apxUSD and the Bitcoin-treasury collateral chain

apxUSD's reserve is reported through Apyx's Accountable dashboard as a collateral mix, and the live split is lopsided. Around 62% sits in STRC, the preferred equity of Strategy (formerly MicroStrategy), with a negligible fraction in Strive's SATA preferred and the rest — about 38% — in cash, short-term U.S. Treasury Bills, and USDC held for liquidity. Redemptions settle in USDC, and under stress the protocol sells preferred-share positions into USDC to meet them. A Protocol-Owned-Liquidity buffer is meant to smooth the secondary market during dislocations.

Preferred equity in a Digital Asset Treasury company is not a Treasury bill. STRC is a dividend-bearing claim on Strategy, a firm whose principal asset is Bitcoin; its value depends on Strategy's ability to service the preferred and on the market's appetite for leveraged Bitcoin exposure. That is a real instrument with a real yield, but it is a corporate-credit-plus-equity claim, and its price behaviour tracks crypto risk far more closely than it tracks the front of the Treasury curve. Calling the resulting dollar 'RWA-backed' is accurate about the asset class and silent about the correlation.

A stablecoin's reserve does its job when it holds value precisely in the moments the peg is tested. T-Bill-backed dollars work because the collateral is, to a first approximation, uncorrelated with the crypto markets where the token trades — when crypto sells off, the bills are still worth par. apxUSD inverts that property. Its dominant leg is a Bitcoin-correlated preferred share, so a crypto drawdown weakens the reserve at the same instant redemption pressure rises. The collateral and the stress share a common factor, and a common factor is exactly what a stablecoin reserve is supposed to avoid.

The buffer could not cover the gap. With roughly 38% in cash and Treasuries against roughly 62% in the volatile leg, even a moderate markdown on STRC pushes net collateral toward — or through — the value of apxUSD outstanding before the buffer is exhausted. The cash sleeve is enough to clear an orderly trickle of redemptions; it is not enough to defend par when the majority of the reserve is falling in lockstep with the broader market. The June low near $0.89 is what that arithmetic looks like under stress.

This is the structural cousin of pmUSD's gold-RWA fragility: in both cases the headline is a real-world asset, and in both cases the peg depends on a layer whose value the token holder cannot independently pin down or convert on demand. apxUSD's chain runs apxUSD → preferred shares → a Bitcoin-treasury company → Bitcoin, and a shock anywhere along it propagates straight up to the token.

A healthy dollar holds par because arbitrage is cheap and symmetric: below $1, someone redeems for a dollar of value; above $1, someone mints and sells. apxUSD weakens both legs. Minting is an attested-minter flow — an OpenZeppelin AccessManager gates a Gnosis Safe (3-of-6) and a backend-signed MinterV0 contract, with an externally-owned account as admin — so supply is created by a small set of controllers on signed orders, not by open arbitrage. Redemption is whitelisted: only approved participants mint and redeem against the protocol directly. Retail holders do not have that door.

For everyone else, the exit is the secondary market, and apxUSD's secondary market is a set of Curve pools that the March upside spike already showed to be thin. When the whitelisted rebalancers cannot or will not buy the discount back to par — because doing so means warehousing a falling, Bitcoin-correlated reserve — the price discovers itself against whatever depth the pools provide. That is why the discount has persisted for weeks rather than snapping back: the participants who can redeem at par are not the participants holding the token, and the ones holding it can only sell into a shallow book.

Transparency around the reserve is real-time but lightly attested. Apyx publishes a live Accountable feed, but Pharos records the proof-of-reserve as coming from a niche attestor with no monthly attestation, and the GENIUS-compliance review returns no public authorization, a Terms of Service that disclaims the customary legal meaning of its own financial terminology, and a BVI domicile that blocks U.S. persons. None of that caused the depeg, but it shapes what a holder can enforce when the price is below par.

The signal was structural before it was a price. Pharos scores apxUSD's collateral quality as RWA with a single-entity governance posture, flags the STRC and SATA legs as high-risk in the live reserve map, and routes the asset through the RWA-credit-fund mechanism rather than treating it as a cash-backed dollar. Those classifications are why a token that spent its first months near par still carried structural caution rather than a clean bill of health.

When the incident opened, the live surfaces moved together. The depeg resolver flagged the June 2 downside event and has kept it open with no recovery print; DEWS sits at WATCH (29/100) and Safety at D (49/100); and the supply trend turned negative — down about 8.6% over thirty days — as redemptions and exits shrank the float to roughly 407.5M against a $392M market cap. The peg and liquidity sub-scores (65 and 60 across 17 pools) capture a token that can still clear sellers, but only at a standing discount. apxUSD also appears inside other protocols' reserves as Pendle PT collateral, so its stress is a dependency worth tracing on the map, not only a price on its own page.

The asset-class label is not the risk profile. 'RWA-backed' tells you what the reserve is made of; it does not tell you whether the reserve holds value when the peg is tested. The question that matters is correlation: a dollar backed by an instrument that falls with crypto carries the same factor risk as the market it trades in, and it will weaken exactly when holders most want par. A buffer sized below the volatile leg cannot change that — it can delay the discount, not prevent it.

Redemption access is part of the peg. When direct mint-and-redeem is whitelisted and retail's only exit is a thin secondary market, the arbitrage that normally pins par is gated to a few participants whose incentive to defend the floor evaporates precisely when the reserve is falling. And when an issuer describes a depeg as expected behaviour rather than a break, take the statement literally: it is telling you the design permits the price to move with its collateral. apxUSD remains live, but its still-open June incident is a clean illustration of what happens when a stablecoin's reserve and its stress share a common cause.