Tokenized Asset Bankruptcy: How Digital Claims Get Treated

A token sitting in a wallet feels like ownership. It has an address, a balance, a transaction history you can verify yourself. That feeling turns out to be legally irrelevant the moment the entity behind that token files for bankruptcy protection. Courts do not look at the blockchain first. They look at contracts, custody records, and the Bankruptcy Code, and only then ask what the token was supposed to represent in the first place.

That gap between what a token feels like and what a bankruptcy court decides it legally is has already cost thousands of people real money. It is about to matter for a much bigger population: holders of tokenized private credit notes, tokenized receivables, tokenized real estate interests, and other on-chain claims on real-world assets that are no longer a niche experiment but a market measured in tens of billions of dollars.

Why This Question Stopped Being Theoretical

For years, “what happens to my tokens if the exchange goes under” was a hypothetical asked mostly by skeptics. Then Celsius Network, Voyager Digital, BlockFi, Genesis Global Capital, and FTX all filed for Chapter 11 within roughly a fourteen-month stretch between mid-2022 and late 2023, and hundreds of thousands of account holders got a very concrete answer. In nearly every one of those cases, people who believed they were simply parking their coins with a custodian discovered they were, in the eyes of the bankruptcy court, unsecured creditors standing in line behind lenders, professional fees, and tax claims.

The reasoning in those cases did not turn on blockchain mechanics at all. It turned on a handful of pages in each platform’s terms of service, on whether wallets were segregated or commingled, and on the Bankruptcy Code’s broad definition of “property of the estate.” That same reasoning now applies directly to a fast-growing segment of finance that has nothing to do with speculative crypto trading: tokenized real-world assets.

Tokenized private credit funds, tokenized trade receivables, fractionalized tokenized real estate, and tokenized short-term Treasury exposure have moved from pilot programs to a market that industry trackers now peg at well over twenty billion dollars in on-chain value, with private credit and Treasury products making up the bulk of it. Every one of those structures involves a token that is supposed to represent a claim on something real — a loan, a receivable, a slice of a building, a security. When the entity issuing or servicing that claim runs into financial trouble, the exact same custody-versus-ownership question that sank Celsius depositors resurfaces, except now the underlying asset is a mortgage pool or a private credit book instead of Bitcoin.

Regulators have noticed the gap. The 2022 amendments to the Uniform Commercial Code, which add a new Article 12 covering “controllable electronic records,” have now been adopted by a majority of U.S. states, giving lenders and token holders a long-missing statutory path to perfect a security interest in a digital asset the same way they would perfect one in equipment or inventory. That statute did not exist when Celsius or Voyager collapsed. It exists now, and it changes the calculus for anyone structuring or buying a tokenized claim in 2026.

What a “Tokenized Claim” Actually Represents in Bankruptcy Court

Before a court can decide who gets paid, it has to decide what the token is. That sounds obvious, but it is the single most contested issue in every one of these cases, and the answer is rarely written down anywhere the token holder actually read.

Token as Evidence vs. Token as Property Right

A token can function in two very different legal roles. In the first, the token is merely evidence of an off-chain legal relationship — a receipt. The actual property right (a promissory note, a membership interest in a special purpose vehicle, a fractional deed) exists in traditional legal documents, and the token is a convenience layer for tracking and transferring that right. In the second, the token itself is the asset, meaning transferring the token is legally sufficient to transfer the underlying right without any additional paperwork, the way Article 12’s controllable electronic record framework now permits.

Courts have consistently ruled that the label the platform used in its marketing material does not control. What controls is the actual contract language in the terms of use, the subscription agreement, or the note purchase agreement the investor signed. A platform that calls its product a “digital deed” in a pitch deck but writes “SPV retains full title and discretion over program assets” in the fine print will lose that argument for the token holder every time.

