Skip to content
PROTOCOL RESEARCH

Re risk

Re's reinsurance yield is real and the plumbing holds up under direct inspection: the 48h timelock answers 172800 on-chain, all 52 Hacken findings across three engagements are marked resolved with none accepted or mitigated, and the Ethereum redemption buffer sits at 27.5% of reUSD supply value today. Exit is what gates the verdict. The senior token's fallback below the 1% floor is quarterly-window-only rather than a rolling queue, the Curve pool the docs call the primary reUSD exit holds $462k, the junior token's first redemption window in protocol history is already oversubscribed against a $1.5M pool, and the observed buffer is roughly half the protocol's own stated policy of at least 50% of the deposit high-water mark.
CautionResearched Jul 20, 2026

Scorecard

watch: Audits & contractsThree Hacken engagements re-read first-hand this run via a text proxy: Re-DeFi (Sep 4 2024, 29 findings, 29 resolved, 0 accepted, 0 mitigated, 42.11% branch coverage), Re-Contracts (Dec 31 2024, 15 findings, 15 resolved, 0 accepted, 0 mitigated, 3 High, test coverage reported N/A with Hacken noting tests rely on mocks so their reliability is compromised), NAV-Oracle (Apr 11 2025, 8 findings, 8 resolved, max Medium, 95.83% branch coverage), plus Certora formal verification Sep 2025 per protocol docs. Nothing accepted or merely mitigated, which is better hygiene than most of this cohort. Against that: contracts are UUPS-upgradeable, the newest published report is about fifteen months old, both Hacken scopes carry explicit exclusions (Re-Contracts excludes PoolRegistry, DepositTokenRegistry, SharePriceCalculator and KYCRegistry; Re-DeFi excludes governance), and the audited token contracts are named USDRE, USDRWA and ReToken. Neither Hacken report names reUSD or reUSDe.
watch: Admin controlTimelockController 0x69dDEa332723cF5407151aAF68B9b076557FCA93 returns getMinDelay 172800 seconds, re-verified live on-chain this run, holding DEFAULT_ADMIN_ROLE and UPGRADER_ROLE over both ICLs and both tokens with no bypass; upgrades executed by a 3-of-5 governance MPC, day-to-day roles split across four wallets (Access Manager is 5-of-8), Fireblocks 3-of-5 for custody sweeps. Against that: the Daily Instant Redemption Vault carries a guardian break-glass emergencyWithdraw, the NAV oracle has a forceNAVUpdate emergency path Hacken flagged as able to manipulate NAV arbitrarily if the emergency updater turns malicious, and signer identities are undisclosed. Token holders hold no direct legal claim on the Section 114 trust assets.
watch: OracleNAV is pushed once daily at 00:00 UTC, monotonic, daily upward moves capped at 20%/365 with per-update deviation enforced on-chain and emergency updates requiring 3-of-5; the sUSDe mint and redeem oracle is capped at the intrinsic exchange rate and redemptions hard-revert if USDe spot is under $0.99. Off-chain and Section 114 trust balances are attested daily by The Network Firm and published via Chainlink, but the most recent full AUP report is October 2025. Hacken's own note is the sharper one: the deviation check means that during high volatility the automated feed will not report accurate data and the system falls back on a manual forceNAVUpdate with no guarantee the fed data is correct. A loss breaking during market hours does not touch NAV until the next daily print, and for reUSDe not until the next quarterly actuarial revision.
weak: Liquidity & exitBuffer is healthy right now: the two Ethereum redemption contracts hold 35,170,194 sUSDe (about $43.6M, 27.5% of reUSD supply value) against a 1% cut-off, so a $50-130k redemption clears same day at a 6bp fee. Three structural problems. Below 1% the contract does not run a rolling FIFO queue, it switches to quarterly-window-only mode. reUSDe is quarterly by design with a 40-day minimum hold, and its first window in protocol history (July 9 to July 22 2026, $1.5M pool) already showed $1.79M requested across 51 wallets, so pro-rata rationing is likely. And 27.5% is roughly half the protocol's own stated policy that at least 50% of the deposit high-water mark stays on-chain as redemption liquidity. Secondary depth is thin: Curve reUSD/USDC holds $462k, reUSDe/sUSDe $554k.
strong: Yield (real vs emission)Live 7-day APY is 6.18% on reUSD and 12.23% on reUSDe (Re API, 2026-07-19), sourced from reinsurance premium plus SOFR or sUSDe basis, not token inflation. The underlying instrument is a $100M Principal-at-Risk Note program issued April 2 2026 paying 30-day average SOFR plus 500bps (M1) and 950bps (M2), with token-level spreads of 250bps for reUSD and 850bps for reUSDe. $RE launched publicly in June 2026 as a governance and utility token with an explicit disclaimer of any claim on revenue, earnings, reserves, collateral or insurance flows; the points program is a separate incentive layer.
watch: Holder concentrationMeasured this run from Re's own public points API (2026-07-18) and it is concentrated by wallet count: 35 wallets hold $25.5M of reUSD as Morpho collateral, 31 wallets hold $14.3M of the Fluid reUSD/USDT LP, 256 hold $30.3M of Fluid reUSD collateral, 6 wallets hold $631k of reUSDe on Morpho. Directly-held reUSD is $6.84M across roughly 1,800 wallet entries on all chains and reUSDe $2.56M across 464; total programme participants are 3,732. Most of the float lives inside a handful of levered DeFi positions rather than in end-holder wallets. Per-address rankings could not be pulled: Etherscan tokenTopHolders is API Pro only and returned NOTOK this run.
watch: Track recordNo protocol-level exploit and no depeg found this run, and nothing for Re among the 588 entries on the DeFiLlama hacks list. Underwriting predates the token: 48 programs, 49 US states, 700,000+ policyholders, a $510.5M portfolio as of June 30 2026 and a 92% all-lines combined ratio to date, all protocol-reported and not independently verified. The on-chain product is 13 to 15 months old (first recorded reUSDe price 2025-04-14, reUSD 2025-06-18) and the very first junior redemption window is still settling as of this writing.
🟢 strong🟡 watch / caveat🔴 weak / fund-loss risk
Verdict is a gate (worst flaw wins), not an average. Our read, not financial advice.

