Fiat, Crypto, and Algorithmic Stablecoins Compared

Why "stablecoin" is a misleading word

The word "stablecoin" describes a goal, not a mechanism. Every token in this category shares one promise: a market price close to one US dollar. How that promise is engineered varies enormously, and so does the chance that it breaks.

Beginners often assume a stablecoin is the crypto equivalent of a dollar savings account. It is not. A US savings account is a regulated deposit liability of an insured bank. Most stablecoins are either claims on a private company that holds reserves, or open smart-contract systems that try to mint and burn tokens in response to demand. Both designs have produced losses. The rest of this article walks through the three main families, what backs them, and what can go wrong.

The three risk families at a glance

Most stablecoins fall into one of three buckets, with a few interesting hybrids.

• Fiat-backed. A centralized issuer holds traditional assets, usually cash and short-dated US Treasuries, and issues tokens 1:1 against those reserves. USDC and USDT are the largest examples.
• Crypto-backed. A decentralized protocol accepts crypto as collateral, lends out the dollar value, and mints a stablecoin against it. DAI from Sky (formerly MakerDAO) is the classic example.
• Algorithmic or synthetic. The token's peg is defended by software, derivatives, or arbitrage loops rather than reserves. This bucket includes the failed TerraUSD as well as the live Ethena USDe.

A fourth bucket, sometimes called "diversified-collateral" or "RWA-backed," blends real-world assets such as tokenized Treasuries with on-chain crypto. Some newer protocols position themselves this way to compromise between the safety of fiat reserves and the on-chain nature of crypto collateral.

Fiat-backed stablecoins: how they work, where they break

Fiat-backed stablecoins are the simplest to explain. A company accepts US dollars (or dollars and Treasuries) from users, mints tokens, and promises to redeem each token for one dollar on request. The blockchain ledger lets the token move 24/7, but the value still comes from a balance sheet held by a regulated or semi-regulated entity.

USDC is issued by Circle, which holds reserves at US banks and in a BlackRock-managed Treasury fund, and publishes monthly attestations from a Big Four auditor. USDT is issued by Tether, which holds a larger and more opaque reserve mix that historically included commercial paper, bank deposits, and now predominantly Treasuries. Both companies hold assets, both claim to be fully backed, and both have survived stress tests.

The risks nobody talks about

Fiat-backed tokens are only as safe as three things: the issuer, its bank, and the regulator.

• Issuer solvency. If the company that holds the reserves fails or commits fraud, the token can become worth less than a dollar. This is exactly what happened to the smaller issuer Basis Cash in 2021 and to several centralized yield products around the same time.
• Bank and custody risk. During the March 2023 US banking scare, Circle disclosed that roughly 8 percent of USDC reserves were parked at Silicon Valley Bank. USDC briefly depegged to about 87 cents before the FDIC backstop was clarified. The token recovered within days, but the episode proved that a "T-bill backed" stablecoin can still move with bank headlines.
• Freeze and censorship risk. Because issuers control the token contract, they can blacklist addresses. That is sometimes a feature for compliance teams and sometimes a nightmare for users in sanctioned jurisdictions.
• Redemption friction. Most retail holders cannot redeem a stablecoin directly for dollars. They must rely on a crypto exchange or an OTC desk, which adds counterparty and slippage risk.

The honest summary: fiat-backed stablecoins are the closest thing the crypto market has to digital dollars, but they are not dollars. They are unsecured claims on a private company. The reserves lower the risk, they do not eliminate it.

Crypto-backed stablecoins: the CDP mechanism

Crypto-backed stablecoins replace a company's balance sheet with a smart contract. To get a DAI, you deposit crypto collateral such as ETH into a Maker vault (now rebranded to a Sky position) and borrow DAI against it. Because ETH is volatile, the protocol requires you to post more collateral than the loan is worth: typically 150 percent or more, depending on the asset.

This ratio is called the collateralization ratio. If your ETH falls in price and the ratio drops below the liquidation threshold, the protocol automatically sells your collateral at auction to repay the loan and keep the system whole. Liquidations are loud and visible on-chain, which is part of the point: they are the safety valve.

What can go wrong

• Cascade liquidations. In a fast crash, falling collateral triggers liquidations, which adds selling pressure, which crashes the price further, which triggers more liquidations. This is exactly what happened during the March 2020 "Black Thursday" event, when ETH dropped sharply and oracle price updates lagged badly. Millions of dollars of collateral were sold for zero in badly designed auction lots.
• Oracle risk. The system needs a reliable price feed for ETH. If that feed is manipulated or delayed, the protocol can be drained. MakerDAO has weathered several oracle incidents, and competitors have not always been as lucky.
• Governance risk. Parameters such as collateral types, debt ceilings, and liquidation penalties are set by MKR/SKY holders through on-chain voting. Bad parameter choices can weaken the peg during stress.

Crypto-backed designs are genuinely decentralized in the sense that no single company can print DAI or freeze your position without governance. The trade-off is that they are exposed to the underlying volatility of crypto markets, and to the engineering bugs that every complex smart-contract system accumulates over time.

Algorithmic and synthetic stablecoins: how the peg is supposed to work

This is where the history of stablecoins gets painful, and where the design space is most inventive.

