The $230B Stablecoin Machine Has One Moving Part Left: A Banker

Stablecoin supply just pushed past $230 billion. The number of US banks willing to custody issuer reserves still sits in the single digits. Read that gap twice, because it explains the next two years better than any price chart. Last month, a quiet commentary started circulating through institutional channels. The headline: Stablecoins Won't Scale Without Banks. Barely three paragraphs. No data. No code. No named projects. A pure posture statement — the kind of thing I normally sweep into the delete folder. I didn't delete it. Whispers before the ticker opens: every treasury desk testing dollar tokens keeps coming back with the same sour note. The token layer is fine. The settlement rails work. But the compliance operations around them run on phone calls to nervous bankers. That is the boundary the market really cares about now. The clock stops, but the chain doesn't. When the chain wants to touch the real world, it needs a banker to answer the phone. This is the weirdest possible moment for stablecoins. On-chain, the asset class looks triumphant — payments volume climbing, Wall Street tokenization pilots multiplying. But the plumbing underneath keeps narrowing. Reserves sit at a handful of supervised banks. Issuance runs through a few qualified custodians. Redemption depends on correspondent networks that have shrunk every year since 2023. That was the year the crypto-banking stack cracked. Silvergate pulled the plug. Signature was seized by regulators. Within weeks, USDC redemption flows showed visible stress, and the market learned a painful lesson: these banks weren't partners, they were lifelines. When lenders decide crypto is too hot to touch, the word "stable" suddenly becomes relative. Look at what emerged from that panic. Two issuers still dominate the float. Their reserve books — Treasury bills, cash, commercial paper — generated billions in annualized interest at recent rates. That yield is the actual business; the token is just a wrapper around the spread. The open question is whether that yield economy broadens to institutions — or whether banks realize the wrapper is the easiest part and build their own. Regulators have been preparing the battleground. Europe's MiCA regime now requires e-money licensing for issuers. US lawmakers keep circulating stablecoin bills, with the GENIUS Act being the latest iteration, tying issuance to charters and reserve transparency. The Federal Reserve keeps repeating one phrase in speeches: dollar tokens should be the province of supervised, regulated institutions. Which is why "no banks, no scale" is quickly becoming the consensus of every treasury desk I speak with. Consensus, in my experience, is the very first thing you should audit. So audit it. Start with the mechanics. What does "banks" actually mean in this context? It is not vague infrastructure gossip. In my exchange seat, I watched institutions attempt larger stablecoin allocations, and every time the checklist had to line up: one, reserve custody with a supervised institution — in practice, among the few lenders that will accept stablecoin reserves; two, a licensed issuance entity with capital buffers and AML obligations; three, a reporting rail that can show auditors where each token is sitting, in real time. The blockchain part is, genuinely, the easiest part of that stack. The other three are banking problems. Liquidity flows where trust is liquid. No credible institution allocates billions to a token because a blog post assures them the reserves exist. They need a bank statement, a custody chain, and a legal entity they can hold accountable. In my years auditing these arrangements, I learned how much of the industry's proof-of-reserves content was theater. Most of the famous "audits" covered a single wallet snapshot, not complete liabilities, and the attestations were quarterly at best, sometimes monthly snapshots with no continuous monitoring. Retail bought it. A pension fund treasury team? They read a page like that and laugh — before calling their bank counsel. Here is the insight that matters: the next wave of stablecoin demand will not be won by better protocol design. It will be won in custody agreements, audit reports, and banking approvals. The code is already commodity. Institutional trust is not. Then there is the quieter layer nobody says directly: the yield. Two dominant issuers sit on a reserve base worth hundreds of billions. At recent prevailing rates, that reserve book yields tens of billions a year. That interest is the real prize, and banks are beginning to want it for themselves. Watch how the conversation shifts when a big lender realizes the token is just a distribution channel for what is essentially a deposit franchise. Reverse-engineer the regulatory logic. If a bank issues a digital dollar that pays interest, the product starts resembling demand deposits or money-market funds — and under securities analysis, the Howey test lights up fast. Money invested. Common enterprise. Expectation