Crypto credit is splitting into four markets, and only one of them looks like a bank

On 17 July 2026, Cointelegraph Research published a working paper that does something the crypto credit market has avoided doing for half a decade: it draws bright lines between four asset classes that practitioners routinely lump together. CeFi lending, DeFi pools, tokenized US Treasuries and private credit backed by real-world assets, the researchers argue, are not variants of one trade. They are four trades with different custody arrangements, different collateral enforceability and different default waterfalls, and conflating them has already cost investors real money.

The taxonomy matters because the next stress event in digital assets will not be a single failure. It will be a mismatch, between how a buyer thought a product was built and how it was actually built, surfacing in the same week across four otherwise unrelated balance sheets.

What the four buckets actually are

Cointelegraph Research defines the divide by asking three questions of each instrument: who holds the collateral, who can seize it, and who gets paid first when the borrower defaults. Most marketing material skips the second and third.

CeFi lending is the oldest and most familiar bucket. A centralised lender takes digital-asset deposits, lends against them, and keeps the spread. Custody sits with the lender, collateral is enforced through the lender's terms of service and bankruptcy estate, and the depositor is an unsecured creditor of a corporate counterparty. This is the structure that collapsed at firms like Celsius, BlockFi and Genesis in 2022.

DeFi pools invert the arrangement. Smart contracts on a public chain hold the collateral, liquidation runs on code, and the depositor is a pro-rata claimant against an on-chain liquidation pool. There is no corporate counterparty to sue, because there is no corporate counterparty at all. The risk model looks completely different: not the failure of a fiduciary, but the failure of an oracle, a price feed or a liquidation engine.

Tokenized US Treasuries sit somewhere between the two. A regulated entity issues a blockchain token that represents a claim on Treasury bills held by a custodian. The token settles like a stablecoin, the underlying settles like a money-market fund, and redemption depends on a named issuer honouring a redemption queue. The architecture borrows from both worlds, and inherits failure modes from each: a smart-contract bug on the token side, a settlement delay or suspension on the fund side.

Private credit backed by real-world assets is the newest of the four and the one Cointelegraph Research flags as the most heterogeneous. Loans to corporates, invoices, receivables or property are pooled, securitised, and issued as tokenised claims on an off-chain pool. Enforcement runs through the legal system of the jurisdiction where the borrower sits. The chain records the transaction; the law governs the default.

Why the conflation is dangerous

The 2022 cycle taught the market that CeFi lenders are not banks, a point now obvious in hindsight. Investors who bought yield on Celsius or Voyager discovered that "yield" was partly the compensation for absorbing a counterparty they had never audited.

The risk now is that the same muscle memory is migrating to RWA private credit. Buy-side mandates pitched as "tokenised yield" or "on-chain credit" often mean a regulated securitisation vehicle wrapped in a token. The legal recovery path in a default runs through a Delaware or Cayman special purpose vehicle, not through a liquidation bot. A token holder who expects to be made whole in minutes may instead be waiting on a trustee, a servicer and a court order.

DeFi pools face the mirror problem. Protocol governance changes, oracle failures and cross-chain bridge exploits all carry the same first-loss default, namely the person who showed up last. Several post-mortems of 2024-2025 DeFi exploits traced the loss not to the underlying borrower but to a liquidation bot racing the price feed, leaving depositors with the residual claim.

The Cointelegraph Research framing makes the mispricing concrete: the spread between a tokenised Treasury yield and the underlying Treasury bill yield is not pure arbitrage. It prices the issuer's redemption risk, the custodian's operational risk, and the smart-contract's bug risk, all of which move on different clocks.

The structural shift already in motion

The four buckets are not competing on a level playing field. US Treasury-backed structures now move on a regulatory rail that DeFi and CeFi do not have. Tokenised Treasuries sit inside money-market-fund-style supervision, with daily mark-to-market and documented redemption mechanics. Private credit funds backed by RWAs increasingly run inside the same family of regulated wrappers that collateralised loan obligations (CLOs) used a generation ago, with the chain replacing the trustee's spreadsheet.

The structural effect is that capital is re-sorting itself by legal regime, not by yield. Treasury-backed products attract balance sheets that need a regulated settlement layer, because that is where their reporting and their counterparty risk model already sit. DeFi pools attract balance sheets that explicitly want code-native enforcement. CeFi lending lives in the shrinking middle: useful for users who want a familiar interface, structurally exposed in a stress event.

That sorting is also visible in the vendor landscape. The same set of issuers that pioneered on-chain credit in 2021-2022 now run separate business lines for tokenised Treasuries and for private credit, with separate compliance teams, separate auditors and separate bankruptcy-remote structures. The unification story that sold the last cycle, "all of these are crypto credit", is no longer one any of them tells.

What the next default actually looks like

If a private-credit fund structured through a tokenised vehicle defaults, the recovery path looks like a securitisation: holders of the most subordinated tranche absorb losses first, the servicer advances collections, and the senior tranches are paid out of the cash waterfall. The token is administrative paperwork; the chain is the spreadsheet.

If a DeFi pool gets exploited, the recovery path is a governance vote and a fork at best, a haircut distributed pro-rata across depositors at worst. There is no servicer, no trustee and no senior tranche.

If a CeFi lender fails, depositors queue in bankruptcy court behind secured creditors, just as they did in 2022.

These are not the same default, and they are not the same recovery. A portfolio that blends all four, as more mandates now do, has a default probability that is a weighted average across four uncorrelated regimes. That can be a feature, or it can be an opaque concentration problem dressed up as diversification.

The honest read of Cointelegraph Research's taxonomy is that crypto credit is not one market coming of age. It is four markets, each at a different stage of legal and operational maturity, trading under a single banner. The investors who make money in the next cycle will be the ones who treat it that way.

Desk note: Monexus reads Cointelegraph Research's note as a taxonomy clarification rather than a forecast, and treats the four buckets as analytical scaffolding for forthcoming desk coverage of tokenised Treasuries and RWA private credit rather than as investment guidance.

Wire provenance

This editorial synthesis draws on the following public wire/social posts:

• https://t.me/s/cointelegraph
• https://t.me/s/cointelegraph