Live facts

auto-sourced now
TVL$264.3M
30d↑1%
Audits2
Last hacknone

DeFiLlama + our exploits feed. Cross-check the dated report against today.

TL;DR

Re routes KYC-gated stablecoin deposits into collateral for reinsurance treaties written by one Cayman reinsurer, Cover Re SPC. reUSD is the senior tranche, reUSDe the mezzanine, and the reinsurer's own capital sits below both. This refresh closed the four gaps we left open in July. Governance holds up under direct inspection: the timelock answers 172,800 seconds on-chain, all 52 Hacken findings across three engagements are marked resolved with nothing accepted or mitigated, and the reUSD redemption buffer is currently 27.5% of supply value, nowhere near its 1% trip point. The TVL discrepancy also resolves cleanly. What got worse rather than better is exit. The senior token's fallback below the buffer floor is not a rolling queue that keeps paying, it is quarterly-window-only mode. The junior token's first redemption window in protocol history opened July 9 with a $1.5M pool and was already oversubscribed. The Curve pool the docs call the primary reUSD exit holds $462k. And 27.5% is roughly half of what Re's own documentation says it keeps on-chain. Caution, on exit rather than on solvency.

Checklist

Audits & contracts. Hacken's index and all three report bodies were read first-hand this run through a text proxy, which closes the 403 gap from July. Re-DeFi (September 4 2024): 29 findings, 29 resolved, 0 accepted, 0 mitigated, 42.11% branch coverage. Re-Contracts (December 31 2024): 15 findings, 15 resolved, 0 accepted, 0 mitigated, including three High severity (a funds-lock via CANCEL_TIMEOUT in cancelDeposit, incorrect allowance handling in the Pool contract, incorrect fee handling on collateral and profit return). NAV-Oracle (April 11 2025): 8 findings, 8 resolved, highest severity Medium, 95.83% branch coverage. Certora ran formal verification in September 2025 per Re's own audits page. Zero findings sit in accepted or mitigated status, which is better hygiene than most of this cohort.

Three caveats the finding counts hide. First, Hacken rated the Re-Contracts test suite coverage as N/A and wrote that the tests "predominantly rely on mock contracts rather than the original implementations, which raises concerns about their effectiveness," so the reliability of those tests is, in Hacken's words, compromised. Second, both scopes carry explicit exclusions: Re-Contracts excludes PoolRegistry, DepositTokenRegistry, SharePriceCalculator and KYCRegistry despite the in-scope contracts depending on them, and Re-DeFi excludes governance. Third, and largest, the audited token contracts are named USDRE, USDRWA and ReToken. Neither Hacken report names reUSD or reUSDe. Nor does any published report we could find map to the instant redemption entrypoint (0x8aEb9453), the Daily Instant Redemption Vault (0x5C454f55), the Surplus Note Registry, or the Katana, Plasma, BNB, Ink and Solana deployments. Contracts are UUPS-upgradeable, not immutable, and the newest published report is about fifteen months old.

