A peg is a commitment that one asset will maintain a fixed exchange rate with another asset. USDT is pegged to the US dollar — one USDT should always be worth one dollar. USDC is also pegged to the dollar. WBTC is pegged to BTC — one WBTC should always be redeemable for one BTC. Pegged assets exist in crypto because they let you hold something with predictable value (stablecoins) or move value between chains (wrapped assets) without being exposed to the price volatility of the underlying market. The peg is the promise that makes the wrapper useful, and the promise is kept by whatever mechanism the issuer has in place to maintain it.
Pegs break. This is the most important thing to understand about them. A peg is only as strong as the redemption mechanism that enforces it, and if that mechanism is stressed beyond what it can handle, the peg loses hold and the pegged asset trades below its nominal value. “Depegging” is the word for what happens when a peg breaks, and the crypto industry has had plenty of experience with it.
How Pegs Are Maintained
Three main mechanisms, each with different failure modes.
Reserve-backed pegs work by holding the pegged-to asset in reserve. Tether holds (roughly) $1 of reserves for every USDT issued, and in principle anyone can redeem USDT for real dollars at par by going through Tether’s redemption process. USDC operates similarly, with reserves held at regulated US banks and audited monthly by Grant Thornton. The peg is enforced by arbitrage: if USDC trades below $1 on a secondary market, someone can buy the cheap USDC, redeem it for $1 at Circle, and pocket the difference. This drives the market price back up. The mechanism works as long as the reserves really exist, the redemption process actually functions, and enough market participants are willing to do the arbitrage.
Crypto-collateralised pegs work by over-collateralising with other crypto assets. DAI (issued by MakerDAO, now Sky) is the canonical example: users lock up $150 or more of ETH in a smart contract and mint $100 of DAI, which can be freely redeemed for the collateral at any time. The over-collateralisation provides a buffer against price drops in the backing assets, and the redemption mechanism enforces the peg. The failure mode is that if the collateral drops fast enough, the protocol cannot liquidate fast enough, and undercollateralised positions start producing bad debt. MakerDAO’s Black Thursday event in March 2020 is the canonical case study.
Algorithmic pegs work by creating incentives that should, in theory, keep the peg without any collateral at all. The most famous attempt was Terra’s UST stablecoin, which was “stabilised” by an arbitrage loop with LUNA — one UST could always be swapped for $1 of LUNA, and one dollar of LUNA could always be swapped for one UST. The theory was that arbitrageurs would keep UST at $1 by buying when it was below and selling when it was above. The theory failed catastrophically in May 2022 when UST lost its peg, the arbitrage loop started minting LUNA faster than the market could absorb it, LUNA’s price collapsed, and within a week roughly $60 billion in total value was wiped out. Algorithmic stablecoins without real collateral backing are now widely considered an unsolved problem, and most credible projects have abandoned the approach.
Wrapped Asset Pegs
WBTC, wETH on non-Ethereum chains, and similar “bridged” representations of an asset on a different chain are also pegged — typically to the underlying asset 1:1, maintained by a custodial or protocol-level bridge that holds the real asset and issues the wrapped version. WBTC, for example, is maintained by BitGo, which holds real BTC and mints corresponding WBTC to qualified institutions who bring in the real BTC for wrapping. Users can then trade WBTC on Ethereum DeFi protocols while trusting that it is redeemable for real BTC on the other end.
The peg risk for wrapped assets depends on the reliability of the bridge. A custodial bridge like WBTC has essentially one counterparty (BitGo); its peg holds as long as BitGo remains solvent and honest. Decentralised bridges have security models that depend on validator sets, proof systems, or economic incentives, and these models have failed expensively many times. The Ronin bridge ($600 million), Wormhole ($320 million), Nomad ($190 million), and several others have had catastrophic failures where wrapped assets on the receiving chain became unbacked, their pegs collapsed, and anyone still holding them lost most of their value.
The Practical Lesson
Treat every pegged asset as a conditional promise, not a fact. USDT and USDC have held their pegs consistently, but both have had temporary wobbles (USDC dropped to ~$0.88 briefly in March 2023 during the Silicon Valley Bank panic, when some of Circle’s reserves were stuck at the failing bank). DAI, WBTC, and other well-backed pegged assets have held up reasonably well but come with their own tail risks. Lesser-known pegged assets, bridged tokens from small chains, and anything algorithmic should be assumed to have a higher probability of depegging at the worst possible moment.
The question to ask about any peg is always: “what is the actual mechanism, and what happens if it is stressed?”. A peg with a good answer to that question is worth trusting for day-to-day use. A peg without a good answer is an accident waiting to happen.