Frax USD (frxUSD) — Risk Report
Moderate risk · 5.5/10 · Strong Treasury backing, undercut by unaudited upgradeable contracts
This report covers frxUSD (
0xCAcd6fd2…586E29), Frax Finance’s current flagship stablecoin — NOT Legacy FRAX (0x853d955a…1b99e). They are separate assets with separate balance sheets and separate contracts. The 1:1 FRAX→frxUSD migration was closed in April 2025 (FIP-430). If you’re looking at the deprecated, under-collateralized predecessor, see the Legacy FRAX report instead.
frxUSD is Frax Finance’s flagship payment stablecoin, launched April 2025 as part of the “North Star” upgrade. It is backed 1:1 by tokenized US Treasuries held with regulated custodians, and it replaced legacy FRAX as the primary Frax dollar. The asset itself has performed well — the peg has held tightly since launch — but the structural risks sit on the contract and exit side, not the backing.
| Peg | Yield | Exit | Status | Chains |
|---|---|---|---|---|
| $1.00 (trades ≈$0.9993) | None (stake to sfrxUSD for yield) | Per-custodian redemption; USDC exit gated by a ≈$10M buffer | Active flagship, float contracting | Ethereum + 10 chains (LayerZero OFT) |
Backing & reserves
The backing is genuinely high-quality — tokenized US Treasuries — but concentrated and not independently audited at the token level.
frxUSD is collateralized by a basket of tokenized Treasury products: BlackRock BUIDL (via Securitize), Superstate USTB and USCC, WisdomTree WTGXX, plus USDC (Circle), AUSD (Agora), and JTRSY (Centrifuge). Per LlamaRisk’s July 2025 review the collateral ratio was about 103.7%, and despite the BlackRock-forward marketing, Superstate funds (USTB + USCC) represented over 90% of backing in aggregate at that time. T-bill quality is excellent; the concentration in a single fund manager is the risk vector. The current public framing emphasizes custodian diversification, but the live split could not be independently re-verified on this pass — treat the ≈90% Superstate figure as the last confirmed reading, not a guarantee of today’s mix.
A consolidated proof-of-reserves feed now exists, but we still have not been able to read it. frxUSD integrated Chaos Labs Proof of Reserves in late 2025 to publish on-chain collateralization attestations as the token expands cross-chain. This is a real transparency improvement over launch — but it has now failed verification on two consecutive passes, by two different failure modes: on 2026-06-08 the frxUSD PoR card rendered empty (”—”) while sibling feeds (USDe, AUSD, USDT0) populated normally, and on 2026-08-13 oracles.chaoslabs.xyz did not resolve at all from our environment. That second failure may well be transient or local to us — but the practical status is unchanged, and it has been unchanged for two months: the PoR “exists but is unverified.” The fallback remains the underlying-fund attestations (EY for USTB/WTGXX, PwC for BUIDL). There is no third-party audit of the frxUSD stablecoin contracts themselves (see Contracts below) — the audits cover the underlying funds, not the token.
Exit liquidity & redemption
This is the most underappreciated risk: redemption is fragmented, and clean USDC exit at scale is shallower than the headline suggests.
Redemptions run through Frax’s deposit/redemption coordinator, and they are non-fungible across custodians — each custodian redeems into its own asset (BUIDL → BUIDL, USTB → USTB, and so on). A retail holder who wants to exit to USDC is funneled through a ≈$10M Superstate USDC buffer. Beyond that buffer, large redemptions push you into BUIDL/USTB shares and a separate institutional off-ramp. In practice, USDC-at-par exit liquidity is much thinner than the ≈$105.5M float implies.
On the secondary market, liquidity has improved but is still mid-tier:
| Venue | Notes |
|---|---|
| Curve (Ethereum) | The DEX center of gravity, but paired more against FRAX / sUSDS than USDC — so deep USDC-exit-at-scale is the residual constraint. |
| Aave V4 | frxUSD became a default borrowable asset on Aave V4 on 2026-04-06, adding a major lending venue and on-chain depth. |
| Cross-chain (LayerZero OFT) | Expanding across 10+ chains, which broadens reach but fragments depth. |
| Tier-1 CEX | None yet. No major centralized listing. |
Daily volume is now about $10.8M (roughly 3.7× the ≈$2.9M seen in April), so the trajectory is improving — but for a sized exit, the binding constraint is still the USDC-pairing depth, not the headline TVL.
