Skip to content
PROTOCOL RESEARCH

Apyx risk

apxUSD has recovered to $0.9177 because STRC rallied 24%, not because anything was repaired: attested collateral is still 92.24%, there is still no atomic redemption, and the dollar-side exit behind a float DeFiLlama marks near $298M is about $1.96M of USDC across every apxUSD/USDC pool plus roughly $253K of Kraken bids whose best bid is $0.83. STRC concentration did not grow, contrary to the draft: it is flat at about 84% of asset reserves and down from roughly 74% to 70% on the wider base.
AvoidResearched Aug 5, 2026

Scorecard

watch: Audits & contractsFour firms verified on the official audits page: Halborn 2026-06, Quantstamp 2026-04, Certora 2026-03, Zellic 2026-03, Quantstamp 2026-02. Public Foundry repo with invariant tests and Slither CI. No bug bounty on any major platform, and ApxUSD.mint() takes no collateral on-chain.
weak: Admin controlGuardian Safe took its Curve LP from 40,890,164 to zero by July 6, removing the permissionless exit mid-drawdown. Verified on-chain today: Admin Safe 0xabdd8c8e is 4-of-6 holding ADMIN_ROLE at zero delay; the 3-of-6 Guardian holds UPGRADER at 3 days. Zero-delay deny list, custody offshore at Alpaca under a BVI entity.
weak: OracleDowngraded on verified evidence the draft missed: ApxUSDRateOracle is manually set via setRate() by ADMIN_ROLE with zero execution delay, no feed, no TWAP and no staleness check, reads 1.000000 today, and prices the Curve pool. Accountable never discloses how STRC and SATA are marked while Nasdaq is closed.
weak: Liquidity & exitNo atomic redemption; direct redemption permissioned at $0.9131 rather than par; apyUSD cooldown 20 days. Dollar exit verified on-chain totals about $1.96M of USDC across all apxUSD/USDC pools, plus about $253K of Kraken bid depth where the best bid is $0.83, under 1% of the float.
watch: Yield (real vs emission)Real cash, not emissions: Strategy raised STRC to a 12.00% annual rate for record dates from July 1, paid as $0.50 semi-monthly. But it funds roughly $1.76B of yearly obligations partly by selling bitcoin, 3,588 BTC for $218.5M in early July and about $105M more in late July.
weak: Holder concentrationCorrected: concentration did not grow. STRC is 84.4% of asset reserves against Forbes' 84% on July 21, and 70.0% on the POL-inclusive base against roughly 74% in July. Flat to slightly lower, and dangerous at either number. apyUSD vault holds 59.3% of Ethereum supply.
weak: Track recordLaunched February 18, 2026, broke peg June 4 in month four, and has closed below par for 62 straight days, 36.9% of the protocol's entire life. The attested ratio has never regained 100%. Supply is down 38.2% from a 524.66M peak.
🟢 strong🟡 watch / caveat🔴 weak / fund-loss risk
Verdict is a gate (worst flaw wins), not an average. Our read, not financial advice.

TL;DR

Apyx issues apxUSD, a synthetic dollar, and apyUSD, its yield-bearing ERC-4626 sibling, across Ethereum, Base and BNB Chain. Both are claims on one off-chain portfolio that is mostly Strategy's STRC perpetual preferred, held at Alpaca brokerage under Preference Foundation, with Apyx itself incorporated in the British Virgin Islands. apxUSD broke peg on June 4, 2026, in month four of the product's life. Sixty-two days later it has not returned to par.

The price genuinely recovered: $0.7176 at the June 26 low, $0.868 when this report was last written on July 2, $0.9177 today, up about 5.7% over the month. This refresh existed to find out whether anything real is behind that. One thing is, and it is not Apyx. STRC rose from $74.57 on June 26 to $92.53 on August 4, up 24.1%, and apxUSD rose 27.9% off its own low over the same window. The token tracked its portfolio. That is what a correctly priced claim on a damaged reserve does, and it is worth being precise about what it means: apxUSD has stopped trading at a discount to its collateral and started trading at its collateral's value. The market price now sits a hair above the protocol's own attested redemption value of $0.9131.

