NUSD Is Shrinking. Can You Verify Its Reserves Yourself? No, You Trust the Attestor.

The most useful moment to ask whether you can verify a synthetic dollar is the moment money is leaving it. A calm book hides the question. A shrinking one sharpens it, because the holders still inside are the ones deciding whether to stay, and the honest basis for that decision is whether they can check the backing themselves before they move, or whether they are trusting someone else's word for it.

So this is a live read of Neutrl's NUSD, a synthetic dollar that has fallen from a February peak near 230 million dollars to about 80 million, roughly two-thirds of it gone, with around 44 percent leaving in just the last thirty days. We picked it because it is a genuinely sophisticated example of the boundary this series keeps finding, not because it has a problem. We are not claiming NUSD is under-backed. Neutrl's disclosures and its attestor say the opposite, and we have no evidence otherwise. The only question here is epistemic: which parts of NUSD's backing can a holder verify without anyone's permission, and which parts require trusting a third party. In plain terms, if capital is moving, can you check the reserves yourself first?

The rules we hold ourselves to, the same ones from our Falcon USDf walk: name Neutrl's real method accurately, allege no wrongdoing, and end on the same boundary in our own product. Neutrl is also one of ten names in our neutral field survey; this is the dedicated, single-name walk that the survey only had room for a paragraph on.

What NUSD is, and why the walk is interesting

NUSD is a delta-neutral dollar with an unusual collateral engine. Instead of holding liquid majors, Neutrl acquires discounted or locked altcoins over the counter, often tokens bought below market from funds or foundations that need liquidity, and hedges each position with an equal short perpetual so the dollar value is intended to stay flat regardless of where the token trades. Liquid stablecoin reserves sit alongside as a buffer. The staked form, sNUSD, passes through the strategy's yield to holders who opt in. It is a clever way to turn OTC altcoin discounts into a dollar-denominated return, and it is also, for our purposes, close to the hardest possible case for holder-side verification, which is exactly why it is worth walking.

The context that makes the walk timely is the outflow. As of June 30, 2026, NUSD's on-chain-tracked supply is about 80 million dollars, down from a peak near 230 million in mid-February, a fall of roughly two-thirds, with about 44 percent of it leaving in the trailing thirty days by DefiLlama's daily series and RWA.xyz's tracker; CoinGecko shows a similar figure near 73 million. One caveat worth stating plainly, because we would want it stated about us: a few exchange and aggregator pages still show a stale, self-reported number two to three times higher, around 205 to 218 million dollars, which is not consistent with the on-chain-tracked series, and we use the lower, chain-tracked figure. Whatever the exact mark, the direction is not in question. Capital has been leaving NUSD for months, and the pace has picked up, not slowed. That is the setting in which the verification question stops being academic.

What you can read for yourself, and where that stops

Start with the part that needs no one. NUSD and sNUSD are tokens; their supply and the vault's share-to-asset ratio are on-chain values you can pull directly, the same way you would for any ERC-20 and ERC-4626 vault. If money is leaving, you can watch the supply fall in real time without asking anyone. That is the liability side, the claim on the reserves, and it is fully legible.

Now turn it around. What backs those tokens? A basket of over-the-counter altcoins, some of them discounted or still vesting, held and hedged across custodians and centralized perpetual venues. None of that lives at an address you can open in a block explorer and read. An OTC-acquired locked token is, almost by definition, not a public on-chain balance you can independently price and confirm; a short perpetual hedge is an open position on an exchange, not a chain state. The moment you move from "how much NUSD exists" to "what is behind it," you stop reading the chain and start reading Neutrl's measurement of assets that are not on it.

Who does the verifying, and how far it gets you

If a holder cannot re-derive the assets, who does? Neutrl's answer is one of the more advanced setups in the field, and it deserves to be named accurately. Reserve transparency runs through Accountable's zero-knowledge Data Verification Network, integrated before launch, which reads balances from Neutrl's custodians, exchange accounts, and contracts and publishes a live Proof of Solvency built from Merkle sum trees and secure enclaves that you read on Neutrl's dashboard. A separate real-time monitor, Hypernative, watches the system and holds a pauser role. This is a real cryptographic attestation pipeline, not a screenshot of a spreadsheet, and it is meaningfully better than the dashboards many synthetic dollars ship with.

Here is the precise boundary, and it is not a criticism, it is a description. A zero-knowledge Proof of Solvency proves that an oracle network observed a set of balances and that they satisfied a solvency relation, without revealing the underlying accounts. What it does not do, and does not claim to do, is let you re-derive those balances yourself. You are verifying that the attestor's network saw what it says it saw and did the arithmetic correctly. You are trusting that the accounts it read are the complete and correct set, that the off-chain positions are what the feed reports, and that the OTC tokens are valued the way the model values them. The proof is cryptographic; the inputs to it are attested, not open. That is a genuinely strong version of "trust an attestor." It is still, structurally, trust an attestor.

