5. Tokenised money-market funds functioning as cash equivalents
The third flavour is the one that gets miscategorised most often, and the miscategorisation has real consequences. Tokenised money-market funds (MMFs) like BUIDL (BlackRock USD Institutional Digital Liquidity Fund issued through Securitize), FOBXX (Franklin Templeton's OnChain US Government Money Fund), and OUSG (Ondo Finance's wrapper of short Treasury exposure) are not stablecoins. They are tokenised fund interests in a money-market fund or similar vehicle.
The distinction is structural. A payment stablecoin is a liability of the issuer, redeemable at par, with reserves held to back the float. A tokenised MMF is a fund interest. The holder owns a share of the fund. The net asset value (NAV) moves, even if it moves slowly and predictably. The yield accrues to the holder, either through a rebasing token design or a price-accruing design or via separate distribution. The fund is regulated as a fund: in the US, the named examples split between two structures. BUIDL is structured under section 3(c)(7) of the Investment Company Act, restricted to qualified purchasers. FOBXX is a 1940 Act registered government money-market fund. OUSG is also a 3(c)(7) private fund. In all cases the apparatus is full: a transfer agent, an administrator, an auditor, an independent board.
Why does the distinction matter? Several reasons that show up in real workflows.
First, eligibility for institutional treasury allocation. A corporate treasurer can usually allocate to a money-market fund under their investment policy. A payment stablecoin sits outside that policy in most cases, because it is not a fund, not a deposit, and not a registered security. A tokenised MMF slots into the existing policy line for cash and cash equivalents. A payment stablecoin needs a separate policy decision.
Second, who can hold it. Tokenised MMFs are typically restricted to qualified purchasers or qualified institutional buyers, with KYC enforced at the token level via permissioned transfer logic. Payment stablecoins are bearer-style, transferable to any wallet that is not on a blocklist. Opposite transferability profiles.
Third, the regulator. For a tokenised MMF, the SEC and the fund's other regulators are in the room. For a payment stablecoin, you are talking to whichever payment, banking, or specialised stablecoin regulator authorised the issuer.
Fourth, yield treatment. Tokenised MMFs distribute yield to holders. Payment stablecoins under most regimes are prohibited or constrained from paying interest tied to the holding period. The US GENIUS Act flatly prohibits interest payments to holders. MiCA prohibits e-money token (EMT) issuers from granting interest related to the length of time the holder holds the EMT, which is the same thing in substance. The deliberate design choice across regimes is to keep payment stablecoins from competing with bank deposits and to keep the perimeter narrow.
Fifth, role in a portfolio. A tokenised MMF is a treasury instrument: where you park cash to earn the front-end yield. A payment stablecoin is a settlement and movement instrument. They are complements, not substitutes.
When you hear someone describe BUIDL as a "yield-bearing stablecoin," translate: they mean a tokenised money-market fund, regulated as a fund, that happens to be transferable on-chain among permissioned holders. Same chart shape near par. Different instrument, different regime, different role.
6. Algorithmic constructions
The fourth flavour is algorithmic stablecoins, where the peg is maintained without full off-chain reserves, typically through a seigniorage or dual-token mechanism. The audience knows the history. Terra/UST is the cautionary tale, and the post-mortem is well covered. Smaller experiments continue, often relabelled as overcollateralised crypto-backed stablecoins to dodge the algorithmic label, which is a reasonable distinction when the collateralisation is genuinely overcollateralised in liquid on-chain assets.
The relevant point for a TradFi reader is that the major regulatory regimes have either explicitly prohibited or effectively foreclosed pure algorithmic constructions. MiCA's reserve and redemption requirements cannot be met by a seigniorage design. The GENIUS framework restricts permitted issuers to those holding cash and short-dated Treasuries. The Hong Kong regime requires high-quality liquid asset reserves at least equal to par. So whatever the on-chain experimentation continues to look like, the production institutional surface for algorithmic stablecoins is essentially zero.