What did not happen is the repair. The attested collateral ratio is 92.24%, still short of par and never once above it since June. There is still no contract that swaps one apxUSD for a dollar. The protocol finished withdrawing its own Curve liquidity on July 6. And the venue that will actually take size, Kraken, is bidding $0.83.

Two things in the previous draft of this refresh were wrong and are corrected here, because both cut against the report's own thesis. First, STRC concentration did not grow. The 84.4% figure is measured against asset reserves; July's roughly 74% was measured against a base that also counted protocol-owned liquidity. On that same wider base the number today is 70.0%. On the narrower base Forbes independently put it at 84% on July 21 against 84.4% now. Flat to slightly lower, either way, not rising. Second, the admin Safe threshold is not contested. Both readings were right about different Safes, verified on-chain today.

One warning carried forward, because it still protects people: X accounts promoting a "Repeg and Compensation Plan" governance vote, including @apyx_FD and @apyxfnd_, are impersonator scams. The official handle is @apyx_fi, Apyx's own post-mortem announces no such vote, and "compensation plus priority recovery access" links after a depeg are a standard wallet-drainer pattern. Do not connect a wallet to any compensation vote.

Checklist

Audits & contracts. Every named firm was checked against Apyx's own audits page and all five engagements are listed there: Halborn 2026-06, Quantstamp 2026-04, Certora 2026-03, Zellic 2026-03 and Quantstamp 2026-02. No invented firms. Worth noting that Yearn's otherwise thorough August assessment does not mention Halborn at all, so the post-depeg engagement rests on Apyx's own disclosure rather than on independent confirmation. The code is public at apyx-labs/evm-contracts with an invariant test suite, audit-remediation tests and Slither in CI, which is better hygiene than most protocols this young. There is still no bug bounty; an exhaustive search across Immunefi, Sherlock, Cantina, HackerOne and Safe Harbor found none. One design fact outweighs any audit finding here: ApxUSD.mint() checks the caller's role and the supply cap, then mints. No transferFrom, no collateral deposit, no on-chain proof of backing at mint time. The contracts held in June because the contracts were never the thing at risk.

Admin control. Red, on a power that was actually exercised. The Guardian Safe took its Curve LP position from 40,890,164 to zero: about 88% between June 1 and June 5 while the token was falling, the remainder by July 6 at block 25474518. The protocol removed the permissionless exit during the exact drawdown that exit existed for. The Safe structure was re-read on-chain this run and the previous ambiguity is resolved. There are two Safes, not one disputed number. The current Admin Safe at 0xabdd8c8ee69e5f5180eb9352aeffc5ceead65e96 returns a threshold of 4 across six owners and holds ADMIN_ROLE with zero execution delay, granted March 20, 2026. The former Admin Safe at 0xf9862EfC1704aC05e687f66E5cD8c130E5663cE2 returns a threshold of 3 across six owners and now holds UPGRADER with a three-day delay, PAUSER at zero delay, UNPAUSER at four hours and YIELD_OPERATOR at zero delay. Our July on-chain read and Yearn's August read were both correct about different contracts. Alongside that sits a shared deny list at 0x2c271ddF484aC0386d216eB7eB9Ff02D4Dc0F6AA, wired into apxUSD, apyUSD and the unlock token, where a Safe can deny an address with zero execution delay while removal carries four hours. It was in active use at the last count, 141 addresses added and 137 still denied on 2026-07-27. No admin function can move or burn a frozen balance, so this freezes rather than seizes. Custody remains off-chain at Alpaca in the name of an offshore foundation, and only DFDV's CEO is publicly named.