Admin control. Re-verified directly this run. getMinDelay() on 0x69dDEa332723cF5407151aAF68B9b076557FCA93 returns 172800, a 48-hour delay with no bypass, and per Re's docs it holds DEFAULT_ADMIN_ROLE and UPGRADER_ROLE over the reUSD ICL, the reUSDe ICL and both tokens. Upgrades are executed by a 3-of-5 governance MPC. Operations are split across four controller wallets: Oracle Config (3-of-5), Redemptions Config (3-of-5), Custodian Manager (3-of-5), Access Manager (5-of-8). Emergency pause runs through the Access Manager plus the Redemptions Config wallet, and unpause requires executive and risk approval followed by a further 48-hour delay clocked from the on-chain approval event. That is a genuinely well-separated setup and it telegraphs privileged changes two days ahead.

Two things keep it yellow. The Daily Instant Redemption Vault has a guardian emergencyWithdraw break-glass, and the NAV oracle has a forceNAVUpdate() path that Hacken flagged explicitly: if the emergency updater role acts maliciously, the function "could be exploited to arbitrarily manipulate the NAV value, undermining the protocol's integrity and affecting all dependent systems." Worth being precise about the mitigation, because our July draft got it wrong: the on-chain deviation check sits in fulfillRequest() on the automated path, and forceNAVUpdate() is the bypass around it. What actually constrains the bypass is a 4-hour minimum between forced updates and a 3-of-5 signing requirement, not a deviation check. The second issue is legal rather than technical. Re's audits page states plainly that reUSD and reUSDe holders "have no direct legal claim on §114 trust assets." Exposure runs through a Principal-at-Risk Note that is an unsecured, limited-recourse obligation of segregated portfolio SP1 only, with no cross-portfolio recourse, no recourse to general assets and no third-party guarantee. That is a structure you accept, not one you verify on-chain.

Oracle. Daily NAV push at 00:00 UTC, monotonic upward, daily moves capped at 20%/365, maximum per-update deviation enforced on-chain, emergency updates requiring 3-of-5. The sUSDe oracle used for mints and redeems is a spot oracle capped at the intrinsic exchange rate so a basis dislocation cannot inflate the quoted price, and the redemption contract hard-reverts if USDe spot trades below $0.99. Off-chain and Section 114 trust balances are attested daily by The Network Firm under Agreed-Upon Procedures and published through Chainlink on Avalanche. The most recent full AUP report is October 2025.

The sharpest criticism here is Hacken's, not ours. The audit notes that because of the deviation check, "during periods of high volatility, the system is not going to report real or accurate price data," and in exactly those periods it depends on the emergency role to feed data manually, with "no guarantee that the fed data is correct or under control." Layer on the design timing: for reUSD the oracle is a day behind, and for reUSDe the tNAV only revises at the quarterly actuarial print. A loss that becomes public intraday leaves lending markets valuing collateral above what a liquidator could realise. Re's own curator guide says as much.

Liquidity & exit. Three facts pull against each other and the structural ones win.

The good one: the Daily Instant Redemption Vault (0x5C454f55) and the Redemption Reserve Custodian (0x9eA38e09) together hold 35,170,194 sUSDe on Ethereum, about $43.61M at the live 1.239836 exchange rate, against a reUSD supply value of $158.60M. That is 27.5%, and the trip point is 1% (about $1.59M). Instant redemptions are capped at 20% of the available pool per day, which today is about $8.72M protocol-wide, so any position in the $50-130k range clears same day at a 6bp fee, executed at the 00:00 UTC NAV. Note that the per-wallet limit is stated two different ways in Re's own documentation: the redemptions page says 10% of that daily pool (about $872k today), while the reUSD product page says no single redemption may exceed 10% of the available buffer (about $4.36M today). At our size the difference does not bite, but it is an unresolved discrepancy in the primary source.

The bad ones. The docs are explicit that if the buffer falls under 1% of supply, redemptions "move to the quarterly queue." First-come-first-served applies only while the buffer lasts; after that the senior token behaves like the junior one. Our July report described a FIFO queue that keeps clearing. That was wrong. Separately, Re's own product overview states that "at least 50% of the high-water mark remains onchain as redemption liquidity." The observed buffer is 27.5% of current supply value. The high-water mark of deposits is not published, so we cannot compute the shortfall precisely, but on any plausible reading of that denominator the buffer is running at roughly half the stated policy. That is a gap worth watching monthly, because the whole senior-tranche liquidity claim rests on it.