The SPV Wrapper and Bankruptcy-Remote Structuring

Most serious tokenized RWA programs route the underlying asset through a special purpose vehicle, or SPV, specifically to keep it out of the operating company’s bankruptcy estate if the parent ever fails. Done correctly, this works: the SPV owns the receivables or the real estate, issues tokens representing beneficial interests in the SPV, and maintains genuine separateness — its own books, its own board resolutions, no commingled cash with the parent.

Done poorly — and this happens far more often than issuers like to admit — the SPV shares service agreements, cash sweeps, or a servicer with the parent, and a bankruptcy court applying substantive consolidation doctrine can pull the SPV’s assets back into the parent’s estate anyway. The token holder’s recovery then depends on the parent’s overall solvency rather than the performance of the specific receivables pool they thought they bought into. This is the single most common structural weakness in tokenized RWA deals marketed as “bankruptcy-remote.”

Property of the Estate: How Section 541 Decides Who Owns the Token

Section 541 of the Bankruptcy Code defines “property of the estate” about as broadly as a statute can. It sweeps in essentially every legal and equitable interest the debtor holds at the moment of filing, wherever located and by whomever held. The practical question in every tokenized-claim bankruptcy is whether the debtor actually held a property interest in the tokens or receivables at issue, or whether it was merely holding someone else’s property as a bailee or custodian.

The Custody Language Test

Judge Martin Glenn’s August 2023 ruling in the Celsius Network case is the clearest illustration of how this plays out. Celsius’s “Earn” program terms of use stated that depositing crypto into Earn transferred “title” to Celsius, which could then lend, stake, or otherwise use the assets. The court held that language controlled: once title transferred, the crypto became Celsius’s property, and depositors became unsecured creditors of the bankruptcy estate rather than owners entitled to get their specific coins back. Celsius’s separate “Custody” program, by contrast, used language preserving customer ownership and was treated differently.

Voyager Digital’s case reached a similar outcome for similar reasons: its terms of use described customer crypto as a loan to the company in exchange for interest, which meant customers were creditors, not bailors, when the company failed. FTX’s Chapter 11 proceedings involved an even messier fact pattern, since customer assets were commingled with proprietary trading funds well beyond what any terms of service disclosed, which pushed the case toward fraud-adjacent recovery mechanisms rather than a clean custody analysis.

The lesson generalizes cleanly to tokenized RWA platforms: a note purchase agreement that says the issuer “may commingle proceeds,” “retains discretion to reinvest,” or “holds assets for its own account pending distribution” is doing the same legal work as Celsius’s title-transfer clause. Token holders reading a slick landing page that talks about “your asset, your token” need to read the actual subscription documents, because that page is not what a bankruptcy judge will cite.

Commingling and the Death of the Bailment Defense

Even where the paperwork says the platform is merely a custodian, courts have repeatedly found that operational commingling defeats the bailment defense in practice. A bailment requires an identifiable res — a specific, segregated thing the bailee holds for the owner. If a platform pools token-holder receivables into a single omnibus custody wallet, sweeps proceeds into a general operating account, or lends against the pool to fund its own working capital, the “specific asset held for you” argument collapses. The receivables become fungible with the debtor’s general assets, and fungible, commingled assets are property of the estate under Section 541, full stop.

This is why on-chain segregation matters far more than most token holders realize. A platform that mints one token per identifiable, on-chain-traceable receivable, held in a dedicated wallet with no discretionary re-lending, has a real shot at keeping those assets out of its bankruptcy estate. A platform that mints tokens against a pooled fund with discretionary reinvestment rights does not, regardless of what the cover page of its offering memorandum says.

UCC Article 12 and the Rise of the Controllable Electronic Record

Before 2022, the Uniform Commercial Code simply had no clean category for a digital asset like a token. Lenders trying to take a security interest in crypto or tokenized claims had to force the asset into ill-fitting categories like “general intangible,” which comes with weak, hard-to-perfect priority rules. The Uniform Law Commission’s 2022 amendments fixed that by creating a new Article 12 defining a “controllable electronic record” (CER) and a companion concept of “control” that functions as the digital equivalent of possession.