Oracle. Downgraded to red this run on evidence the previous draft did not have. The rate oracle that prices the Curve pool, ApxUSDRateOracle at 0xa2ef2e7bf32248083e514a737259f3785ea8d37d, is manually set through setRate(). It has no price feed, no TWAP and no staleness check. It reads exactly 1.000000 on-chain today. The permission on that setter is role 0, ADMIN_ROLE, and the target admin delay is zero, which means the 4-of-6 Admin Safe can change the number that prices the main pool instantly, with no timelock and nothing to observe it going stale. A rate of 1.0 is normal for a stablecoin pool invariant and is not itself the defect. The defect is that it is a hand-set input on a fund-loss surface with no delay in front of it. On the credit side, the apyUSD/apxUSD Morpho market keys off the redemption rate rather than spot, and it again booked no bad debt through the whole event; that design decision has now been tested twice and held. Other Morpho markets ran a stale self-managed spot oracle during the depeg with a Chainlink migration described as in progress, and completion remains unconfirmed. The other hole is valuation. Accountable's feed reaches enclave-level verification with Nitro attestation material and sits at verifiability level 4 on Accountable's own five-level scale, where level 5 is zkTLS, but it publishes no source-level timestamps and never says whether STRC and SATA are marked at last trade, a broker quote, a modeled fair value or a bid-side liquidation mark when Nasdaq is shut. Nasdaq being shut is precisely when the June wicks printed. Apyx's own dashboard has already shown an inflated NAV once, from a STRCx pricing bug, while users transacted against it.

Liquidity & exit. Still the defect that sets the verdict, and the previous draft understated it in one direction and overstated it in another. Both are corrected here from direct on-chain reads. No contract swaps one apxUSD for a dollar of collateral, so no arbitrage forces the price to par. Direct redemption is permissioned, gated to approved counterparties, and quoted at redemption value rather than $1: $0.9131 on August 3. apyUSD exit adds a cooldown of 1,728,000 seconds, exactly 20 days, at a 0.1% fee, with no yield accruing while you wait. On depth, the claim that the Curve pool was drained is true of the old pool and only the old pool. 0xE1B96555 holds about $11.0K, confirmed on-chain. A newer Curve apxUSD/USDC pool at 0x6F63deEDc9870D6c16FC644C6654748352cdc87c holds 7,418,710 apxUSD against 980,037 USDC, about $7.79M nominal. That pool does not appear in Curve's factory-stable-ng registry, which is why one reviewer missed it; it was confirmed here through DexScreener and then by reading the token balances directly. But the newer pool is not the whole exit either, and counting only it understates the total. Adding every apxUSD/USDC venue gives about $1.96M of USDC: $980.0K in the new Curve pool, $780.1K and $192.7K in two Uniswap V4 pools, $1.7K in PancakeSwap V3 and $1.4K in the old Curve pool. The sharpest number is one no previous version of this report contained. apxUSD is listed on Kraken as APXUSD/USD and APXUSD/USDC, and the book there is bidding $0.83, roughly 9.5% below the $0.9177 DEX mark and below the $0.9131 attested redemption value, with about $253K of total bid-side depth across both pairs. The only venue that looks like a real exit prices the token materially worse than the screen does. Total dollar-side exit is therefore about $2.2M against a float DeFiLlama values near $298M, under 1%. Volume across 37 CoinGecko tickers on 8 venues was about $1.15M in 24 hours, roughly 0.4% of market cap. The $18.8M apyUSD/apxUSD Curve pool does not help, since it moves you between two claims on the same reserve, and the sUSDat pools pair apxUSD against another DAT-backed synthetic with correlated risk. One inference, flagged as inference: Accountable reports $44.8M of protocol-owned liquidity, and total observed DEX liquidity across all 19 apxUSD pairs is about $37.8M. Those are the same order of magnitude, which suggests most of the visible depth is the protocol's own and can be withdrawn at will, exactly as it was in June and July.

