Twelve. Not twelve hundred — twelve.
According to ARK Invest research director Lorenzo Valente, that is roughly the number of stablecoins that have crossed the billion-dollar threshold since 2021, up from single digits four years ago. Above ten billion dollars, the field narrows to two names: Tether and Circle. In 2022, at the peak of that cycle, four stablecoins carried more than ten billion dollars in circulation. Today, only two remain. Valente's projection for the next two years is starker still: perhaps a single stablecoin exceeds five hundred billion dollars in market value, and it will not be a newcomer.
I read those numbers on a morning when the majors were printing green and a dozen new stablecoin whitepapers were circulating in my inbox. The contrast was the point. In a bull market, every category looks expansive. The data says the opposite about the asset class that underwrites all the others. Stablecoins are not fragmenting. They are consolidating, and the consolidation is not a technical accident.
The least glamorous, most load-bearing layer
Stablecoins are the quotation currency of crypto. They are the settlement leg in most DeFi operations, the working capital of market makers, the collateral that does not move, and increasingly the payment rail for cross-border transfers, remittances, and tokenized treasuries. When people say "on-chain dollars," this is what they mean: a fiat-denominated liability issued by a centralized entity, represented as a token across many chains.
Beneath the surface of that definition lies a structural fact that rarely makes it into a pitch deck. The technical barrier to issuing a stablecoin is close to zero. The trust barrier is close to absolute. Anyone can deploy an ERC-20 contract with a mint function and an attestation PDF. What almost nobody can do is be believed when they say the reserves exist, in the size and form required, redeemable on demand, in the jurisdiction that matters.
I want to be precise about what ARK is actually claiming. Valente describes "significant network effects." That is a market-structure observation dressed in technical clothing. Nothing about stablecoin cryptography explains why the count above ten billion dollars fell from four to two. Contracts are copyable in an afternoon. Trust, distribution, and permission are not.
Where the moat actually lives
If issuance technology is a commodity, why can a well-funded entrant not simply fork the code, hire the audits, and buy the market share? I have watched three teams try. The answer is that the moat sits in three places, none of them the smart contract.
First, liquidity depth. Market makers go where spreads are tightest and slippage is lowest, which is where volume already is. This is a reflexive loop, and it is expensive to enter from the outside — you must subsidize depth for long enough that someone else's order book becomes the marginal venue. Most treasuries cannot fund that war of attrition.
Second, integration surface. A stablecoin is not adopted; it is wired in. Trading pairs, collateral parameters, payment gateways, custody screens, OTC settlement desks, wallet defaults. Each integration is a contractual and operational commitment, not a click. When I led product for a privacy-focused payments startup in Berlin, I learned that the cost of replacing a settlement asset inside a live system is rarely the protocol — it is the reconciliation, the legal review, and the treasury policy that surround it. Switching costs in stablecoins are administrative before they are cryptographic.
Third, regulatory permission. This is the moat that has quietly become the tallest. Frameworks in the United States and the European Union now contemplate one-to-one high-quality reserve backing, segregation of assets, redemption guarantees, and periodic attestation. Every one of those requirements is a fixed cost, and fixed costs favor scale. A two-hundred-million-dollar issuer and a hundred-billion-dollar issuer pay for the same compliance apparatus. Only one of them can amortize it.
Stack those three moats and the twelve-name club stops looking mysterious. It starts looking inevitable.
The reserve-income economics nobody markets
There is a second structural reason the field is consolidating, and it is rarely stated plainly: stablecoin issuers are not, in economic terms, protocol companies. They are short-duration asset managers with a distribution network bolted on. Revenue comes from interest earned on reserves — predominantly short-dated government obligations and bank deposits — not from protocol inflation or token emissions.
That single fact reframes everything. Holders receive no share of reserve yield. Value accrues to the issuing entity or to its equity. The business is levered to the dollar rate environment, which means a falling-rate regime compresses issuer revenue without necessarily compressing issuer market share. Those two variables move independently, and conflating them is a common analytical error.
During the 2022 drawdown, I withdrew from public writing for six months and retreated to a cabin in Jutland, where I audited twelve failed lending contracts line by line. The common thread was not malice. It was over-leveraged design that had substituted speculative yield for real-world utility. Stablecoins are not that failure mode — they are not Ponzi structures — but they share one vulnerability with it: their economics are duration-sensitive, and their resilience is only tested when the duration turns.