What “Control” Means for a Tokenized Claim

Under Article 12, a person has “control” of a CER if they have the power to enjoy substantially all the benefit of the record, the exclusive power to transfer it (or to share that power with a limited number of others), and the ability to identify themselves as having that power. In plain terms: whoever holds the private keys, or has contractual and technical rights functionally equivalent to key-holding, has control. Perfecting a security interest in a CER by obtaining control is analogous to perfecting a lien on a car by taking possession of the title, and it generally beats a security interest perfected only by filing a UCC-1 financing statement.

For tokenized RWA claims, this matters enormously. A lender that takes control of the tokens representing a receivables pool — rather than merely filing paperwork against the SPV’s general assets — jumps to a stronger, harder-to-dislodge priority position if that SPV later fails. Sophisticated warehouse lenders financing tokenized private credit deals have started building control-based security agreements into their term sheets specifically because of this.

Qualifying Purchaser Protection and Its Limits

Article 12 also introduces a “qualifying purchaser” concept, under which a buyer who gives value, obtains control, and takes without notice of a competing claim can take the CER free of most prior property claims, similar to good-faith-purchaser protections that already exist for negotiable instruments. This protection is powerful for someone buying a token on a secondary market, but it has a hard limit that trips up a lot of retail buyers: it only protects transfers of the CER itself under control-based rules. It does nothing to fix a defective underlying structure, and it does nothing for a token holder whose issuer never gave them control in the first place, which describes most retail-facing tokenized RWA platforms today. Most of these platforms retain the master private keys and simply issue account-level ledger entries to buyers, meaning the buyer never had “control” under Article 12 to begin with and the qualifying-purchaser shield never engages.

State-by-state adoption timing is also a live wrinkle. A majority of states have enacted the 2022 amendments as of 2026, but a handful of larger states adopted later or with variations, and choice-of-law clauses in note purchase agreements determine which state’s version of Article 12 (if any) actually governs a given deal. A token holder should know which state’s law their subscription agreement selects and whether that state has adopted Article 12 at all, because in a state that has not, the old “general intangible” categorization and its weaker perfection rules still apply.

The Bankruptcy Priority Waterfall for Tokenized Claims

Once a court has decided what a tokenized claim legally is, distribution follows the ordinary Chapter 11 priority waterfall under Sections 507 and 726 of the Bankruptcy Code. Secured creditors get paid first out of their specific collateral, up to its value. Administrative claims — the professional fees, trustee costs, and post-petition financing that keep a Chapter 11 case running — come next. Certain priority unsecured claims, like limited wage claims and some tax obligations, come after that. Everyone else, including ordinary trade creditors and token holders who were found to be unsecured, splits whatever is left on a pro rata basis. Equity holders are last and are wiped out in the large majority of cases that reach this stage.

The chart below shows how differently the same pool of estate assets gets carved up depending on whether tokenized claims are found to be secured (via a perfected Article 12 security interest) or unsecured (via commingled, title-transferring custody, the Celsius pattern).

A Worked Example: $30 Million in Tokenized Notes, One Chapter 11 Filing

Numbers make the abstract legal distinction concrete. Consider a hypothetical issuer, “Meridian Receivables SPV,” that raised $30 million by selling 600 tokenized private credit notes at $50,000 face value each, backed by a pool of small-business receivables. Meridian files Chapter 11. The receivables pool, after collection costs and a handful of defaults, liquidates for $24 million in cash available to the estate.

Ahead of any distribution to token holders, the estate has to pay:

• A secured revolving credit facility with a first-priority lien on the receivables pool: $6,000,000, paid in full from collateral proceeds.
• Chapter 11 administrative claims (professional fees, DIP financing costs): $1,200,000.
• Priority unsecured claims (capped wage claims and certain taxes): $300,000.

That leaves a residual estate of $24,000,000 − $6,000,000 − $1,200,000 − $300,000 = $16,500,000 to satisfy everyone else.