Yield: real vs emission. The yield is real cash from a dividend, not token emissions, and the rate is now pinned precisely. Strategy's June 29 Digital Credit Capital Framework raised the STRC annual dividend rate to 12.00% for record dates on or after July 1, 2026, and moved to semi-monthly payment; the tape confirms it, with $0.479 on June 30 and $0.50 on both July 15 and July 31 against monthly $0.958 through June 15. Twelve percent on a $100 par instrument is a junk-bond coupon, and it is paid because the instrument is distressed rather than because the issuer is generous. Strategy's preferred dividends and interest run about $1.76B a year, a figure that falls straight out of its own disclosure that a $2.55B reserve covers 17.4 months. It is funding that partly by selling bitcoin: 3,588 BTC for $218.5M between June 29 and July 5, and about $105M more in the week before August 3, alongside an $8.2B second-quarter loss on bitcoin's decline. The never-sell line is gone. Separately, there is an emission on the demand side even though there is none on the yield side: the APYX Pips program, Season 2 running May 23 to October 11, 2026, pays users to leave capital in place. Some of the supply that stayed through the depeg was being paid to stay.

Holder concentration. The orchestrating question for this refresh was whether the STRC concentration came down, and the honest answer is that it is flat, not rising. The previous draft got this backwards by comparing two different denominators. On the August 3 Accountable snapshot, asset reserves of $217.04M were $183.29M STRC, $33.74M cash and equivalents, and $6.8K other, putting STRC at 84.4% of the base the collateral ratio is measured against. Forbes, using Artemis data on the same basis, put it at 84% on July 21. That is flat. Counting protocol-owned liquidity in the base, which is the denominator behind July's roughly 74% figure, today gives 70.0% STRC, 12.9% cash and 17.1% POL. That is slightly down. Neither number describes a diversified reserve and neither supports a claim that the exposure grew. On the token side, the apyUSD vault holds 184,994,984 apxUSD against an Ethereum supply of 312,073,515, which is 59.3%, read on-chain today. The fullest public holder reconciliation beyond that is explicitly incomplete. One more concentration worth naming: an Apyx operations Safe at 0x37B0779A66edc491df83e59a56D485835323a555 holds 582,774 STRCX, around 35% of all STRCX in existence and the only directly on-chain-verifiable slice of the backing.

Track record. Launched February 18, 2026. Broke peg June 4, 2026, in month four. Below par every day since, 62 days as of this refresh, which is 36.9% of the protocol's total life. The attested collateral ratio has never regained 100%: roughly 84% in late June, 90.7% on July 21, 92.24% on August 3. The trend is genuinely up and that deserves acknowledgment; a ratio climbing toward par is still not a ratio at par. Supply has fallen from a peak of 524.66M tokens on May 28 to 324.37M, down 38.2%, executed through on-chain burn and burnFrom calls. That is two things at once: a real transparency improvement, since redemptions used to settle entirely off-chain and are now observable, and a picture of holders leaving in batches. Things that held should be said plainly. Redemptions were processed proportionally across the basket rather than draining the liquid leg first, the apyUSD unlock window prevented a run, the redemption-rate Morpho market took no bad debt, and the post-mortem admitted the NAV bug rather than burying it. Wolf & Company's monthly attestations are real and current through June, opinion dated July 22. That is partial credit on a failed record, not a pass.

Worst case

Bitcoin takes another leg down, MSTR compresses, and STRC hands back the 24% it just recovered. A reserve that is 84% STRC marks down with it and the collateral ratio, currently 92.24%, falls back through the June lows. No atomic redemption exists to arbitrage that gap closed. Price discovery then happens across roughly $1.96M of USDC and a Kraken book already bidding $0.83, where a few hundred thousand dollars of selling moves the mark, while approved counterparties redeem at redemption value and everyone else queues or sells into that book. If the protocol pulls its liquidity again, and reported POL is the same order of magnitude as all visible DEX depth, the exit thins further at exactly the wrong moment. The looping layer amplifies rather than absorbs: Artemis counted 116 liquidations across roughly $7.2M of Apyx and Saturn collateral between early June and July 16, and estimated that a further 10% decline puts about $5.7M of Apyx-related debt in liquidation range, assuming borrowers neither repay nor top up. Liquidator bots then sell into the same thin book already absorbing nervous holders. A separate, smaller path needs no market move at all: the Curve pool's rate oracle is hand-set by a 4-of-6 Safe with no timelock and no staleness check. The backstop is thinner than the "Nasdaq-listed backer" line suggests: DFDV closed at $2.75 on August 4, and in a solvency event holder claims route through a BVI foundation, not through DFDV. None of this requires an exploit, a rogue insider, or a contract bug. It requires Strategy to have a bad quarter, which it just did, to the tune of $8.2B.