Secondary markets do not compensate. Curve reUSD/USDC, which the docs call the "primary reUSD exit; deepest stable-to-stable liquidity," holds $461,555. reUSD/sUSDe holds $2.34M, reUSD/sUSG $238k, and the reUSDe/sUSDe twocrypto pool $554k against $20.1M of reUSDe outstanding. The docs' slippage table (under 5bp below $50k, 5 to 20bp from $50k to $500k) does not describe a $462k pool, and the docs themselves warn that the StableSwap curve must be reconfigured as NAV drifts above $1.00 or "will produce misleading slippage estimates." Fluid's reUSD/USDT LP, at about $14.27M per Re's own points data, is the deepest venue in practice.

PT-reUSD exit mechanics (Pendle expiry 2026-12-10)

At expiry. The market's accounting asset is USDC (assetPriceUsd 0.99987), so one PT-reUSD-10DEC2026 redeems for SY-reUSD worth $1, which unwraps to roughly 0.918 reUSD at the current NAV of 1.08984. You do not receive USDC. You receive reUSD, and you land inside Re's redemption regime.

Then reUSD to USDC. Instant redemption via redeemInstant(shares, minPayout) requires KYC, on both minting and redeeming. With the buffer at 27.5% and today's caps, a $50-130k redemption is a rounding error against the daily ceiling: same-day, calm conditions, 6bp fee, at the 00:00 UTC NAV. The payout asset needs a caveat. The docs say you "receive your deposit asset," but Re's product overview also says redemption liquidity is held "primarily in sUSDe, with some USDC, USDe, or other assets temporarily held depending on what users mint with. All of those are ultimately converted into sUSDe." Every token in the Ethereum buffer we measured is sUSDe. So a USDC-denominated exit plausibly requires one further swap, and the reUSDe window settles explicitly in sUSDe. Under stress the picture inverts. There is no partial-fill queue below the 1% floor, only the next quarterly window, and the secondary route is a $462k Curve pool. Without KYC the instant path does not exist at all and you are a price-taker on Curve or Fluid.

Queue history. No episode of the reUSD instant buffer being exhausted surfaced this run, and none is documented; treat that as not-found rather than clean. The junior side has one data point and it is fresh. The first reUSDe redemption window in the protocol's history ran July 9 to July 22 2026 with a $1.5M pool and pro-rata fill above it, claims open July 23 to August 22, settled in sUSDe. Re's points API showed $1,787,481 of reUSDe in the redemption queue across 51 wallets as of July 18, already 19% above the pool. One detail worth flagging that we missed in July: if the 30-day claim period elapses without claiming, the sUSDe payout for that window is forfeited and only the original reUSDe is returned. Final fill ratio is not yet knowable, but the first junior exit in protocol history was rationed on a pool worth 7.5% of the reUSDe float.

Early exit via Pendle. Cheap at our size. The PT market (0x13285bcb) holds $7.57M of AMM liquidity; a live quote for a $130k sell_pt returns 0.159% price impact and $106 of swap fee, about 1.7% of the pool. Implied APY is 10.30% against a 6.37% underlying, a 3.78% PT discount, PT price $0.9621.

Morpho loop unwind. The market is PT-reUSD-10DEC2026 against USDC at LLTV 91.5%, not 92%: $40.16M supplied, $34.75M borrowed, $39.92M of collateral, $5.41M of idle liquidity, 86.5% utilisation, 8.04% borrow, zero bad debt. The carry is thin. Fixed PT yield 10.30% minus 8.04% borrow leaves 226bp of gross spread before leverage, so the whole trade lives or dies on the borrow rate. If lenders withdraw the $5.4M of idle USDC, utilisation runs toward 100% and Morpho's adaptive IRM ratchets the rate up until the market clears; anything above 10.30% flips the loop to negative carry while your collateral is locked to a December maturity. The rate at target is already 8.29%. The exact path at full utilisation is refresh. The crowding is the second problem: PT-reUSD-10DEC2026 total supply is 78,324,515 tokens (six decimals, read on-chain), about $75.36M at the current PT price, of which $41.28M sits as Morpho collateral across the USDC and USDT markets. That is five and a half times the $7.57M of Pendle AMM depth. A single-name unwind at our size is trivial. A crowded one is not.