Scenario A: Tokens Ruled Unsecured (the Celsius Pattern)

Meridian’s note purchase agreement let it “reinvest and commingle” receivables proceeds pending distribution, and it operated a single omnibus wallet rather than segregating tokens against identifiable receivables. A court applying the Celsius reasoning finds the tokens were never held in a true bailment; the receivables were property of the estate, and token holders are general unsecured creditors.

Token holders’ $30,000,000 in face-value claims now compete against $2,500,000 in ordinary trade creditor claims for the $16,500,000 residual: a combined pool of $32,500,000 in unsecured claims chasing $16,500,000.

Recovery rate = $16,500,000 ÷ $32,500,000 = 50.8%. Each $50,000 token recovers roughly $25,400 — a loss of nearly half the invested principal, plus however many months of Chapter 11 proceedings it takes to actually see a distribution check.

Scenario B: Tokens Perfected Under UCC Article 12

Now assume Meridian instead issued its tokens as controllable electronic records, gave each purchaser genuine control consistent with Article 12, and its note purchase agreement created (and its counsel properly perfected) a security interest in the receivables pool in the token holders’ favor, ranking immediately behind the revolving lender.

Token holders now stand as secured creditors with respect to the $16,500,000 of residual collateral value. Their claim ($30,000,000) exceeds that collateral value, so they recover the full $16,500,000 available — but as a senior secured claim, ahead of the $2,500,000 in ordinary trade creditors, who are left with nothing from this pool and must look elsewhere in the case for any recovery at all.

Recovery rate for token holders = $16,500,000 ÷ $30,000,000 = 55.0%, or roughly $27,500 per token. That is a meaningful improvement over Scenario A, and it comes entirely from a change in legal characterization — not from the receivables performing any differently. The trade creditors’ recovery, meanwhile, falls from 50.8% to effectively zero from this collateral pool, which is exactly the kind of shift that makes perfection status a genuinely adversarial issue inside these cases rather than a technicality.

Custodial Tokens vs. Perfected Security Interests vs. Bankruptcy-Remote SPV Notes

Common Mistakes That Turn Token Holders Into Unsecured Creditors

• Trusting marketing language over the subscription agreement. “Your asset, your token” on a landing page means nothing if the note purchase agreement grants the issuer discretion to commingle or reinvest proceeds.
• Assuming self-custody rhetoric applies to pooled RWA products. “Not your keys, not your coins” was coined for direct crypto custody. Most tokenized private credit and real estate platforms never gave the buyer keys or Article 12 control to begin with, so the slogan does not transfer over.
• Ignoring which state’s law governs. A choice-of-law clause pointing to a state that has not adopted the 2022 UCC amendments can leave a would-be secured token holder stuck with weaker general-intangible perfection rules.
• Confusing an SPV wrapper with actual separateness. The existence of an SPV in the offering documents proves nothing on its own; shared servicing agreements, commingled operating accounts, or a parent guarantee that never got disclosed can all support substantive consolidation.
• Skipping the servicer and custodian due diligence. A token can be perfectly structured on paper and still fail in practice if the entity actually holding the receivables or the keys is thinly capitalized, unaudited, or affiliated with the issuer in ways that create conflicts during a wind-down.
• Treating a secondary-market purchase as automatically protected. Article 12’s qualifying-purchaser shield only helps a buyer who obtains real control and lacks notice of competing claims — not a buyer who simply receives a ledger entry from a platform that itself never had control to pass along.

A Pre-Investment Checklist for Tokenized Claim Holders

1. Read the note purchase agreement or subscription documents, not the marketing site, and look specifically for commingling, reinvestment discretion, and title-transfer language.
2. Confirm whether the issuer uses a genuine SPV with independent directors, separate books, and an arm’s-length servicing agreement, or whether it merely references an SPV without operational separateness.
3. Ask directly whether tokens are issued as UCC Article 12 controllable electronic records with a perfected security interest, or as simple ledger entries with no perfection at all.
4. Identify the governing state law in the choice-of-law clause and confirm whether that state has adopted the 2022 UCC amendments.
5. Check whether receivables or collateral are segregated on-chain per claim, or pooled into a single omnibus structure the issuer can reallocate.
6. Look for an independent audit or attestation of the underlying collateral pool, ideally from a firm with no other financial relationship to the issuer.
7. Model your own recovery under both an unsecured and a secured scenario, the way the worked example above does, rather than assuming either outcome by default.
8. Verify whether existing senior secured lenders (warehouse facilities, credit lines) already have priority claims on the same collateral pool your tokens reference.

