Crypto credit splits into four stacks, and each one asks a different question of the next lender

At 13:01 UTC on 17 July 2026, a research note from Cointelegraph circulated across crypto trading desks with a deceptively simple claim: not all crypto credit is built on the same foundations. CeFi lending, DeFi pools, tokenised Treasuries and real-world-asset private credit each rely on different forms of custody, collateral and enforceability, and a shock that flattens one stack leaves the other three standing.

The note lands at a moment when the industry's lending surface has never been larger or more fragmented. The next lender to fail will not fail for the same reason as the last one. That is the story worth sitting with, because regulators, auditors and depositors are still working off a single template.

Four stacks, four collateral theories

Centralised-finance lending, the model that blew up at Celsius, BlockFi, Genesis and the perimeter of FTX, treats the lender's balance sheet as the credit instrument. Collateral is rehypothecated, custodied by the same firm that books the loan, and enforced through contracts in a common-law jurisdiction that may or may not be honoured when the firm is in Chapter 11. The Cointelegraph note restates the obvious but durable point: a CeFi loan is, at root, an unsecured claim on a corporate counterparty whose own solvency is the variable being bet on.

DeFi lending inverts the structure. The protocol holds the collateral in a non-custodial smart contract and lets the code enforce liquidation at a price threshold, without any corporate counterparty between lender and borrower. The collateral theory is mechanical and visible: the loan is over-collateralised by a crypto asset whose mark is set by an oracle, and the position is closed by an automated auction. There is no balance sheet to raid, and no CEO to sue. The Cointelegraph note frames this as a distinct enforceability regime, not as a less risky cousin of CeFi.

Tokenised Treasuries and money-market funds sit between the two. The collateral is a traditional short-duration government instrument, the custody is a regulated bank or qualified custodian, and the on-chain wrapper does not change the legal claim on the underlying bond. The note flags this stack as the one where the loudest marketing meets the oldest plumbing.

RWA private credit is the newest and the strangest. A loan to a non-crypto borrower, originated by an institutional lender, is bundled, assigned to a special-purpose vehicle and represented on-chain as a token that pays a coupon. Here the enforcement regime is the original loan agreement, in whatever jurisdiction the borrower lives, plus the structural protections of an SPV. The on-chain token is a settlement layer, not a credit engine.

What the marketing blurs

Headline figures across the industry routinely roll all four stacks into a single "on-chain credit" or "DeFi TVL" line. The Cointelegraph note is one of the sharper recent reminders that the number has at least four moving parts, each with a different loss profile in a stress event. A DeFi liquidation cascade on a memecoin does not threaten the holder of a tokenised T-bill. A borrower default on a private-credit RWA does not move a Compound utilisation rate. A CeFi balance-sheet failure does not drain a MakerDAO vault.

The trouble is that the secondary market treats them as fungible risk on a bad day. A large liquidation on one venue can move oracle prices for collateral that another venue has already accepted, and a CeFi counterparty that becomes insolvent can still be a node in a settlement path that touches an RWA. The note stops short of quantifying the contagion channels, and that is its most useful contribution: it names the seams without overstating them.

A regulatory template that is catching up

Regulators in the United States, the United Kingdom, Singapore and the European Union have, to varying degrees, accepted the proposition that fully collateralised on-chain lending is not the same activity as CeFi balance-sheet lending. The Cointelegraph note does not name any rule, but the framing matches the direction of travel: tokenised Treasuries and money-market funds are being slotted into existing securities and money-market frameworks, private-credit RWAs are being handled case by case by securities regulators, and CeFi lending is the stack still waiting on a clean supervisory home in most jurisdictions.

That asymmetry is the one to watch. The stacks with the cleanest legal wrappers are growing the fastest, and the stack with the muddiest supervisory status is the one that still looks most like the activity that collapsed in 2022. The incentive structure for new entrants is to choose the wrapper that minimises regulatory friction, not necessarily the one that minimises credit risk. A tokenised T-bill with bank custody is not safer than a CeFi loan simply because the marketing uses the word "on-chain".

What the next failure will look like

The Cointelegraph note's implicit forecast is that the next lender to fail will be diagnosed after the fact as a category error. A private-credit RWA platform that defaults will be reported on as a DeFi story; a tokenised money-market fund that breaks the buck will be reported on as a CeFi story; a CeFi lender that rehypothecates collateral will be reported on as an exchange story. The four-stack frame is a corrective: it asks which custody, which collateral and which enforcement regime actually applied.

For lenders, the practical discipline is to read the offering documents against the marketing. For depositors, it is to know which of the four stacks their position sits in, and to price the counterparty accordingly. For regulators, it is to stop using a single activity label for four different activities. The note does not solve any of those problems, but it names them in the same paragraph, which is more than most industry research manages.

Desk note: this piece follows Monexus's standing approach to crypto-credit coverage, separating the lending stack from the marketing and reading both against the underlying enforceability regime. Sources are limited to the Cointelegraph research note circulated on 17 July 2026 and the Telegram thread that carried it.

Wire provenance

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

• https://t.me/cointelegraph