On the Aave Umbrella analogy. It holds at the level of shape and breaks everywhere it matters. What holds: both are two-tier structures where a junior layer is paid a spread to absorb losses first so the senior can hold par, and in both cases the junior is thin relative to the senior (reUSDe is $20.11M against $158.60M of reUSD, about 12.7%, with reinsurer capital sitting below it). What breaks. Umbrella's loss trigger is an observable on-chain deficit in a specific Aave reserve, and slashing executes in code that anyone can read; Re's trigger is an actuary confirming underwriting losses off-chain, applied at a quarterly tNAV revision you cannot see coming. Umbrella's senior holders can withdraw any block subject to utilisation; reUSD holders have a capped daily buffer that falls back to a calendar, and reUSDe holders have a calendar only, pro-rata, first window oversubscribed, with a 40-day minimum hold. Umbrella stakers are slashed on-chain; reUSDe's loss absorption is contractual, inside a limited-recourse note against a Cayman segregated portfolio, with no direct claim on the trust assets. And Umbrella's tail risk is correlated with crypto, while Re's is homeowners and casualty lines, which is genuinely diversifying, except that Re's on-chain buffer and its junior payouts are both denominated in sUSDe, so it reimports Ethena exposure through the back door. Structurally similar. Not the same thing to underwrite.

Yield: real vs emission. Real, and it holds up. The Re API reports 6.18% on reUSD and 12.23% on reUSDe on a 7-day basis (2026-07-19). The economic source is a $100M Principal-at-Risk Note program issued April 2 2026, five-year term to April 2 2031 with rollover by mutual agreement, paying 30-day average SOFR plus 500bps on M1 and 950bps on M2, payable annually in arrears and subject to CIMA solvency and collateral deferral gates with catch-up. Token-level spreads are 250bps for reUSD and 850bps for reUSDe, so the gap between note coupon and token yield is the protocol's take. $RE went public in June 2026 with 159.6M circulating of a fixed 1B (16%, CoinGecko, 2026-07-19, price $0.396, market cap $63.2M); its own tokenomics page states it "does not represent equity, debt, profit-sharing, dividend or fee-distribution rights, or any claim on the revenue, earnings, insurance flows, reserves, collateral, or other assets" of Re. One correction to our July draft: we described 95% of the day-one community allocation as vesting over three years conditional on locked TVL. The actual schedule is that claimants at or below 150M points receive 100% at TGE, larger claimants receive roughly 10% to 37%, and the vested remainder splits into six six-month tranches over 36 months, with a TVL holding requirement attached to tranches one through four only.

Holder concentration. Now measured, from Re's own public points endpoint (2026-07-18), and the answer is that reUSD is mostly not held by holders. Directly-held reUSD is $6.84M across roughly 1,800 wallet entries on all chains (of which $6.38M and 727 wallets on Ethereum) and reUSDe $2.56M across 464. Everything else sits inside DeFi positions with small wallet counts: $25.5M of reUSD as Morpho collateral across 35 wallets, $30.3M as Fluid collateral across 256, $14.3M in the Fluid reUSD/USDT LP across 31, $2.34M in the Curve reUSD/sUSDe LP across 34, and $631k of reUSDe on Morpho across 6. Total programme participants are 3,732. That is a young product whose float is concentrated in a few dozen levered wallets, which is exactly the population that unwinds together. Per-address rankings are still missing: Etherscan's tokenTopHolders endpoint requires API Pro and returned NOTOK this run.

Track record. No protocol-level exploit, no depeg, and nothing for Re among the 588 entries on the DeFiLlama hacks list. One thing needs disambiguating and it now has a firm answer: a February 4 2026 exploit is documented against a SingleAdapterRouter contract at 0x169a5eff whose verified source tree is strategia_ETH/REUSD/Vault_reUSD.sol, draining roughly $25.8k. We pulled that contract's verified source this run and none of Re's reUSD, reUSDe or ICL addresses appear anywhere in it, and it is not in Re's published contract set. Attribution is unresolved and it should not be read as a Re incident. Separately, the Curve pool listed as "reUSD/scrvUSD" with $6.99M is Resupply's reUSD (0x57aB1E00), a different token entirely, as is the June 2025 Resupply exploit. On the underwriting side, June 30 2026 reporting shows a $510.5M portfolio across 48 programs and 49 US states with more than 700,000 policyholders, $149M of it written in the prior 45 days, homeowners share up from 10% to 18%, and Re's docs claim a 92% all-lines combined ratio to date and a sub-100% ratio in every underwriting year since inception. All protocol-reported. First recorded prices are 2025-04-14 for reUSDe and 2025-06-18 for reUSD, so the on-chain product is 13 to 15 months old.