Bottom line

Avoid, and the recovery does not change it. The honest reading of the last month is that apxUSD stopped trading at a discount to its own collateral and started trading at that collateral's value: $0.9177 in the market against $0.9131 of attested redemption value. That is the market functioning correctly on a reserve that is still about 8% short. Coverage did improve, from roughly 84% in late June to 92.24%, and that should be said without hedging. What did not improve is everything structural. STRC concentration is flat at about 84% of asset reserves. There is still no atomic redemption. The protocol's own Curve LP went to zero. The oracle pricing the main pool is a hand-set number behind no timelock. And the venue that would take real size is bidding 83 cents. The one improvement that is Apyx's own doing, burns now being visible on-chain, is a reporting change rather than a solvency change. Re-rate when four things are true together: the attested ratio holds at or above 100% for a sustained window, Apyx 2.0's Redemption Value model lands in the docs and in a contract instead of a blog post, real USDC depth and a Kraken bid near par return, and the market price sits at par because the mechanism forces it there rather than because the portfolio rallied. Until then the $1 on the ticker is still a design goal, not a promise.

Data appendix

  • Supply and price (DeFiLlama stablecoin API + CoinGecko, 2026-08-05): apxUSD price $0.9177 (DeFiLlama) and $0.9193 (CoinGecko). DeFiLlama's totalCirculating is 324.37M tokens and its totalCirculatingUSD is $297.67M; these are token count and market value respectively and must not be interchanged. CoinGecko independently gives 324,376,538 supply and a $298.21M cap. Cross-chain reconciliation holds: Ethereum 312.07M plus Base 9.77M plus BNB 2.44M equals about 324.28M. Token supply is contracting: 361.57M on July 6, 339.19M on July 27, 324.37M today, down 38.2% from a peak of 524.66M on May 28. 24h volume $1,152,374 across 37 tickers on 8 venues. ATL $0.717601 on 2026-06-26 (CoinGecko); the July report recorded $0.699 from the same source, so CoinGecko appears to have revised it, and DeFiLlama's daily mark for that window is $0.749, which Yearn corroborates at about $0.75.
  • Public float, definitional note: DeFiLlama counts all 324.37M tokens as circulating. Accountable's own attestation treats 31.93M of inventory and 44.84M of POL as not circulating, which puts non-protocol-held supply at 235.30M on Ethereum, or about 247.5M tokens and $227.1M including Base and BNB. Both bases appear below; the exit-depth ratio is quoted against the larger DeFiLlama figure.
  • TVL: DeFiLlama still has no TVL series. Slug apyx returns "Protocol not found"; apyx-protocol returns protocol metadata (id 7618, category RWA) with an empty chainTvls.Ethereum.tvl array. Re-confirmed this run. Yearn states the same: "Not tracked by DeFi Llama."
  • Collateralization (Accountable proof-of-solvency, snapshot 2026-08-03 11:26 UTC, via Yearn): asset reserves $217,035,429.31 against circulating supply $235,301,450.21, ratio 92.2372%, redemption value $0.9131. Separately: inventory $31,929,748.28, POL $44,842,316.33, total reserves $293,807,493.92, total supply $312,073,514.82. Both identities were recomputed here and reconcile exactly, and the $312,073,514.82 total supply matches the Ethereum totalSupply() read on-chain today to the cent. Trajectory: about 84% late June, 90.7% on July 21 (Artemis via Forbes), 92.24% on August 3. The live feed at accountable.apyx.fi and api.accountable.apyx.fi returns 403 to automated fetching, so this is the most recent independently retrieved snapshot rather than a today reading.