Peg performance
frxUSD has held its peg tightly since launch — this is the strongest part of the picture. Per LlamaRisk, the maximum deviation has been about $0.00985, with only one wick beyond 0.5% (June 26, 2025) that recovered within hours. Hourly volatility has been tighter than USDe. Current price is about $0.9993 (roughly 7bps below par, 2026-08-13). The risks in this report are structural (contracts, redemption), not operational — the token has done what a stablecoin should so far.
Contracts & admin
This is where the score is lost. Three things compound:
- No third-party audit of the frxUSD stablecoin contracts. LlamaRisk states this explicitly. The contracts are upgradeable proxies. Frax’s in-house “Security Cartel” reviewed the FIP-430 upgrade path, and ChainSecurity audited the FXB-side upgrade — but the frxUSD ERC-20 itself has no public third-party audit report. For an upgradeable stablecoin, that is the single biggest contract-level risk.
- A multisig with no timelock controls upgrades — now 4-of-7, previously 3-of-5. The owner (
0xfFFffF4F3baC444b2C0ecf2A1840d018bE783937, verified on-chain 2026-08-13: a 4-of-7 Safe) can pause, upgrade the implementation, and modify parameters with zero delay. Combine that with “upgradeable + unaudited” and you have a high-trust configuration. - An unconfirmed December 2025 “stealth patch” allegation. A single-source Medium post (Donnyoregon) claims Frax silently deployed a contract patch between Dec 5–16, 2025 to fix a zero-value-ticket vulnerability without crediting the bounty submitter, with on-chain bytecode reportedly diverging from the verified Etherscan source; Token Sniffer flagged it. No Frax public response has surfaced. We can’t confirm the specific claim — but the 4-of-7 / no-timelock / upgradeable / unaudited setup is exactly what would enable such an action silently, which is why it’s worth flagging.
The owner address changed, and we found it by re-reading the chain
A previous version of this report named 0xB174…3f27, a 3-of-5 Safe, as the controlling multisig. That is no longer true. Verified on-chain 2026-08-13:
| Previously published | Verified 2026-08-13 | |
|---|---|---|
| Owner | 0xB1748C79709f4Ba2Dd82834B8c82D4a505003f27 | 0xfFFffF4F3baC444b2C0ecf2A1840d018bE783937 |
| Configuration | 3-of-5 Safe | 4-of-7 Safe (v1.3.0) |
| Timelock | none | none — unchanged |
| Signer overlap | — | zero (isOwner() on the new Safe returns false for the old one) |
The old Safe still exists and is still 3-of-5; it simply no longer controls frxUSD.
There are two honest readings here and you should hold both.
The direction is better. Four independent keys out of seven is a higher bar to compromise than three out of five, and key compromise is the most likely failure mode for an upgradeable, unaudited stablecoin contract. That is a real improvement, and it is why the Issuer score moves from 5.0 to 5.5.
The manner is not reassuring. This was a complete, zero-overlap replacement of the controlling signer set, on a contract that is an upgradeable proxy with no public third-party audit and no execution delay — and we found it by routinely re-reading the chain, not from any announcement or governance record. We could not locate the record authorising the handover. That absence does not mean no such record exists; it means we cannot distinguish “a maturing issuer widening its multisig through proper process” from “a signer set replaced quietly” using on-chain state alone. Both are consistent with what we can see. Treat the improvement as provisional until the authorising record surfaces.
The binding criticism is unchanged. A timelock, not a wider signer set, is what would actually fix this configuration. Every property that made the December 2025 stealth-patch allegation credible is still present today.
Frax Finance itself is an established team (Sam Kazemian, 5+ years, active development on frxUSD, Fraxtal L2, and frxETH) — the issuer-level track record is real. The contract-trust profile is the offsetting concern.
