We build kUSD, a synthetic dollar issued on Base, so Frax USD is a peer of ours and in some venues a competitor, not a customer. Read what follows as a technical recomputation, not a verdict from a neutral referee. Kerne is not an auditor, a rating agency, or a regulator, this report scores nothing and ranks nothing, and it is not legal advice, not a compliance determination, not the statutory examination, and not an audit or a solvency opinion. It reads frxUSD against the public statutory criteria of the GENIUS Act (Public Law 119-27, enacted July 18, 2025) and against public on-chain data.
We picked frxUSD for a plain reason: it is one of the more transparent tokenized-dollar designs in its size class, so a fair read of it is mostly a record of what it does well, with a few specific, checkable gaps rather than an indictment. That is the shape a gap report should take. This is the same six-criterion readout we produce for a paying issuer at the GENIUS gap report, run here in full on a token that is not a customer, so you can see exactly what the deliverable is.
Every figure below is sourced. The on-chain numbers we read ourselves, and we give you the command to reproduce them. Where a point turns on off-chain data or an unsettled legal question, we say so instead of guessing.
Criterion one: is it in scope, and who is the issuer
The threshold question the Act turns on is whether a token is a payment stablecoin at all, meaning it is redeemable by its issuer at a fixed value. frxUSD points to the in-scope side of that line more cleanly than a yield-bearing synthetic dollar does. It is built as a fully collateralized dollar, redeemable one to one through governance-approved custodians, and the base token itself pays no interest, a point we return to below. A yield-bearing synthetic dollar, by the prevailing reading, may sit outside the payment-stablecoin definition; frxUSD does not obviously do so.
The harder half is the identity of the issuer, and it is a genuine one under a statute that contemplates a single permitted issuer. At the token level, frxUSD is minted and redeemed through the Frax Protocol, with the right to mint gated to a set of governance-approved custodians, each onboarded by a Frax Improvement Proposal with a supply cap. But governance did not leave the legal issuer undefined. Through a ratified proposal, FIP-432, the Frax DAO delegated frxUSD's compliance, collateral management, and regulatory responsibilities to Frax Inc, formerly FinresPBC, whose legal name is Financial Reserves and Asset Exploration Inc., a Delaware public-benefit corporation that public reviews describe as a prospective applicant for a federal payment-stablecoin charter. So the issuer is layered: a protocol and a DAO at the token layer, and a named Delaware corporation as the regulated entity. Which of those the Act treats as the issuer is a legal question for Frax's counsel, not one this report decides.
Criterion two: what actually backs it
frxUSD is asset-backed, a full break from the algorithmic FRAX it succeeded. What backs it is a basket of tokenized cash-equivalent funds held through regulated custodians, not a single asset. The headline launch partner in January 2025 was BlackRock's USD Institutional Digital Liquidity Fund (BUIDL), tokenized by Securitize, a fund of US Treasury bills, overnight repo, and cash. BUIDL is not the whole story, though, and calling it the dominant backing would be wrong. Independent reviews by LlamaRisk and Chaos Labs describe BUIDL together with Superstate's USTB as roughly ninety percent of the backing, with the governance-approved set also including Superstate's USCC, WisdomTree's WTGXX, Circle's USDC, Agora's AUSD, and Centrifuge's JTRSY. The basket is deliberately diversified to limit exposure to any one issuer.
Here is where the verifiability line falls, and it is the crux of the whole report. The reserve assets are themselves ERC-20 tokens, and a portion sits in on-chain custodian vault contracts, so part of the backing is visible on-chain. But a material share is held off-chain, as book entries on regulated custodians' ledgers, and the assets inside each tokenized fund, the actual Treasury bills, repo, and cash, live entirely off-chain on the fund managers' books. There is no single on-chain address a holder can read to recompute total backing. The authoritative composition figure is published by the reserve manager and cross-checked by a third-party feed rather than re-derived by the holder. That is a common and legitimate structure. It is also precisely the boundary a treasury desk needs mapped before it sizes a position.
Criterion three: monthly reserve-composition disclosure
Section 4 asks for the composition of reserves, published monthly, in a form a reader can check. On cadence and on-chain presence frxUSD clears that bar and then some. The FraxFacts dashboard publishes reserve data on-chain that updates continuously, and Chaos Labs runs an on-chain proof-of-reserves feed that validates cross-chain supply against reserves and publishes independent collateralization checks. That is ahead of a monthly document.
The gap is in the nature of the authoritative statement rather than its frequency. Per LlamaRisk's review, the full reserve-composition breakdown is self-reported by the reserve manager as spreadsheets, not as an assurance report from an accounting firm. So a reader gets a live on-chain feed plus a self-published composition, which is a lot, and the composition itself is still the issuer's own statement rather than an independently examined one. What is checkable on-chain today is the token supply and the on-chain portion of the reserve tokens. What is not is the off-chain balance and the fund-level holdings, which arrive as reported figures.