  • Reserve composition: STRC $183,285,475.70, cash and equivalents $33,743,189.09, other $6,764.52. STRC is 84.45% of asset reserves; on a base that also counts POL it is 69.99%, with cash 12.89% and POL 17.12%. Forbes and Artemis independently put STRC at 84% of reserves on July 21 against gross reserves of $307M and direct Stretch exposure of $196M. Concentration is therefore flat on the narrow base and down on the wide base, not rising.
  • STRC collateral (Yahoo Finance, close 2026-08-04): $92.53, 7.47% below its $100 par, recovered 24.08% from the $74.57 low on 2026-06-26. Strategy's June 29 Digital Credit Capital Framework raised the STRC annual dividend rate to 12.00% for record dates on or after July 1, 2026, added a USD Reserve policy set at a minimum of 12 months of preferred dividends and interest, and authorized $1.0B repurchase programs for Digital Credit Securities and for class A common plus a BTC Monetization Program of up to $1.25B. Dividend tape confirms the change: monthly $0.958 through June 15, then $0.479 on June 30 and $0.50 on July 15 and July 31, an annualized $12.00 on par. Reserve reported at $2.55B on June 28 (17.4 months of coverage, implying about $1.76B of annual obligations) and about $4B by early August (about 27 months). Bitcoin sales: 3,588 BTC for $218.5M between June 29 and July 5, plus about $105M in the week before August 3. Q2 preferred dividends $400.7M; Q2 loss on bitcoin $8.2B.
  • Audits: verified against docs.apyx.fi/resources/audits, which lists Halborn 2026-06, Quantstamp 2026-04, Certora 2026-03, Zellic 2026-03 and Quantstamp 2026-02. All four firms are real engagements; none invented. Yearn's August assessment covers Quantstamp, Certora and Zellic but does not mention Halborn. Public repo apyx-labs/evm-contracts with invariant tests and Slither CI, no license specified. No bug bounty, per an exhaustive search across Immunefi, Sherlock, Cantina, HackerOne and Safe Harbor. Report contents not read line by line this run.
  • Admin / custody (Safes re-read on-chain 2026-08-05): Admin Safe 0xabdd8c8ee69e5f5180eb9352aeffc5ceead65e96 returns getThreshold() 4 with 6 owners, holds ADMIN_ROLE at zero execution delay, granted 2026-03-20. Guardian/Upgrader Safe 0xf9862EfC1704aC05e687f66E5cD8c130E5663cE2 returns getThreshold() 3 with 6 owners, no longer holds ADMIN_ROLE, and holds UPGRADER (3-day delay), PAUSER (0), UNPAUSER (4h), YIELD_OPERATOR (0). The prior "4-of-6 versus 3-of-6" discrepancy was two different Safes, and is resolved. Guardian Safe Curve LP went 40,890,164 to under 10,000,000 by block 25252968 (June 1 to 5) and to 0 by block 25474518 (July 6). Deny list 0x2c271ddF484aC0386d216eB7eB9Ff02D4Dc0F6AA, zero delay to add and 4 hours to remove, 141 added and 137 denied as of 2026-07-27. MinterV0 0x2c36e1ad holds MINT_STRAT_ROLE with a 60-second execution delay. Custody at Alpaca under Preference Foundation; Apyx incorporated in the British Virgin Islands; operations Safe 0x37B0779A66edc491df83e59a56D485835323a555 holds 582,774 STRCX, about 35% of STRCX supply. Backer DeFi Development Corp (Nasdaq: DFDV) closed at $2.75 on 2026-08-04; Forbes describes the fall as $42.50 to about $2.70, while Yahoo's two-year weekly closes peak at $34.25, so the $42.50 is Forbes' figure and is likely an intraday high.
  • Oracle: ApxUSDRateOracle proxy 0xa2ef2e7bf32248083e514a737259f3785ea8d37d returns 1.000000 on-chain today. Manually set via setRate(); getTargetFunctionRole(oracle, setRate) is 0 (ADMIN_ROLE) and getTargetAdminDelay(oracle) is 0, so the Admin Safe can change it instantly. No feed, no TWAP, no staleness check. Used by the Curve StableSwap-NG pool for pricing. Accountable DVN registry: verifiability 4 of 5, connectors 3, frequency live, live since 2026-04-23; level 4 is secure-enclave verification, level 5 is zkTLS. Accountable's supply_split itemizes Ethereum only.