So let me be blunt about the risk taxonomy, because it matters more than the market-cap chart. The failure mode for a stablecoin is not emissions collapse. It is reserve insufficiency, redemption pressure, and depegging — three events that arrive together and fast. Counting how many tokens exist above a threshold tells you nothing about how any of them would behave on the day the queue forms.
What I learned building a settlement rail I did not control
In 2018, at that Berlin payments company, my team integrated zero-knowledge proofs for transaction verification and refactored the consensus layer to cut gas costs by roughly forty percent while preserving anonymity. We shipped a beta to five thousand early adopters and the numbers held.
The technical success is not the lesson. The lesson is what we discovered about dependence. Settlement finality in our system rested on a stablecoin we did not issue, could not audit directly, and could not replace without renegotiating our entire downstream integration surface. We had built a privacy product on top of a trust assumption we did not control and had no leverage to enforce.
Trust is transitive. You can architect as decentralized a system as you like upstream; the moment the settlement leg is a custodial liability, your trust model is somebody else's balance sheet. This is not a criticism of stablecoin issuers — it is a description of the topology that every payment, DeFi, and treasury application inherits by default. Truth is not what is seen, but what is trusted — and most of the industry has outsourced the second half of that sentence to two companies.
The institutional translation problem
In 2024, following the approval of spot Bitcoin products in the United States, I joined a Nordic fintech firm to design a custody architecture for institutional clients that preserved non-custodial principles. The resistance was not technical. It came from executives who had spent careers in risk management and wanted answers to three questions: who can freeze the asset, who reconciles the ledger, and who is liable when the two disagree.
I ran twenty structured interviews with CTOs and translated cryptographic guarantees into the vocabulary of fiduciary duty. The architecture we proposed offered compliance reporting without exposing private keys and secured a pilot worth two million euros. What I took from it is that institutions do not buy decentralization; they buy auditable accountability with a decentralized option attached.
That demand profile concentrates the stablecoin market further. A regulated institution will not build on an offshore issuer with irregular attestation cadence, however competitive the yield. Circle's compliance posture and Tether's offshore heritage serve genuinely different geographies and different user bases, and I hold that view with moderate confidence — the public record on reserves is uneven, and each side has its own disclosure incentives. But the direction is clear: when regulation forces disclosure, it narrows the field. Compliance is a filter, and filters concentrate.
The AI-identity work I led in 2025 taught me the same governance lesson from a different direction. When we built reputation scoring into a decentralized identity protocol, we found that automated trust systems entrench whatever they measure unless humans intervene. We routed fifteen percent of reputation updates through manual review by a diverse community panel. Stablecoin trust is a reputation system too — one whose "proof" is an attestation and whose "review panel" is a handful of auditing firms. We have automated the scoring and outsourced the judgment.
The contrarian angle: concentration is not the same as stability
Here is where I part company with the triumphant reading of Valente's numbers. The consolidation is being narrated as validation — as evidence that the market has matured and the winners have earned their position. I think that framing mistakes a competitive outcome for a systemic property.
A two-name club means that an enormous share of on-chain notional exposure settles through two balance sheets, two legal structures, two attestation cadences, and two freeze authorities. That is not a mature market. That is a single point of failure with a market-cap chart drawn over it. The industry that spent a decade preaching the elimination of counterparty risk has quietly rebuilt its most critical dependency on two counterparties — and then called the dependency "network effects."
The metric itself is the second problem. Counting how many stablecoins exceed a given size threshold measures distribution, not resilience. The question that matters is narrower and more uncomfortable: who holds the mint keys, and who holds the freeze function? Those authorities are the actual architecture. A hundred billion dollars of "decentralized finance" that can be frozen by a compliance officer is not a governance model — it is a service-level agreement.
And there is a third blind spot, which I think is the most important one. The next serious challenger to this duopoly almost certainly will not appear in the stablecoin category at all. It will arrive as a tokenized money market fund, a yield-bearing deposit token, or a bank-issued settlement instrument — assets that distribute reserve income to holders, which is precisely what the incumbents refuse to do. Those instruments escape the current data taxonomy, so they will be invisible in exactly the reports that are being cited as proof of stability. By the time they show up in the count, the count will already be wrong.
Takeaway
The question worth watching over the next two years is not which stablecoin reaches five hundred billion dollars. It is who is permitted to mint, under what disclosure regime, and who bears the loss when a redemption queue forms on a Sunday.
The number to track is not twelve. It is two — and then to ask what happens to the rest of us when two becomes one. Truth is not what is seen, but what is trusted. In a bull market, the industry counts the tokens. It should be counting the signatures.