Worst case

A loss year pushes Cover Re's combined ratio well above 100%. Reinsurer capital absorbs first, then reUSDe, then reUSD, and Re's own modelling puts reUSDe impairment at roughly 0.9% probability under a 115% combined ratio and reUSD impairment at roughly 0.03% under 135%. But the mechanism that bites long before impairment is the exit. Redemption demand drains the buffer below 1% of supply and the senior token stops being redeemable on demand entirely, switching to a quarterly calendar; the junior token was already on one and its first window was rationed at $1.5M against $1.79M requested. Secondary escape is a $462k Curve pool. Meanwhile a widening discount to NAV on the DEX does not touch the NAV oracle until the next 00:00 UTC print, and for reUSDe not until the next actuarial revision, so lending markets keep marking collateral above realisable value and liquidators bid accordingly. For a PT-reUSD position specifically, the maturity is fixed at December 10 2026 and redeems into reUSD, not USDC, so a stressed exit means selling reUSD into thin pools or waiting for a window. If the $41.3M Morpho loop unwinds at once into $7.57M of Pendle depth, the PT discount does the work the AMM cannot. And underneath all of it, your legal position is an unsecured limited-recourse note against one Cayman segregated portfolio, with daily attestations of trust balances you cannot independently verify and no direct claim on those assets.

Bottom line

Caution, and the reason has moved. The governance and audit story survived a deliberate attempt to break it: the timelock is real and answers on-chain, every audit finding is resolved with none accepted or mitigated, the buffer is fat today, and the TVL numbers reconcile once you know what each source counts. What does not survive is exit. A senior stablecoin whose stress fallback is a quarterly calendar is not a liquid instrument, a $462k primary DEX pool is not a fire escape, and a buffer running at roughly half the protocol's own stated 50% policy is the number to watch each month. Size this as an off-chain credit and insurance position with a fixed maturity you should plan to hold to, not as dry powder. For PT-reUSD specifically, the entry and the early exit are cheap and the loop carry is thin enough (226bp) that the borrow rate, not the reinsurance, is what will actually determine the outcome.