  • Exit depth (on-chain balances read 2026-08-05, cross-checked against DexScreener): Curve apxUSD/USDC 0x6F63deEDc9870D6c16FC644C6654748352cdc87c holds 7,418,710 apxUSD and 980,037 USDC, about $7.79M nominal; this pool is absent from Curve's factory-stable-ng registry. Uniswap V4 apxUSD/USDC pools hold about 780,055 and 192,710 USDC. PancakeSwap V3 0x1D8177897FC90819CF644fa84B3247AC690985D5 holds 3,001,707 apxUSD against 1,707 USDC. Old Curve pool 0xE1B96555BbecA40E583BbB41a11C68Ca4706A414 holds 10,444 apxUSD and 1,399 USDC, about $11.0K. USDC total across all apxUSD/USDC venues: about $1,955,900. Kraken lists APXUSD/USD and APXUSD/USDC; best bid $0.83 on both, total bid-side depth about $252,800. Combined dollar-side exit about $2.21M, or 0.74% of the DeFiLlama float. Curve apyUSD/apxUSD 0xe41be7B3 holds about $18.78M and provides no dollar exit; STRCx/apxUSD holds about $983. Total DEX liquidity across 19 apxUSD pairs is about $37.8M nominal against $44.8M of reported POL, which suggests but does not prove that most visible depth is protocol-owned.
  • Redemption mechanics: No atomic on-chain redemption. ApxUSD.mint() checks only the mint role and the supply cap, with no transferFrom and no collateral deposit. Direct redemption permissioned, gated to approved counterparties, priced at redemption value. apyUSD unlock delay 1,728,000 seconds (20 days) with a 0.1% admin-settable fee and no yield accrual during cooldown. Apyx 2.0 (announced June 15, 2026) proposes a single Redemption Value floor plus an RFQ for approved counterparties; the docs Peg Stability Model page still describes the old overcollateralization-plus-arbitrage mechanism with no mention of Redemption Value, so it remains blog-only.
  • Attestations: Wolf & Company monthly Independent Accountant's Reports under AICPA standards are published through June 2026 (report dates June 17 and 30, opinion dated July 22, $193,307,068 total: $109,512,763 at Alpaca plus $83,794,305 of self-custodied on-chain STRCx). Scope covers reported assets' existence, ownership, custody and valuation at two dates per month; it does not opine on apxUSD liabilities or collateral coverage, and the bank or custodian for cash and cash-equivalent balances is still not named, so the cash sleeve remains outside attested scope. The July report is not yet due.
  • Recent-news scan: Forbes, July 28, 2026, "DeFi 'Stablecoin' Operations Built On Saylor's Preferred Stock Are Now On Shaky Ground" (Artemis data: 90.7% coverage on July 21, STRC 84% of reserves, 116 liquidations on about $7.2M of Apyx and Saturn collateral from early June to July 16, about $5.7M more at risk on a further 10% decline; Apyx and Saturn did not respond to requests for comment). Yearn risk-score assessment updated August 1 and corrected August 3, 2026, score 3.73/5.0. TID Research last revised 2026-07-13. Apyx's blog shows no August post; the most recent entries are "Making Smart Contracts Observable, Finally" (July 13) and a July 1 defense of Strategy written when STRC was near $84. No exploit, custody loss or contract failure identified through 2026-08-05. The impersonator-scam warning stands: no official repeg or compensation vote exists, and accounts promoting one (@apyx_FD, @apyxfnd_, GovernanceApyx) are not Apyx; the official handle is @apyx_fi.

Maintained monthly. Methodology: DeFi Research Instruction v2.

← all protocols