For five years, "stablecoin" meant one thing. It was a dollar on a blockchain that you didn't have to think about. In 2026 that sentence stopped being true, and the panda noticed nobody filed a press release about it.
The thesis: the onchain dollar category has quietly split into three markets that look identical to retail but serve completely different buyers. The first persona is the payment dollar, owned by USDT and USDC, and it still grows. The second is the treasury dollar, where BlackRock's BUIDL and its rivals tokenize T-bills onto chains that didn't exist three years ago. The third is the yield-bearing dollar, where Ethena's USDe and similar products pay you a coupon by running a structured product under the hood. According to the CoinGecko global market dashboard, the total crypto market cap sits at $2.20T as of June 10, 2026, with USDT at $186.82B in market cap and USDC at $75.01B. That's the payment dollar. The other two markets, treasury and yield, now add roughly another $30-40B between them, and they're not stealing share from USDT. They're stealing share from each other.
What is an onchain dollar in 2026, exactly?
The honest answer is that nobody on the protocol side wants to define it cleanly, because clean definitions invite regulators. So let's do it cleanly anyway.
A payment dollar is a 1:1 fiat-backed token that does not pay you yield, that issues against bank deposits and short-duration paper, and whose entire reason to exist is to move money across rails that don't ask permission. USDT and USDC are this. Per CoinGecko's Tether page, USDT trades at $0.999278 with a $186.82B market cap on June 10, 2026. The product is settlement. The yield it generates internally goes to Tether and Circle. You hold it to send it, not to earn from it.
A treasury dollar is a tokenized representation of a money market fund or a basket of T-bills. The yield is the T-bill yield, minus a manager fee. The reserves are auditable and held by a regulated custodian. BlackRock's BUIDL, Franklin Templeton's BENJI, Ondo's USDY all live here. According to the DefiLlama RWA dashboard, tokenized US Treasuries on public chains crossed $14B in onchain notional during Q2 2026, with growth concentrated on Ethereum mainnet. The buyer here is institutional: a DAO treasury parking idle stables, a market maker collateralizing perps with a yield-bearing asset, a fintech routing client cash through a SEC-recognized wrapper.
A yield-bearing dollar is a stablecoin that pays you yield from somewhere that isn't a T-bill. Ethena's USDe pays the funding rate of perpetual futures via a delta-neutral short. Frax sFRAX pays a blended algorithmic rate. Maker's sUSDS pays from DAI's surplus. The yield can be higher than treasuries in a bull market, can flip negative in a flush, and the regulatory category is, generously, "we'll see." That uncertainty is not a bug. It's the whole moat. We covered the latest stress test on that thesis in Ethena's USDe and the funding rate stress test.
How three markets pretend to be one
The reason this fork doesn't get covered is that the user experience looks identical. You hold a token. The token says "1 USD." You see a balance in dollars. The fact that one of those three tokens is a money market fund share, another is a delta-neutral structured product, and the third is just digital cash is invisible at the wallet layer. Spoiler: we saw this one coming.
The buyers are not the same person.
USDT and USDC are bought by anyone who needs to move dollars across borders, fund a perp account, or settle a trade. The marginal buyer is global retail. The depth is in CEX order books, not DeFi pools. According to CoinGecko's USDC page, USDC sits at $75.01B in market cap with negligible 24h price drift, which is exactly the boring stability profile a payment dollar should have.
BUIDL is bought by institutions that need the SEC-recognized treasury exposure, the 1940 Act fund wrapper, and the auditable reserves. Retail can't even buy it directly. The marginal buyer is a CFO parking working capital, or a tokenization platform that needs T-bill collateral to back a structured offering. The growth path runs through Wall Street rails being plugged into Ethereum, not through CEX listings.
USDe is bought by yield-hungry DeFi users who want a synthetic dollar that pays roughly the funding rate of perpetual futures. The marginal buyer is a yield farmer rotating between Pendle PT markets and Aave loops. The depth is in DeFi pools, not on exchange order books. The product is fragile in flush regimes and dominant in greed regimes. That fragility is why it sits in a different bucket than the first two.
These three users overlap less than crypto Twitter assumes. A CFO won't touch USDe. A yield farmer won't bother with BUIDL's gated KYC. A remittance corridor won't switch from USDT to anything cheaper if the cheaper option needs an attestation flow. The category is forked because the buyers were always different. The 2024-2025 framing of "stablecoin growth" lumped them together, and now the numbers force the split.
Why does this matter for DeFi yields and ETF treasuries?
Two reasons, both structural.
First, the treasury dollar is sucking yield-sensitive capital out of DeFi pools. Until 2025, the easiest way for a DAO or a market maker to get 4-5% on idle stables was to lend on Aave or farm Pendle PTs. With BUIDL or USYC paying a comparable yield with an SEC-blessed wrapper and zero smart contract risk, the operational answer flipped. DAOs are now defaulting to treasury dollars for the cash-management bucket, and only deploying to DeFi for the alpha bucket. According to the DefiLlama yields page, the median stablecoin DeFi yield in mid-2026 has compressed toward 3-5%, almost exactly the T-bill yield minus an operational risk premium. That's not a coincidence. That's the treasury dollar acting as a yield floor.