Data appendix

  • TVL (reconciled): DeFiLlama on-chain $251,435,671 (api.llama.fi/tvl/re, 2026-07-19); by chain Ethereum $247.07M, Base $4.35M, Avalanche $5,066, Arbitrum $6,883. DeFiLlama's stated methodology is custodian wallets plus redemption reserves plus off-chain assets tracked via oracle. Re's own API for the same day: on-chain capital $68,801,125 plus off-chain capital $179,108,156 equals $247,909,280, a 1.4% gap to DeFiLlama. The two therefore do reconcile, contrary to our July report. Re's headline total of $566,835,239 adds $318,925,958 of premium receivables, which DeFiLlama does not count; Re's June update states the same split as $70.97M / $177.67M / $318.93M at the June 30 snapshot. Token float is smaller than both: reUSD product TVL $159,316,667 and reUSDe $20,113,282, sum $179.43M, so roughly $68M of the capital base is reinsurer and cell capital rather than user deposits. Verified independently: summing the six Ethereum protocol addresses (the two ICLs, the two ICL custodial wallets, the Daily Instant Redemption Vault and the Redemption Reserve Custodian) gives 47,252,428 sUSDe worth $58.59M, plus USDC $8,730,083, USDT $441,002 and USDe $245,373, for $68.00M against the reported $68.80M on-chain capital; the residual is the non-Ethereum chains.
  • Supply / NAV: reUSD 145,525,261.68 tokens at NAV 1.089835 (supply value $158,598,536); reUSDe 14,469,723.66 at NAV 1.390025 (supply value $20,113,282). Both re-read on-chain totalSupply and matched the Re API exactly. 7-day APY 6.18% and 12.23%.
  • Redemption buffer (gap closed): Daily Instant Redemption Vault 0x5C454f5526e41fBE917b63475CD8CA7E4631B147 holds 2,187,092.41 sUSDe; Redemption Reserve Custodian 0x9eA38e09F41A9DE53972a68268BA0Dcc6d2fAdf8 holds 32,983,101.33 sUSDe. Total 35,170,193.74 sUSDe, $43,605,279 at convertToAssets 1.239836, equal to 27.49% of reUSD supply value against a 1% trip point of $1,585,985. Implied daily cap about $8.72M protocol-wide. Fee 6bp (0.06%, confirmed in docs). Ethereum only; Avalanche, Arbitrum and Base have their own vault and custodian pairs which were not summed.
  • Stated buffer policy vs observed: docs state the buffer is sized on "50% of the highwater mark of reUSD total deposits" and that "at least 50% of the high-water mark remains onchain as redemption liquidity." Observed 27.49% of current supply value. The high-water mark itself is not published, so the shortfall is indicative rather than exact. Watch monthly.
  • Redemption mechanics (corrected): instant while the buffer holds, first-come-first-served, 20% of the available pool per day; below 1% of supply the contract switches to quarterly-window-only mode, not a rolling queue (docs.re.xyz/minting-and-redemptions/redemptions and products/about-reusd). Per-wallet limit stated inconsistently across two doc pages: 10% of the daily pool (redemptions page) versus 10% of the available buffer (about-reusd page). KYC is required to mint and to redeem.
  • reUSDe first redemption window: requests July 9 to July 22 2026, claims July 23 to August 22, $1.5M pool, pro-rata above it, settled in sUSDe, 40-day minimum hold, unclaimed payouts forfeited after the 30-day claim period. Re's points API showed $1,787,481 queued across 51 wallets as of 2026-07-18. Final fill ratio refresh.
  • Audits (gap closed, bodies read): Hacken index plus all three report pages. Re-DeFi, Sep 4 2024, 29 findings, 29 resolved / 0 accepted / 0 mitigated, 42.11% branch coverage, 10 of 29 rows published (4 Medium, 6 Low), scope excludes governance. Re-Contracts, Dec 31 2024, 15 findings, 15 resolved / 0 accepted / 0 mitigated, 3 High, coverage N/A with tests relying on mocks, scope excludes PoolRegistry, DepositTokenRegistry, SharePriceCalculator and KYCRegistry. NAV-Oracle, Apr 11 2025, 8 findings, 8 resolved / 0 accepted / 0 mitigated, max Medium, 95.83% branch coverage, documented risk that forceNAVUpdate() could arbitrarily manipulate NAV if the emergency updater turns malicious. Audited token contracts are USDRE, USDRWA and ReToken; reUSD and reUSDe are not named in any Hacken scope. Certora formal verification Sep 2025 (protocol docs). Reserve attestations: The Network Firm daily AUP, most recent full report October 2025.
  • Admin/governance: TimelockController 0x69dDEa332723cF5407151aAF68B9b076557FCA93, getMinDelay() = 172800 verified live this run, no bypass, holds DEFAULT_ADMIN_ROLE and UPGRADER_ROLE. Governance MPC 3-of-5 for upgrades. Controller wallets: Oracle Config 0x49BC5A880f77247A348764DdB95951cd9212A0ee (3-of-5), Redemptions Config 0xEE16bE0374f2eFb34218affC1a8EbEe9310c47f8 (3-of-5), Access Manager 0x80a62B72dF1136aCBc57141FB67Aa46812fECAFc (5-of-8), Custodian Manager 0x9b6d7f2de2E4569297C7e88531E47679cEbE6eC9 (3-of-5). Fireblocks MPC 3-of-5 for custody sweeps. Guardian emergencyWithdraw on the Daily Instant Redemption Vault. Pause via Access Manager plus Redemptions Config; unpause requires exec and risk approval plus a further 48h delay. Signer identities refresh.
