiary. Budget a $5m floor, bespoke capital, 12 months of expenses in HQLA, weekly reporting.
• A stablecoin below $10bn. State status is formally available and practically not: no regime is SCRC-certified, and breaching the threshold triggers a 360-day migration.
• Moving money or crypto without bank status. FinCEN RMSB registration plus an MTL in every state of presence, with a BitLicense for New York and DFAL for California on top.
• Securities, tokenisation, derivatives. Broker-dealer and FINRA, ATS for a venue, TA-1 for a registry, RIA or ERA for management, CFTC for derivatives.
• No license of one's own. A sponsor bank, appointed-agent status under a licensed transmitter, or payment facilitator on the agent-of-payee exemption — cheaper to enter, dearer to supervise.

Without a License: the Sponsor Bank and Its Price

The BaaS model is alive, but more expensive and fully transparent to examiners. The interagency guidance on third-party relationships of 6 June 2023 and the joint statement on deposit products delivered through third parties of 25 July 2024 turned sponsor selection into two-way diligence. On an industry register's tally, nine of fifteen major sponsor banks have or have had published federal actions: Cross River, Sutton, Piermont, Evolve, Silvergate and a second Lineage consent order dated 24 June 2026, with the Blue Ridge and Thread orders terminated in November and December 2025.

The direction of enforcement matters more than the count: crypto deregulation is not BSA deregulation. In an OCC order from April 2026, published on 21 May, the findings against a community bank that had "significantly expanded" its processing capability relative to its size come down to not understanding its fintech partners' customer base, transaction flows and geographic exposure; the remedy runs to an independent consultant, a CDD rebuild and a SAR lookback. The economics of renting access are set out in the pieces on BaaS and sponsor banks and renting a license, and one alternative route to dollar settlement in the analysis of Puerto Rico banks.

The legal character of the 2025–2026 reversal is worth holding onto: it was executed through withdrawals, not new permissions. OCC Bulletin 2025-2 of 7 March 2025 removed the supervisory non-objection requirement and restored the earlier interpretive letters; the Fed and FDIC withdrew their SR letters and joint statements on 24 April 2025; the Fed policy statement of 22 December 2025 rescinded the January 2023 presumption. The permissive base is the old one, and a future administration can reverse it with the same instruments, without full rulemaking.

The Calendar to 2028