The exact point you must trust someone

Put the walk together and the handoff is clean. Up to the supply of NUSD, you trust no one. At the reserves, you take on a set of things at once, none of them alleged to be wrong, all of them real dependencies:

1. The attested account set is complete. The Proof of Solvency covers the accounts the network is pointed at. You are trusting that this is all of them, and that the custody and exchange balances exist and are controlled by Neutrl, as the attestor observed them.
2. The OTC and hedge legs hold their reported value. Discounted or locked altcoins and their short-perp hedges are not passive on-chain collateral. Their mark depends on a valuation model and on exchange positions no holder can read. When the collateral is inherently illiquid and off-chain, the gap between "attested value" and "value you could realize under stress" is a thing you are trusting, not checking.
3. The snapshot still holds now. Between proofs, you are trusting that nothing material moved since the last one.

Neutrl has also had the ordinary operational scares that make holders care about what they can independently check. In February 2026 an attacker used a flash loan to exploit an access-control flaw in Neutrl's peripheral Pendle Standardized-Yield integration, a permissionless internal-balance burn in the redeem path, and drained about 45,900 dollars from NUSD and USDC liquidity providers; it was a smart-contract bug in an integration, not a shortfall in the OTC backing, and it hit the Pendle wrapper rather than Neutrl's core issuance contracts. In March 2026 the project separately suffered a DNS hijack of its front end, a website-layer attack that did not touch the contracts or the reserves. Neither was a backing shortfall, and we present them as what they were. They are the kind of event where, if the reserves are something you can only read through an attestor, your tools in the moment are that attestor's last proof and the depth of a market, and not a check you can run yourself.

Kerne's own boundary, disclosed here

It would be dishonest to run this walk on Neutrl and not on ourselves. Kerne issues a synthetic dollar too, kUSD, and it has a boundary of the same class. The difference is not that we are "better attested." It is that on the reserve leg, kUSD is attestor-optional where NUSD is attestor-required.

kUSD's collateral lives on Base and is readable with raw on-chain calls, and we publish an hourly Proof of Reserves signed with an EIP-191 key that you can re-derive yourself against the chain, with no third-party attestor standing between you and the on-chain figure. You can run it at /verify and read the breakdown at /api/por. Then comes our own private point, and it is the direct analogue of Neutrl's off-chain legs: kUSD's delta-neutral hedge runs on Hyperliquid, a single venue, and those positions are self-reported and signature-bound rather than independently re-derivable on-chain. That is our attestor-equivalent, the exact line where a kUSD holder stops verifying and starts trusting us, and an independent third-party attestation of that leg is being scoped. We are also pre-audit and at Genesis scale, with a live modeled skUSD yield, currently in the low teens while funding is positive and normalizing toward the high single digits across a full funding cycle, rather than a large realized distribution. We did not remove the trust boundary that every off-chain-hedged synthetic dollar has. We moved it to the smallest surface we could and named it on the same page as the part you can check. The whole field, including where kUSD ranks and where it does not, is in the synthetic-dollar scorecard, and our own open gaps live at /legible.

What proof of solvency actually proves

None of this makes NUSD a bad synthetic dollar. A zero-knowledge Proof of Solvency with an independent real-time monitor is, by the standards of the field, a strong transparency setup, and the OTC-discount engine behind it is genuinely inventive. The point is narrower, and it is the same point every post in this series lands on. "Proof of solvency" and "proof of reserves" almost always mean a network or a firm proved it, from inputs you cannot open, and you are reading their result. That is valuable. It is not the same thing as a holder re-deriving the number from the chain whenever they want, and when capital is on the move, the difference between those two is the difference between checking and hoping. If you hold NUSD, or anything else, run this walk on it: read the supply yourself, then find the exact line where you have to trust the attestor, and decide with your eyes open which side of that line your money is on.

Figures are as of June 30, 2026 and nothing here is investment advice. NUSD's on-chain-tracked supply of about 80 million dollars, its fall of roughly two-thirds from a February peak near 230 million, and the roughly 44 percent decline over the trailing thirty days are per DefiLlama's stablecoin series, cross-checked against CoinGecko (about 73 million) and RWA.xyz; stale, higher self-reported figures near 205 to 218 million on some exchange pages are noted and not used. Neutrl's reserve model, OTC-acquired discounted or locked altcoins hedged with short perpetuals plus liquid stablecoin reserves, and its transparency stack, Accountable's zero-knowledge Data Verification Network and live Proof of Solvency with Hypernative as monitor and pauser, are per Neutrl's own documentation and Accountable's and Hypernative's writeups on the Neutrl integration. The February 2026 exploit of about 45,900 dollars in Neutrl's Pendle Standardized-Yield integration and the separate March 2026 DNS hijack are per the DarkNavy and SmartContractsHacking incident analyses and CryptoTimes; both were an integration bug and a front-end attack respectively, in the periphery rather than Neutrl's core contracts, and neither was a reserve shortfall. Kerne is not affiliated with Neutrl, Accountable, or Hypernative. Kerne's own claims resolve to live endpoints: the hourly signed Proof of Reserves at /api/por/signed, its on-chain leg at /api/por, and the live risk surface at /api/risk-status. A /verify pass proves an attestation is authentic and fresh; it is not an audit and not a solvency opinion.