Not all stablecoins are stable in the same way

“Stablecoin” describes an objective, not a mechanism. Three broad designs exist and they fail differently.

Fiat-backed

Reserves held in cash and short-dated government paper. The risk is straightforward: whether the reserves exist, what they are actually invested in, and whether redemption works under stress.

This is a counterparty and disclosure question, not a crypto question. Read the attestations — specifically the maturity profile of the reserves, because a duration mismatch during a redemption run is the classic failure.

Crypto-collateralised

Over-collateralised with volatile assets, typically at 150% or more. The mechanism is liquidation: if collateral falls, positions get closed automatically.

The failure mode is a fast drawdown where liquidations cannot clear, which has happened. These designs are transparent — you can verify the collateral on-chain — and structurally fragile in exactly the conditions where you would want them to hold.

Algorithmic

Maintains the peg through supply adjustment and arbitrage incentives, with little or no external collateral.

This category has a poor record. The failure mode is reflexive: confidence falls, the mechanism requires confidence to work, and the unwinding is fast. Several have gone to zero within days.

The practical point

A peg holding is not evidence the design is sound. Every failed stablecoin held its peg right up until it did not, and the ones that failed fastest had held longest.

Not financial advice.

Reported at CoinDesk and Cointelegraph; analysis ours.