Last week I pulled the daily issuance data across the eleven largest dollar-pegged stablecoins. Two names — Tether and Circle — accounted for 85% of circulating supply. That reading sits within a hand's width of the highest concentration this category has recorded since it was invented. Nobody on the desk blinked. That is the interesting part. Markets price concentration only when it converts into a withdrawal event; until then, the same statistic reads as stability rather than fragility.
I have watched this movie before. In 2018, while peers chased ICO pumps, I spent the winter modeling vesting schedules and burn rates across fifteen emerging DeFi protocols. I found flawed tokenomics in three prominent names and predicted their dump cycles before the charts confirmed it. The lesson was never "be bearish." The lesson was that structural concentration stays invisible right up to the moment it becomes the only thing that matters. Stablecoins have now crossed that line, and the tape is treating it as background noise.
Trade the news, trade the reaction. The news is 85%. The reaction is still forming.
To understand why 85% is a load-bearing number rather than a trivia stat, you have to be precise about what stablecoins actually are inside the current architecture. They are not a product category. They are the settlement layer. Every perpetual swap, every DeFi borrow, every exchange withdrawal finalizes against a dollar token that lives on a chain and clears at the speed of a block confirmation.
The mechanics are deceptively simple. An issuer takes a dollar, mints a token, and promises redemption at par. The trust assumption is not cryptographic — it is custodial. There is no zero-knowledge proof verifying that reserves match liabilities. There is an attestation, a banking relationship, and a legal entity. That is the entire security model, and it has not materially changed in eight years.
This is where the sector's vocabulary has drifted away from its own engineering. "Decentralized finance" is, in practice, a settlement stack built almost entirely on centralized issuance. The base layer of DeFi — the layer every yield strategy ultimately depends on — is two private companies operating under two regulatory regimes, with two treasury desks and two redemption policies.
The concentration itself is old news. What is new is that it has climbed while everything around it matured. Rollups proliferated. Liquidity fragmented across dozens of venues. Bridge architecture was reinvented twice. Through all of it, the dollar token that everyone actually uses did not diversify at all. The infrastructure got more sophisticated; the money did not.
Let me put the number where it belongs. In antitrust analysis, a Herfindahl-Hirschman Index above 2,500 marks a "highly concentrated" market. The stablecoin market, on published supply data, sits well above that line. Two actors, one index, a structure that in any other financial sector would trigger immediate regulatory review.
The reflexive objection is always the same: "Concentration is fine because these are interchangeable dollars." That is wrong, and the reason is mechanical rather than ideological.
When you hold USDT you are not holding a dollar. You are holding a claim on Tether's balance sheet. When you hold USDC you are holding a claim on Circle's. Different claims, different reserve compositions, different banking rails, different redemption risk. The market prices them as equivalents only because, so far, redemption has cleared. That equivalence is a convention, not a property. Conventions are cheap to maintain in calm markets and brutally expensive to defend under stress.
Here is the transmission channel. A single issuer de-peg does not stay contained. Watch the sequencing. The token trades below par. Arbitrage desks that normally buy the discount and redeem at the issuer for the spread hesitate, because redemption now has a queue. The discount persists. Lending protocols using the token as collateral face a collateral-ratio cascade. Liquidation engines fire. The liquidations push the token further down. The venue that prices everything against that token — which is most of them — now operates on a broken reference rate.
That is the plumbing. None of it requires the issuer to be fraudulent. It only requires the issuer to be slow. Two redemption desks with banking hours and manual approvals, supporting a market that trades 24/7 at machine speed, is not a hedge. It is a latency mismatch dressed up as a stablecoin.
I have argued for years that oracle feed latency is DeFi's Achilles' heel. The concentration problem is the same disease with a different symptom. The risk is not the smart contract. The risk is the gap between what a protocol believes the price is and what the market actually clears at. When 85% of your settlement asset depends on two off-chain operations, you have built a global market on top of two part-time settlement engines.
The data-availability debate deserves the same cold look. The industry spent three years and a great deal of capital arguing about modular DA layers and rollup throughput. Meanwhile the actual bottleneck sat one layer below the smart contracts, inside two companies' treasury operations. Nobody shipped a research note about that. It was not novel. It was not investable. It was just true.
Let me quantify what "near historic highs" actually implies. Track concentration as a time series and the peaks line up with the crisis events. Concentration spikes when confidence in alternatives collapses and capital flees toward the two names everyone still trusts. That means the concentration metric is not only a risk reading — it is a fear reading. It rises precisely when the market is least able to absorb an issuance shock. The number is a thermometer that only runs hot once the patient already has a fever.
Liquidity dries up when fear sets in. Run the direction in reverse, and the same mechanism holds: when fear sets in, liquidity concentrates in the two venues everyone believes are safe. That is exactly where we are.
There is a governance point the price charts hide. A peer-to-peer currency that most users would describe as issueless — and functionally, in their heads, it is — is issued by two entities with unilateral mint and burn authority. I watched this dynamic up close during the DeFi yield trough of 2020. Liquidity does not equal value, and a treasury that can mint more of its own liabilities on demand is not a treasury in any conventional sense. It is a supply schedule with a brand name. When one entity controls that schedule, the "decentralized" governance story is decorative.
For holders, the practical implication is uncomfortable: when the two largest issuers can unilaterally alter the supply of the unit of account, position sizing stops being a market decision and becomes a governance decision.
The consensus contrarian take is that decentralized stablecoins will fill the gap. I don't buy the timing, and I don't buy the framing.
Decentralized alternatives hold roughly 15% combined. Every attempt to scale them has hit the same wall: capital efficiency. A fully collateralized or algorithmic dollar requires overcollateralization, which makes it expensive to mint and fragile under deleverage. It performs in a bull market and struggles in precisely the conditions where you need it most. That is not a technology gap. It is an economic one, and it does not close because someone ships a better curve.
The deeper contrarian point is that 85% may be structurally stable rather than precariously high. Settlement is a natural-monopoly business. Networks that clear value tend toward concentration because liquidity begets liquidity — the buyer goes where the seller already is. We accept this at the card-network level, at the clearinghouse level, at the correspondent-banking level. Crypto does not get exempted from the economics just because its ledger is decentralized.
So the correct question is not "how do we push concentration below 50%?" The correct question is: what oversight structure makes the incumbent issuers behave like regulated utilities rather than unregulated counterparties? That is a different roadmap. It requires attestation standards, enforceable redemption guarantees, and a resolution mechanism for the failure case. Boring infrastructure. Exactly the kind of work nobody writes a thesis about.
A word on the intent-based wave, since it keeps getting pitched as a fix for market-structure fragility. Routing orders through solver networks does not eliminate extraction; it relocates it. The MEV that used to be visible on-chain becomes a private auction among a handful of solvers. Nothing has been decentralized. The concentration simply moved from the mempool to a whitelist.
Watch two numbers, not the price. First, the combined share of the top two issuers: a sustained move below 80% would mean diversification pressure is real, and a move above 90% would mean the market has already selected its single point of failure. Second, decentralized stablecoin TVL as a share of total stablecoin supply: it has never cleared 20% in a durable window. Until it does, the decentralization narrative remains a marketing layer sitting on top of a two-company settlement system.
Everything else is noise. The reaction is what matters.