Seigniorage models and the Terra collapse

The "algorithmic" family originally referred to seigniorage systems, named after the government privilege of printing money. The best-known example was TerraUSD (UST), paired with a volatile token called LUNA on the Terra blockchain. The mechanism was elegant on paper. When UST traded above one dollar, users could burn one dollar of LUNA to mint one UST, increasing supply and pushing the price down. When UST traded below one dollar, users could burn UST to mint one dollar of LUNA, decreasing supply and pushing the price up.

This arbitrage loop only works as long as someone wants to hold the volatile side of the trade. In May 2022, after a series of large withdrawals, the loop reversed viciously. Minting LUNA to absorb UST supply diluted LUNA's value until both tokens collapsed to near zero. Billions of dollars were wiped out, and the event still defines how serious people talk about algorithmic stablecoins.

The lesson is not that arbitrage is broken. The lesson is that a peg backed only by the promise of an arbitrage loop is backed by the willingness of the marginal holder to absorb losses. When fear takes over, that willingness disappears instantly.

Synthetic dollars: the Ethena USDe approach

A newer design, sometimes called "synthetic," tries to avoid the seigniorage trap. Ethena's USDe does not promise to redeem for a dollar from reserves. Instead, it runs a delta-neutral position: for every USDe minted against deposited ETH or BTC, the protocol shorts an equivalent amount of those tokens via perpetual futures. The P&L of the short offsets the P&L of the collateral, so the dollar value is supposed to stay roughly flat.

The yield that USDe advertises comes from perpetual funding rates. In normal bull markets, perpetual contracts trade at a premium to spot, so longs pay shorts a fee every eight hours. Ethena collects this funding rate and shares it with USDe stakers as yield.

This design is genuinely clever, but it has risks that are easy to miss.

• Funding can flip negative. When traders are aggressively short, funding goes negative and the protocol pays the long side. During prolonged bearish regimes, USDe yield can turn negative.
• Exchange and counterparty risk. The short leg lives on centralized derivatives venues. If access is restricted or a venue fails, the hedge can break.
• Liquidation and basis risk. If ETH spikes hard and the hedge is imperfect, the protocol can realize losses. Liquidity in perpetual markets is deep, but it is not infinite.
• Regulatory risk. A synthetic dollar that promises yield may eventually be treated as a security or a derivatives product in major jurisdictions.

USDe is not the same thing as UST, and it is not the same thing as USDC. Calling it "algorithmic" because it does not hold dollars in a bank misses the point. Calling it "safe" because it has survived a bear market also misses the point. The honest framing is that USDe is a bet that perp funding rates will stay positive on average, backed by the protocol's ability to manage its hedge book.

Diversified and hybrid designs

Some protocols try to get the best of multiple worlds. Frax originally used a fractionally algorithmic model, partly backed by collateral and partly by a mint-and-burn loop. Over time it has tilted toward more traditional collateral. Other projects wrap tokenized US Treasuries or money-market funds together with crypto collateral to blend stability and decentralization.

The honest read on these hybrids is that they are experiments. They can lower the worst-case failure mode by diversifying reserves, but they also multiply the number of things that can go wrong: custodians, tokenization issuers, oracles, governance votes, smart-contract bugs. More moving parts is not automatically safer. Sometimes it is the opposite.

Decentralization is a spectrum, not a flag

Newcomers often ask which stablecoin is "most decentralized." The honest answer is that decentralization is a property of each component: who mints, who custodies, who freezes, who upgrades the code, and who decides parameters. A token can be decentralized on one axis and centralized on another.

• USDC is decentralized in the sense that anyone can hold and transfer it without permission, but the issuer can blacklist addresses and the supply is controlled by Circle.
• DAI is more decentralized on the issuance and freeze axes, but MKR/SKY governance still sets parameters and can vote to add or remove collateral types.
• USDe runs non-custodial smart contracts but depends on centralized exchanges for its hedge leg, so its censorship-resistance is partial.

Treating decentralization as a binary flag, or assuming that "more decentralized" always means "safer," has cost a lot of people money. The Terra system was technically decentralized. So were several DeFi protocols that lost user funds to governance attacks.

How to think about which one to actually use

If you are choosing a stablecoin for payments, savings, or trading, the right question is not "which one is best." The right question is "what failure mode am I willing to live with?"

• If you want the cleanest dollar exposure and you trust the issuer's audits, USDC and USDT are the default. They have the deepest liquidity and the longest track record, but you are trusting Circle or Tether and their banking partners.
• If you want non-custodial dollar exposure and you accept crypto-market risk, DAI-style systems let you hold a dollar position without trusting a company. You are trusting smart-contract code, governance, and liquidation engines instead.
• If you are optimizing for yield and you understand derivatives, USDe and similar synthetic designs can pay well. They can also pay badly when funding flips or hedges break.

Whatever you pick, treat the "stable" in stablecoin as a default assumption that can break. Size your positions accordingly. The people who lost the most in past failures were the ones who treated their stablecoin balance like a checking account.

Stay ahead of stablecoin risk

Stablecoin designs evolve quickly, and so does the news around them. Tracking depegs, reserve attestations, governance votes, and funding-rate shifts manually is a losing game. Zippfeed surfaces stablecoin headlines with sentiment scoring (bullish, neutral, or bearish) and an importance rating, so you can spot peg stress, issuer trouble, and protocol upgrades before they hit your portfolio.