Crypto Lending: What Replaced the Firms That Failed

Several large crypto lending firms failed within months of each other in 2022. What replaced them differs structurally, and the differences are worth stating precisely.

What failed

The model was: take customer deposits, promise a yield, lend the assets to institutional borrowers, keep the spread.

The failures had common features.

Unsecured or undersecured lending. Loans made against reputation rather than collateral, in some cases to a small number of borrowers.

Concentration. A large share of the book with few counterparties, so one failure was existential.

Maturity mismatch. Deposits redeemable on demand, loans with longer terms. A withdrawal run cannot be met regardless of solvency.

Opacity. Depositors had no visibility into where their assets went.

Yield competition. Firms competed for deposits on advertised rate, which forced increasingly risky lending to fund it.

What operates now

Over-collateralised on-chain lending. Borrowers post collateral worth substantially more than the loan, with automatic liquidation if it falls below a threshold. Transparent, verifiable on-chain, and capital inefficient by design.

The risks are different rather than absent: oracle manipulation, contract bugs, and liquidation mechanisms failing during extreme volatility.

Institutional secured lending. Bilateral, collateralised, with legal agreements and independent custody of the collateral. Closer to conventional securities lending.

Exchange-operated margin. Venues lending to their own customers against positions held on the platform, with automated liquidation.

What is genuinely better

Collateral is standard. The unsecured lending that caused the failures is largely gone from the retail-facing part of the market.

Transparency where on-chain. Anyone can verify collateralisation ratios in a protocol continuously.

Regulatory attention. Products offering yield to retail depositors now attract supervision in most major jurisdictions, which has removed some of the worst offerings.

What has not changed

Yield still comes from a borrower. Any product paying a return is lending your assets to someone who may not repay. The advertised rate is a claim about risk-adjusted return, and the risk part is rarely quantified.

Platform risk remains. A centralised yield product is an unsecured claim on a company, whatever the marketing says.

Complexity hides leverage. Layered products can rebuild the maturity mismatch that caused the failures, in a form that is harder to see.

The question that still works

Where does the yield come from, and what happens to it if the borrower fails?

An answer naming a mechanism you can verify is an answer. An answer using the words “strategy”, “arbitrage” or “market neutral” without specifics is not, and that distinction has been sufficient to avoid every major failure in this category so far. For anything you intend to trade rather than hold, a regulated European platform is the part of the setup worth getting right first.