  • Oracle: daily NAV push 00:00 UTC, monotonic, daily up-cap 20%/365, per-update deviation enforced on-chain, emergency updates require 3-of-5 and are subject to a 4-hour minimum interval. sUSDe spot oracle capped at intrinsic rate; hard revert if USDe spot is below $0.99. reUSD NAV oracle 0x72b5760cfbe437dd01409f44055fdfb8f8121b46, reUSDe 0x7844441f13C61CBD3d297F5d2D7fcaBf17f0b843, reserves reader on Avalanche 0xc79a363a3f849d8b3F6A1932f748eA9d4fB2f607.
  • Pendle PT-reUSD-10DEC2026: market 0x13285bcbc27f92b47b4edb99d744c07b48c977c0, PT 0xeCfaFdC7741323a945A163ed068B5a3C43483957 (6 decimals), SY-reUSD 0x9487bd5a3b16ecb5f3184453e3ee75b800141648, accounting asset USDC (0.99987), underlying reUSD. AMM liquidity $7,570,488, implied APY 10.297%, underlying 6.370%, PT discount 3.781%, PT price $0.9621. Live $130k sell_pt quote: 0.159% price impact, $105.87 fee, 1.7% of pool. PT total supply 78,324,514.64 read on-chain, about $75.36M, of which $41.28M sits as Morpho collateral.
  • Morpho (2026-07-19): PT-reUSD-10DEC2026/USDC LLTV 91.5% (curator brief said 92%), supplied $40,161,191, borrowed $34,749,514, idle liquidity $5,411,678, collateral $39,918,605, utilisation 86.53%, borrow APY 8.04%, rate at target 8.29%, bad debt $0. PT-reUSD-10DEC2026/USDT collateral $1,362,544. Comparable markets: reUSD/USDC LLTV 91.5% at 5.02% borrow ($25.41M supplied), reUSD/USDT 91.5% at 4.26%, reUSDe/USDC LLTV 77% at 8.93%. Rate path at full utilisation refresh.
  • Secondary depth (Curve API, 2026-07-19): reUSD/USDC factory-stable-ng-466 $461,555; reUSD/sUSDe 534 $2,336,854; reUSD/sUSG 909 $238,414; reUSDe/sUSDe twocrypto-169 $553,949. Fluid reUSD/USDT LP about $14.27M (Re points API). Caution: the $6,990,862 "reUSD/scrvUSD" pool (factory-stable-ng-392) is Resupply's reUSD 0x57aB1E0003F623289CD798B1824Be09a793e4Bec, a different token.
  • Holder concentration (gap closed): Re points API, 2026-07-18. Direct reUSD $6.38M / 727 wallets on Ethereum plus $457,558 / 1,073 across Avalanche, Base, Plasma, BSC, Ink, Arbitrum and Katana; direct reUSDe $2.56M / 464. Morpho reUSD collateral $25.52M / 35; Fluid reUSD collateral $30.30M / 256; Fluid reUSD/USDT LP $14.27M / 31; Curve reUSD/sUSDe LP $2.34M / 34; Morpho reUSDe collateral $631k / 6; Pendle YT reUSD Dec2026 notional $82.09M / 158. Total participants 3,732. Etherscan tokenTopHolders returned NOTOK (API Pro required) this run, so per-address rankings are refresh.
  • Counterparty and legal: Cover Reinsurance SPC Ltd, Cayman Class B(iii), CIMA-regulated, acting for Segregated Portfolio SP1. Resilience (BVI) Ltd (incorporated Oct 23 2024) issues the tokens; Resilience Foundation (Cayman, incorporated Oct 7 2024) governs and acts as agent for holders; Resilience Inv SPC purchases the notes through separate portfolios (1 for reUSD, 2 for reUSDe). $100M Principal-at-Risk Note program issued April 2 2026, matures April 2 2031, coupon 30-day average SOFR plus 500bps (M1) and 950bps (M2), unsecured and limited-recourse to SP1 only, subordinated to cedent and policyholder claims, CIMA-approved. Token holders have no direct legal claim on Section 114 trust assets. Drawdowns and repayments emit NoteDraw / NoteRepay via the Surplus Note Registry.
  • Underwriting (protocol-reported): June 30 2026 portfolio $510.5M across 48 programs, 49 US states, 700,000+ policyholders; $149M (29%) written in the prior 45 days; homeowners share up from 10% to 18%; capital position grew $77.45M in June, nearly 16% over May. Docs report a 92% all-lines combined ratio to date and a sub-100% ratio every underwriting year since inception, and a $358M premium mix of small business commercial 41%, commercial auto 28%, workers' compensation 18%, homeowners 12%, personal auto 1%. Stress modelling (Nov 2025 LP memo): reUSDe impairment about 0.9% at a 115% combined ratio, reUSD about 0.03% at 135%. Reinsurer equity refresh; our prior "$77M equity" figure conflated the June capital-position increase of $77.45M, now confirmed as an increase rather than a level.
  • $RE: 159,600,000 circulating of a fixed 1,000,000,000 (16%), price $0.396, market cap $63.2M, ATH $1.078 on 2026-06-20 (CoinGecko, 2026-07-19); public launch June 2026 across 27 venues. Ecosystem allocation 50%, stake-to-vote, explicitly no claim on revenue, reserves, collateral or insurance flows. Points-allocation vesting: 100% at TGE at or below 150M points, roughly 10% to 37% for larger claimants, remainder in six six-month tranches over 36 months with a TVL holding requirement on tranches one through four.
  • Recent-news scan (2026): no Re entry on the DeFiLlama hacks list (588 entries checked). No reUSD or reUSDe depeg found. The February 4 2026 SingleAdapterRouter exploit at 0x169a5effcae91ab33bc9e97f49b513b81008c453 (verified source tree strategia_ETH/REUSD/Vault_reUSD.sol, about $25.8k) contains none of Re's token or ICL addresses and is not a Re-published contract; attribution refresh. The June 2025 Resupply exploit hit an unrelated token also named reUSD. Maintained monthly. Methodology: DeFi Research Instruction v2.

← all protocols