Read the two decision letters side by side and you find a contradiction that most of the coverage skipped.
Revolut, a global payments firm operating in more than 39 countries, got a national bank charter while projected digital assets at under 2% of total revenue and stated it does not intend to hold any on its balance sheet.
OpenReserve, a startup with a $25 million seed round and no bank yet, got the harder full-service charter to do the opposite: take deposits, make loans, and, in limited operational amounts, hold crypto as principal.
The firm with the deepest crypto ambitions is the one being told, in writing, exactly how much capital it must post before it can lend a dollar. That's the story. Not the charter.
The firm with the deepest crypto ambitions is the one being told, in writing, exactly how much capital it must post before it can lend a dollar. That's the story. Not the charter.
Both letters landed on September 2 under Comptroller Jonathan Gould, whose OCC has been reading firms working on "novel technology" into the federally supervised banking system at a pace that drew a National Bank Act challenge from Sen. Elizabeth Warren back in May.
On the surface, it reads as a clean win. Two firms with crypto-related plans reportedly cleared for national charters in a single week. The permissive era, confirmed.
Then you get to the numbers bolted onto OpenReserve's approval, and the win starts to look like a leash.
OpenReserve must raise at least $210 million in initial paid-in capital, net of organizational costs, within 12 months — a deadline that lands in September 2027.
It must open within 18 months, by March 2028. Miss either mark and the preliminary approval expires.
And through its first three years of operation, it must hold a Tier 1 leverage ratio of at least 12%.
That last condition reframes everything, and it's worth being precise about why.
The Tier 1 leverage ratio is capital divided by total assets — it caps how many dollars of assets a bank can carry per dollar of capital, regardless of how safe those assets look. A 12% floor arguably means OpenReserve can, by the author's own calculation, hold roughly eight dollars of assets for every dollar of Tier 1 capital.
By conventional standards, a well-capitalized bank reportedly clears that leverage bar at 5%, meaning the OCC has told OpenReserve to hold more than double the usual standard, and to keep holding it for three years, though this benchmark and comparison are not established in the underlying decision letters.
Leverage is not incidental to banking. It's the engine. A bank creates credit by lending out multiples of its capital base; the ratio between the two is the throttle on how fast that engine runs.
Set the throttle at 12% and you have deliberately slowed the machine.
Now recall what OpenReserve is selling. The a16z-led thesis, reportedly laid out when the firm raised its seed, is a "continuous" bank — one that never closes, settling on a programmable ledger instead of the batch rails of ACH and wire transfer.
The pitch is explicit that stablecoins can move money continuously but can't create credit, while banks create credit but can't run natively on-chain — and OpenReserve exists to fuse the two: regulated credit built on top of always-on money flow.
Credit creation is the whole point. And credit creation is precisely what a 12% leverage floor constrains.
That's the tension that arguably follows. It didn't reject the "always-on bank." It admitted it into federal supervision and then attached capital conditions that may determine whether the model works.
The permanent-uptime thesis assumes the bank is always doing something — settling, lending, recycling capital around the clock. A high leverage floor means every one of those loops has to be backed by an unusually thick capital cushion. Continuous is cheap to promise; continuous at 12% Tier 1 is expensive to fund.
So the real experiment isn't whether a blockchain-native bank can get a charter. It's whether the growth model survives the capital regime.
Compare that to the path OpenReserve chose to reject. Coinbase, Circle, Ripple, Paxos, and BitGo all took national trust charters, which permit custody and fiduciary work but block deposit-taking and lending.
Choubey reportedly framed the divergence this way: "We chose the national bank path deliberately." The trust charter is the low-friction option — no deposits, no lending, no leverage ratio to defend. OpenReserve walked past it precisely to do the two things that trigger the capital regime.
That is a firm choosing the constraint. Which is why the constraint, not the charter, is the news.
Revolut's letter shows the other posture entirely. Its charter is full-service, and it does not intend to exercise trust powers — but it has drawn a hard perimeter around crypto.
Digital asset services are projected at under 2% of total bank revenue across the three-year de novo period. The bank does not intend to hold any digital assets on its balance sheet. Custody runs in a nonfiduciary capacity through a UK affiliate, and the whole reason Revolut wants the charter is cost and efficiency — replacing its FDIC-insured partner-bank model, not making a bet on crypto as principal.
Revolut, in other words, took the charter and left the crypto exposure outside the bank. OpenReserve took the charter to bring crypto inside it.
The OpenReserve letter is specific about what "inside" means. The bank will hold gas-fee crypto on its balance sheet as principal — the OCC confirmed national banks may hold digital assets needed to pay network fees for a reasonably foreseeable need.
It will also receive fees denominated in digital assets, deducted from customer trades, staking rewards, or transfers — and it will convert those to fiat within one business day, or hold them for a permissible purpose like gas.
That one-business-day conversion, which OpenReserve itself represented, is arguably a supervisory tell. The OCC is willing to let a national bank touch crypto as principal, but only in tightly scoped, quickly-cleared amounts. No accumulation, no speculation, no balance-sheet exposure beyond operational need.
Between the leverage floor and the conversion window, the pattern is consistent: yes to the activity, no to the risk that usually comes with it.
There's a third way to read the strategic fork, and it belongs to SoFi. Rather than buy a charter to reach crypto, SoFi reportedly took what could be read as a rent-not-buy approach, wiring itself to Kraken Prime for what is arguably institutional crypto plumbing and listing a SoFiUSD stablecoin.
Rent versus buy. SoFi gets crypto exposure without submitting its balance sheet to the OCC's capital math. OpenReserve bought the hardest possible version of the charter and now owns the capital math outright.
That's the wager worth watching. The OCC under Gould is often described as the permissive regulator, and the two September 2 approvals fit the frame. But permissive on entry is not the same as permissive on terms.
The agency let a VC-backed, blockchain-native startup into the national banking system precisely so it could hold that startup to bank capital discipline — a $210 million floor, a 12% leverage cap, and a clock.
If the "continuous bank" thesis is real, it has to prove it can compound at double the standard leverage cushion, on a deadline, under federal supervision. If it can't, the constraint will have done its work before the bank ever opens its doors.
Either way, the answer arrives by September 2027 and March 2028. The charter was the easy part. The capital regime is the test.