Growth & adoption
The float is now contracting, not flat. ≈$65M (Jul 2025) → ≈$125M (Apr 2026) → ≈$124M (Jun 2026) → ≈$105.5M (Aug 2026) circulating — down roughly 15% in nine weeks. A previous version of this report called this “plateaued”; that framing is now too kind.
What makes it worth a second look is the combination: the float is shrinking while integrations keep broadening (Aave V4 as a default borrowable asset, more OFT chains). A stablecoin adding venues and losing float usually means the new venues are not where the demand is. More than a year after launch it remains a sub-top-50 stablecoin, well behind newer entrants like USDS and USDe.
(Circulating figures are DefiLlama’s Frax USD series — the one to watch is frxUSD, id 235. Legacy FRAX is a different token with a different balance sheet under id 6; the two are frequently confused.)
Who it’s for · Who should avoid
Reasonable for:
- Frax-ecosystem users who want Treasury-backed dollar exposure and value the tight peg, sized to the ≈$10M USDC-buffer reality (i.e. not relying on instant at-par USDC exit for a large position).
- Holders comfortable with the contract-trust trade-off in exchange for high-quality T-bill backing.
Avoid / size down if:
- You need deep, instant, at-par USDC exit for a sized position — the redemption fragmentation and ≈$10M buffer are the binding constraint.
- You weight unaudited upgradeable contracts under a no-timelock multisig heavily — 4-of-7 is better than the 3-of-5 that preceded it, but instant execution is the structural risk here, and that has not changed.
- You’re confusing it with Legacy FRAX — they are different assets (see the disambiguation note at the top).
What to watch
- Chaos PoR feed. If the frxUSD card on the Chaos dashboard starts populating with live numbers, that closes the current “exists but unverified” gap and is a genuine backing-transparency uplift.
- Backing concentration. The ≈90% Superstate exposure (USTB + USCC) vs the “diversified custodians” marketing — watch for the live split to be re-confirmed either way.
- Multisig / upgrades. The 4-of-7 owner (
0xfFFf…3937, in place as of 2026-08-13) can upgrade with no delay. Any implementation upgrade, another owner change, or resolution of the Dec 2025 patch allegation would be material — as would a timelock, which is the change that would actually move this score. - The authorising record for the August 2026 owner migration. A Frax governance post or vote documenting the handover to the 4-of-7 Safe would resolve the open question above. Its appearance would firm up the Issuer score; continued absence keeps the improvement provisional.
- USDC-exit depth. The Superstate USDC buffer and Curve USDC-pairing depth are the real exit constraint — more than headline float.
- FXB redemption asset. Note that Frax Bonds (FXBs) currently redeem to Legacy FRAX on mainnet, not frxUSD — the Fraxtal upgrade making frxUSD the FXB underlying was still being audited at last check. Verify before relying on it.
This report describes frxUSD as of August 2026, based on public Frax/LlamaRisk documentation and direct on-chain reads (owner, threshold, signer overlap and supply re-verified 2026-08-13). Frax Finance has not engaged on this report. The backing sits partly off-chain with regulated custodians and tokenized-fund issuers; figures rely on those issuers’ attestations plus on-chain data. The Chaos Labs PoR feed could not be retrieved on either of the last two passes. Corrections welcome at info@tidresearch.com.
Revision history: 2026-08-13 — admin migration found on-chain: the frxUSD owner moved from the 3-of-5 Safe 0xB174…3f27 to a 4-of-7 Safe 0xfFFf…3937 with zero signer overlap and still no timelock; no authorising governance record located. Issuer 5.0 → 5.5, overall held at 5.5. Circulating restated ≈$124M → ≈$105.5M (down about 15% in nine weeks), reframing “plateaued” as contracting. Price ≈$0.9997 → ≈$0.9993. Chaos PoR unverified for a second consecutive pass, this time unreachable rather than empty. 2026-06-10 — initial publication (multisig verified 3-of-5 on 2026-06-08; liquidity 5.0 → 5.5 on the Aave V4 listing).