Second, the yield-bearing dollar is now a regulated-product question. The SEC and CFTC have spent eighteen months arguing over whether sUSDe is a security, a swap, a commodity, or a registered fund. The European MiCA framework already exempts tokenized funds and stops short of stating clearly that delta-neutral synthetic dollars are something else. The result is regulatory arbitrage by geography: yield-bearing dollars dominate in jurisdictions that don't ask, and underperform in jurisdictions that do. Spoiler: that's not a stable equilibrium.
For context on how the payment dollar bucket itself stratified, see our earlier piece on why USDT structurally won the stablecoin wars. The same forces apply at a different layer of the stack.
| Persona | Example tokens | Primary buyer | Yield source | Regulatory category |
|---|---|---|---|---|
| Payment dollar | USDT, USDC | Global retail, CEX traders | None (goes to issuer) | Money transmitter |
| Treasury dollar | BUIDL, BENJI, USYC | DAOs, CFOs, institutions | US Treasury bills | 1940 Act fund |
| Yield-bearing dollar | USDe, sFRAX, sUSDS | DeFi yield farmers | Funding rates, on-chain spreads | Unsettled |
Counterarguments
Two real objections deserve fair treatment.
The first is "this is one market, not three, because users will treat them as interchangeable and arbitrage them at the margin." There is truth here. USDe trades against USDC on Uniswap, BUIDL can be redeemed to USDC by qualified buyers, and Pendle markets price the implied yield differential between sUSDe and treasury wrappers. The arbitrage layer exists. But arbitrage compresses prices, not categories. You can arbitrage the spread between a money market fund and a checking account too. Nobody calls them the same product. The category boundary is in the risk profile and the buyer base, not in the spot price.
The second is "tokenized treasuries are just a phase, killed the moment Fed funds drop below 2%." Possible. In a zero-rate environment, the treasury dollar loses its yield premium, retail rotates back to DeFi for the alpha, and USDe-style products dominate because funding rates stay positive in greed regimes. The panda concedes this scenario, with one caveat: even with no yield premium, the institutional buyer of BUIDL doesn't disappear, because the SEC wrapper itself is the product. A CFO still wants the auditable, 1940 Act-compliant version of cash over the smart-contract version, regardless of the absolute yield. The Q2 2026 BUIDL growth happened while the Fed was already on a pause path, which tells you the buyer cares about more than yield.
A third minor objection: "this is just RWA, you've renamed it." Not quite. RWA is a basket that includes private credit, tokenized real estate, tokenized commodities, all of which have nothing to do with the dollar question. The treasury-dollar bucket is the subset of RWA that competes with USDT and USDC for the same wallet slot. The rest of RWA is solving different problems for different buyers.
What to watch next
Three datable signals between June and December 2026.
By August 31, 2026: watch whether the total tokenized US Treasury notional crosses $20B on the DefiLlama RWA tracker. If yes, the treasury-dollar bucket has graduated from experiment to institutional default, and the DeFi yield floor will stay around T-bill minus 100 bps. If no, the thesis softens and the segment stays a niche allocation.
By October 31, 2026: watch whether the SEC issues any formal guidance on yield-bearing stable products. The current docket has at least three pending no-action requests that, depending on outcome, will either legitimize sUSDe and sFRAX onshore, or push them offshore. Either path forks the on-ramp market further.
By December 31, 2026: watch whether USDC's market cap recovers above $90B. If Circle's payment-dollar dominance rebounds, the three-market thesis stays clean. If USDC keeps bleeding share to USDT and the yield-bearing bucket, the payment-dollar segment may itself be forking again, which would be a story for 2027.
Outside the dashboard, the more interesting signal is whether any major chain treasury starts denominating its operating budget in BUIDL or USYC rather than USDC. That would be the unambiguous tell that the treasury dollar has won the institutional cash-management slot. We saw similar yield-stack dynamics in Pendle's PT markets becoming the new DeFi collateral layer, and the lesson there was the same: when a yield primitive wins, it becomes infrastructure, not narrative.
BSC sits in an awkward spot in this fork. Per the DefiLlama BSC chain page, BSC TVL stands at $5.14B as of June 10, 2026, down 4.91% week over week. None of the three onchain-dollar buckets has consolidated on BSC the way they have on Ethereum. The payment dollar (USDT) is there at scale, the treasury dollar (BUIDL, USYC) is not, the yield-bearing dollar (USDe, sFRAX) has thin liquidity. For a chain that defines itself by retail throughput and cheap fees, missing the treasury-dollar wave is the larger structural risk. The full topic cluster on this is at our stablecoins pillar. The panda has no prediction for the BSC catch-up timeline. The point is just that when categories fork, the chains that host all three buckets accrue authority, and the chains that host only one slowly become single-purpose rails. That is a thesis worth watching, and the kind of structural question this Dadacoin blog keeps poking at while everyone else writes the same daily-pulse take.