Criterion four: examination and certification
The Act's monthly requirement is an examination by a registered public accounting firm, an attestation engagement, together with executive certification. It is not a full monthly audit, and it is not the same as a holder re-deriving a number. On this criterion frxUSD's picture is mixed and worth stating precisely, because it is easy to get wrong. The underlying tokenized funds carry their own auditors: BUIDL is audited by PwC, and Superstate's USTB and WisdomTree's WTGXX by Ernst and Young. The frxUSD contracts themselves were reviewed by the security firm Zellic in 2025. At the token level, though, per LlamaRisk, no independent accounting firm has yet issued an assurance opinion on the reserve manager's own statements; those remain self-reported. One monthly accounting attestation that circulates in coverage, by the firm BPM, attaches to USDB, a stablecoin held as one reserve component, not to frxUSD itself, and even that USDB attestation is reported as not yet publicly released. So as of mid-2026, frxUSD rests on audited underlying funds and a self-reported issuer-level composition, without a publicly disclosed monthly examination of the issuer's own reserve statement by a registered public accounting firm. That is a specific, nameable gap against the statutory shape, and it is one an issuer can close.
Criterion five: the issuer-yield question
Section 4(a)(11) bars the issuer from paying interest on the stablecoin. frxUSD is built to sit on the right side of that line: the base frxUSD token pays holders nothing. Yield reaches a holder only if they move into sfrxUSD, a separate staked-wrapper token, and through the protocol's strategies and AMOs. That base-token-plus-staked-wrapper split is the single most common way tokenized dollars handle the yield prohibition, and it is also exactly where the law is unsettled: whether a staked wrapper issued by the same protocol, or the rewards routed to it, count as the issuer paying interest is an open interpretive question that the final rules may sharpen. frxUSD keeps the payment token clean; the gray area is the wrapper, and we flag it as gray rather than pretend it is settled.
Criterion six: what a holder can re-derive
This is the axis our free tools grade every dollar on, and it is where frxUSD lands as a genuine hybrid. What a holder can recompute alone, right now, without trusting anyone, is the supply. frxUSD is a LayerZero omnichain token, natively minted and redeemable on two base chains, Ethereum and Fraxtal, and represented as a bridged token elsewhere, so the honest supply figure is a cross-chain sum. We read the two largest legs ourselves. On Base, the frxUSD contract at 0xe5020a6d073a794b6e7f05678707de47986fb0b6 reported a total supply of about 368,595 frxUSD at block 48,590,836. On Ethereum, the contract at 0xcacd6fd266af91b8aed52accc382b4e165586e29 reported about 104.96 million at block 25,525,807. Anyone can reproduce either read with one call, for example on Base:
cast call 0xe5020a6d073a794b6e7f05678707de47986fb0b6 "totalSupply()(uint256)" --rpc-url https://mainnet.base.org
Divide by ten to the eighteenth for the human figure. Summed across chains, that is an aggregate near 108 million dollars as of July 13, 2026, in line with the roughly 108.5 million that CoinGecko and CoinMarketCap report and the roughly 107 million DefiLlama attributes; the three differ by method, which is worth knowing before you cite one. What a holder cannot recompute alone is the backing. The off-chain custodial balances and the assets inside each tokenized fund are read through the reserve manager's disclosure and the third-party proof-of-reserves feed, not off a single contract. For frxUSD the verifiable-versus-trusted split is favorable by the standards of tokenized-RWA dollars, and it is still a split, which is the whole reason to name it rather than wave at it.
What this report is and is not
This is a technical recomputation of frxUSD against the GENIUS Act's public criteria and public on-chain data, published as an example of the report Kerne produces for issuers. It is not legal advice, not a compliance determination, not the statutory examination, and not an audit, a rating, or a solvency opinion. A token appearing here is not endorsed by us and is not a customer of ours, and Kerne is not affiliated with Frax or with any custodian or auditor named above. Every factual claim is drawn from public sources: the token contracts on Ethereum and Base, Frax's documentation at docs.frax.com/frxusd, the public LlamaRisk pegkeeper review and Chaos Labs token review, and the underlying funds' own disclosures. Where those sources disagree on a number, we said so.
You do not have to trust this reading. The free tool at verify any stablecoin runs the on-chain half of it on frxUSD, or on us, in your own browser. Kerne is early and not yet externally audited; our first external review, by Hexens, has fieldwork underway as of mid-July 2026 and is not complete. We hold our own kUSD to the same bar we apply here, and we would rather you check us than take our word for any of it.
Verify it yourself
Run the same check on any reserve, or have it run for you.
Paste any issuer's signed attestation into the free verify tool and recover the signer, rehash the figures, and check freshness in your own browser. For a machine-signed, point-in-time read of an address you name, delivered on the page in about two minutes, the instant self-serve read is $29; a human-reviewed read is $149. A teardown like this one, commissioned on any target you name, is $499. An independent read of a counterparty you hold or allocate to is $2,500. Attestation tooling, not an audit, and not a solvency opinion.