Key Takeaways

• Bankruptcy courts decide what a tokenized claim is by reading contracts and custody records, not by reading the blockchain.
• Title-transfer and commingling language, the same pattern that turned Celsius and Voyager depositors into unsecured creditors, applies just as directly to tokenized real-world-asset platforms.
• The 2022 UCC Article 12 amendments give issuers and lenders a real path to perfect a security interest in a tokenized claim through “control,” but most retail-facing platforms have not implemented it, leaving buyers with unsecured exposure by default.
• In a representative $24 million estate scenario, moving a $30 million tokenized claim from unsecured to properly perfected secured status raised recovery from roughly 51% to 55% of face value, while cutting unrelated trade creditors’ recovery from the same pool to zero.
• A bankruptcy-remote SPV only protects token holders if the SPV maintains genuine operational separateness; shared cash management or servicing with a parent can expose the SPV’s assets to substantive consolidation.
• Due diligence on a tokenized claim should focus on the subscription agreement, the governing state’s UCC adoption status, and whether real “control” was transferred, not on the smart contract’s code.

Frequently Asked Questions

Does holding a token give me any special legal protection if the issuer goes bankrupt?

No. A token is only as protective as the legal rights it was designed to carry. If the underlying agreement transferred title to the issuer or allowed commingling, the token holder is typically an unsecured creditor regardless of how the asset is technically represented on-chain.

What is a “controllable electronic record” under UCC Article 12?

It is a category created by the 2022 Uniform Commercial Code amendments for digital records like tokens, where a party can have legally recognized “control” — roughly the digital equivalent of possession — that can be used to perfect a security interest more strongly than older, weaker “general intangible” rules allowed.

How did the Celsius and Voyager bankruptcies affect tokenized real-world assets?

Those cases established the core reasoning courts now apply broadly: terms of service language describing title transfer or commingling turns depositors into unsecured creditors, while genuine custody with segregation can keep assets out of the bankruptcy estate. That same analysis extends to tokenized private credit, receivables, and real estate platforms, not just crypto exchanges.

Does a bankruptcy-remote SPV structure guarantee my tokenized claim is safe?

No. An SPV only protects token holders if it maintains real operational and financial separateness from its parent. Shared cash accounts, shared staff, or an undisclosed guarantee can lead a bankruptcy court to substantively consolidate the SPV’s assets into the parent’s estate anyway.

How can I tell if my tokenized claim is secured or unsecured before an issuer ever runs into trouble?

Look at the subscription or note purchase agreement for explicit security-interest language, ask whether a UCC financing statement or an Article 12 control arrangement was actually put in place, and confirm which state’s law governs the deal and whether that state has adopted the 2022 UCC amendments.

References

1. United States Bankruptcy Code, Title 11, Sections 541, 507, 726, and 1129 (property of the estate, priority of claims, distribution, and plan confirmation).
2. In re Celsius Network LLC, Case No. 22-10964 (Bankr. S.D.N.Y.), ruling on “Earn” program asset ownership, August 2023.
3. In re Voyager Digital Holdings, Inc., Case No. 22-10943 (Bankr. S.D.N.Y.), terms of use and customer claim treatment.
4. In re FTX Trading Ltd., Case No. 22-11068 (Bankr. D. Del.), customer property and commingling findings.
5. Uniform Law Commission, Uniform Commercial Code Amendments (2022), Article 12 — Controllable Electronic Records.
6. American Bankruptcy Institute, commentary on digital asset custody and Section 541 property-of-the-estate analysis.