Goldman and Citizens Flag 3 Bank Crypto Businesses. Here's the $160M Fee Pool Nobody's Pricing

Want the short answer or the long one? Short one first. The money isn't in trading. It's in plumbing.

Here's the setup. Bank regulators and the SEC have spent the last several months clearing the runway for national banks to hold crypto in custody, settle transactions with stablecoins, and run tokenized infrastructure under charters they already have. Crypto Twitter's reaction was instant and predictable. Ape the L1s. Buy the exchange tokens. Same playbook as every cycle before this one.

That's the wrong trade. Real talk: Goldman Sachs and Citizens both ran the numbers on the rule change and landed on the same three words. Custody. Tokenization. Settlement. Not prop desks, not memecoins. The boring middle of the stack where every institutional dollar has to pass through exactly once.

The Raw Numbers

Start with the headline figure. Roughly $230 billion sits in stablecoins right now, up from about $130 billion in January 2024. Tokenized Treasuries, the cousin nobody tweets about, blew past $7 billion this year. BlackRock's BUIDL fund alone holds north of $2.5 billion. BNY Mellon, State Street and JPMorgan are all positioned to scale into that flow, and none of them need a single retail customer to do it.

Now the part that should wake you up. Custody fees run somewhere between 3 and 8 basis points annually on assets under custody. Run the math on $200 billion of institutional crypto flowing into bank vaults over the next 24 months. That's $60 million to $160 million a year in recurring, price-agnostic revenue. No market risk. No inventory. Just holding keys and filing paperwork.

Compare that to Coinbase, which pulled in roughly $1.3 billion in custody and other revenue last year while holding around $200 billion in custodial assets. Banks don't need to beat that number. They need to siphon 20% of the inflow and they've built a franchise.

Tokenization is the bigger prize. Goldman analysts pegged tokenized real-world assets as a multi-trillion-dollar category by 2030. Even the conservative versions of that model put bank-issued deposit tokens at the center of institutional settlement flow, because that's the only version of this where a CFO's compliance officer sleeps at night.

Why This Time Isn't 2020

Look, we've been here before. The OCC under Brian Brooks issued guidance back in 2020 saying national banks could custody crypto. It got walked back. Banks retreated. Remember Signature and Silvergate? Both died in 2023 without ever getting to scale the model.

Anon, let me explain the difference. In 2020 it was an agency letter with no statute behind it. Today it's a repeal of SAB 121 in January 2025, OCC interpretive letters in early March 2025, and actual legislation. The GENIUS Act got signed on July 18, 2025. That's a law with an 18-month implementation clock and federal rulemaking attached. Letters get reversed. Statutes don't.

That's the whole ballgame. Bank compliance departments don't move on guidance. They move on legal certainty. Write stablecoin reserve rules into the US code and suddenly the largest balance sheets on earth can touch this asset class without a legal memo that says 'probably fine.'

Here's my first hot take. The market is still pricing this as a crypto story. It's a banking story.

My second one is sharper. Crypto-native custodians charging 50 to 100 basis points for the same service are on a two-year clock. When BNY Mellon will hold your Bitcoin for 5 basis points and your auditor already trusts them, the premium evaporates. Coinbase and Anchorage don't die. They get repriced. There's a difference, and it hurts either way.

What the Analysts Are Actually Saying

According to the Goldman team, the ruling opens three distinct business lines at once. Custody for institutions that legally can't hold assets with a non-bank. Tokenization rails for funds, treasuries and private credit. And stablecoin settlement for cross-border flows that currently take two days and a correspondent bank. Citizens analysts went a step further, flagging that broker-dealers now have room to build onchain products directly on top of bank rails instead of routing around them.

Traders are watching the regional banks and that's the signal most people are missing. Citizens itself, Huntington, KeyCorp, Fifth Third. These are names with deposit bases, existing trust charters, and a desperate need for fee income that isn't net interest margin. Crypto custody is a fee business with a fixed cost structure. That's exactly what a regional bank wants in a rate-cutting cycle.

Is a $40 billion regional bank really going to out-execute Coinbase on crypto custody? Nope. Not on product. But it doesn't have to. It just has to be the only option that a pension fund's investment committee will approve without a three-hour meeting.

What's Next

Three concrete things to track.

First, the GENIUS Act rulemaking calendar. Federal regulators have until roughly January 2027 to finalize stablecoin issuer and reserve standards. Every draft that leaks moves the deposit token trade. Watch for the Treasury study on payment stablecoins, which lands well before the deadline.

Second, Q4 and Q1 earnings calls. Listen for the words 'digital asset custody' on regional bank calls. State Street has publicly targeted custody launch this year. BNY already has the exemption it needs. The moment a mid-cap bank says those three words on a call, the repricing starts.

Third, watch Fed master account applications from crypto-adjacent trust companies. That's the quiet gate. If a tokenized settlement entity gets direct Fed access, bank-issued deposit tokens become the default settlement layer for institutional crypto and the third-party stablecoin market splits in two.

Here's the thing about all this. The chain doesn't lie, and the chain is telling you where the fees go. Not to whoever ships the fastest DEX. To whoever holds the assets and signs the settlement. That's a bank charter business now, and it's a lot bigger than people realize.

Position accordingly. Or keep aping the L1s and wonder why the chart doesn't move when the news is good.