of profits. Profits derived from the efforts of others. All four factors glow green. So compliant bank stablecoins will likely be structured as zero-yield stored-value instruments, a kind of digital checking account. And that, ironically, pushes the design logic away from the yield-bearing models crypto natives are building — toward something far more conservative. I have seen this pattern before. Right before the SEC approved the spot Bitcoin ETFs, the unusual options volume on Coinbase told me institutional money was positioning for something imminent while the public narrative was still full of doubt. Same discipline applies here. When a central bank or a major lender says nothing will change, look at what insiders are quietly building. Banks are filing tokenization patents. They are testing internal settlement coins. They do not need permission to prepare. Insider sentiment matches the theory. At a Miami side event during a stagnant stretch for markets, I was talking with a few protocol engineers and a stablecoin operations lead. The conversation drifted to re-staking risk and then, quietly, to the deeper fear. "I don't look at smart-contract risk," he said. "I look at Monday. If the bank opens a conversation with 'we can't continue this relationship,' nothing else matters." Precision like that stays with you. It maps onto the current split among the top issuers. One major player spent its capital building mainstream anchor relationships, pushing toward a reserve structure that mirrors traditional money funds, including custody relationships with established global banks. Another built scale on offshore and less regulated channels — the exact structure that grows brittle under new legislation. The business model race is essentially a race to become the most boring, most domesticated issuer. Speed is the only currency that matters. When the regulatory window opens — and everything current suggests it is opening — the entity with the strongest banking relationships snaps to scale first. The one with the weakest starts burning energy hunting for offshore alternatives and hoping the legislative tide turns before the capital does. One more data point usually gets ignored: this framing is heavily American. Step outside the US and the story looks less linear. In emerging markets, dollar-token usage is exploding for reasons that have nothing to do with US bank approval. Those users want off-ramps from broken local banking systems. Some of the fastest-growing usage is, in a sense, the opposite of bank-driven growth — it is banklessness as a feature, not a bug. Which brings me to the part of the thesis that is wrong: the word "banks." The binary framing — banks or bust — conveniently ignores the middle path that regulation has already built. MiCA does not force stablecoin issuers to be banks. It requires an e-money license: a supervised, regulated non-bank entity with capital reserves, redemption duties, and audits. Licensed is not the same thing as bank. Several legislative models under discussion in the US similarly contemplate state-chartered or nonbank licensed issuers alongside full bank issuance. The ecosystem would do well to keep those two categories distinct in its mental models. I also think the rush to "banker knows best" creates a fragile narrative. The 2023 episode demonstrated that the banking layer itself can be the fragile component — lenders become chokepoints through liquidity events and reputational panics. And there is a governance question the narrative never mentions: what happens when the institution issuing, custodying, and auditing the reserves is the same entity? Conflict-of-interest risk expands precisely when the supervisor is also the competitor trying to displace existing issuers. The smart version of the thesis is shorter. Stablecoins need some supervised, audited, trustworthy infrastructure to reach institutional scale. What kind of institution owns that infrastructure is a separate question, and its answer determines whether the banking system ends up consuming stablecoins — or whether stablecoins simply become term deposits wearing modern clothes. Watch the signals, not the marketing. First: the GENIUS Act and related US legislation — if those bills pass with bank-charter mandates, the debate is settled. Second: whether the Fed explicitly demands a bank charter for issuers of dollar tokens. Third: whether bank-affiliated stablecoins capture even a meaningful share of the top-ten market. If they do, the current issuers will have become a regulated feature of the banking system, looking nothing like what founded them. The clock is running on that transformation. The next cycle's question is not whether banks will issue stablecoins — they are already piloting quietly. The real question is whether today's largest issuers become the banks' facility, and whether a banked stablecoin regime still leaves room for the permissionless rails that made this market matter in the first place. Trust no one, verify everything, move fast. And choose your bankers accordingly.