Stablecoins explained: pegs, reserves and what can go wrong

Here's a slightly funny thing about crypto: one of its most useful inventions is a coin that's designed to be boring. While everyone argues about which token will 10x, stablecoins quietly do the opposite. They try very hard to stay worth exactly one dollar, forever.

That boringness is the point. Most useful crypto activity (trading, lending, sending money abroad, paying for something) is way easier when at least one side of the deal isn't swinging 15 percent in an afternoon. A stablecoin gives you something dollar-like that still moves on a blockchain, with the speed, 24/7 availability and programmability that comes with it.

But "stable" describes a goal, not a guarantee. How each stablecoin tries to reach that goal differs enormously, and those differences matter much more than the shared label suggests.

What stablecoins are actually used for

Before the mechanics, it helps to know why people care. Stablecoins have become some of the most heavily used tokens in all of crypto. Common uses include:

• Trading. Traders park money in stablecoins between trades instead of cashing out to a bank, and many trading pairs are priced against them.
• DeFi. Lending protocols, liquidity pools and yield strategies are built heavily around stablecoins. See our guide to DeFi yield.
• Cross-border payments. Sending stablecoins abroad can be faster and cheaper than a traditional international transfer, especially for small amounts.
• Savings in unstable currencies. In countries with high inflation or currency controls, people use dollar stablecoins to hold value in something steadier than the local currency.
• Business payments. Some companies use them to settle invoices or pay contractors across borders.

How a peg is actually held: arbitrage

A peg isn't magic. For most stablecoins, it's held in place by people chasing small, low-risk profits. Here's the basic idea with a fiat-backed coin that lets approved users redeem tokens for dollars:

The whole system depends on one thing: people believing the redemption promise is real. If traders start doubting that a token can actually be swapped for a dollar, nobody wants to do the buy-at-$0.99 side of the trade, and the price can slide much further.

Three designs, three very different bets

1. Fiat-collateralized: backed by cash and bonds

The issuer, a company, holds reserves such as cash, bank deposits and short-term government debt, roughly equal in value to every token in circulation. When you redeem a token (usually through the issuer or an approved partner), the issuer sends dollars and destroys the token. This is the simplest model, and it's by far the most widely used. The largest stablecoins, such as Tether's USDT and Circle's USDC, work this way.

The bet here isn't about code. It's about trusting a company. Does the issuer actually hold what it claims? Are the reserves as safe and liquid as advertised? Can it pay out quickly if everyone wants to redeem at once? Because proof usually comes from periodic reports by accounting firms (called attestations) or audits, rather than continuous public verification, the risk sits in the accuracy and honesty of that reporting and in the safety of wherever the reserves are held.

2. Crypto-collateralized: backed by other crypto

Instead of dollars in a bank, a protocol holds other crypto assets in smart contracts. Because crypto is volatile, these systems are usually over-collateralized: to borrow $100 of stablecoins, you might need to lock up $150 or more of another asset. If the value of that collateral falls too close to the debt, the protocol automatically sells it (liquidates it) to protect the peg. DAI, created by the MakerDAO project (now part of the Sky ecosystem), is the best-known example.

This model swaps trust in a company for trust in code and markets. The smart contracts have to work correctly, price feeds (called oracles) have to be accurate, and the collateral market has to stay liquid enough to sell positions in an orderly way, even during a fast crash when every crypto asset tends to fall at the same time.

3. Algorithmic: held up by incentives

Algorithmic stablecoins have little or no direct collateral. Instead they use incentives, typically creating and destroying a second, freely floating token, to nudge the price back to $1. When the stablecoin trades below $1, the system lets you swap it for $1 worth of the second token; when it's above, the reverse.

This model has the least backing and the worst track record. It works only while people believe the second token is worth something, and that belief is exactly what disappears first in a crisis.

Real times pegs broke

This isn't theoretical. Each design has had a public, painful test.

TerraUSD (UST), May 2022: the algorithmic collapse

UST was a large algorithmic stablecoin paired with a token called LUNA. In May 2022, a wave of selling pushed UST below $1. Holders rushed to swap UST for LUNA, the system printed huge amounts of new LUNA to honor those swaps, LUNA's price crashed, and confidence vanished. Within days UST was worth a fraction of a dollar and LUNA was close to zero, wiping out tens of billions of dollars of value and dragging down lenders and funds exposed to it. It remains the clearest example of a confidence spiral.

USDC, March 2023: when the bank wobbled

USDC is fiat-backed, and Circle disclosed that about $3.3 billion of its reserves sat at Silicon Valley Bank when that bank failed in March 2023. Over a weekend, USDC dropped to around 87 or 88 cents on some markets as people worried those reserves were stuck. It recovered once US authorities said all SVB deposits would be protected. The lesson: even "safe" reserves are only as safe as the institutions holding them.

DAI, March 2020: the crash test

On 12 March 2020, crypto prices collapsed as markets reacted to the pandemic. Ethereum's network was so congested that MakerDAO's liquidation auctions didn't work as designed; some collateral was sold for almost nothing, and the system was left with bad debt that had to be covered. DAI itself traded above $1 for a while because people scrambled for it to repay loans. The system survived, but it showed how crypto-collateralized designs get tested hardest exactly when markets are at their worst.

How pegs break: the four repeat offenders

Reserve doubt

Questions about whether fiat-backed reserves really exist, are safe, or can be reached quickly. Doubt triggers redemptions, and if the issuer can't pay fast enough, the price slips.

Collateral collapse

For crypto-backed coins, a sharp crash that outruns the liquidation system, especially when everything falls at once and buyers disappear.

Confidence spirals

For algorithmic coins, a wobble becomes self-fulfilling. As the peg slips, the reason to hold the token weakens, which weakens the peg further.

Contagion

Stablecoins are wired into exchanges, lending protocols and liquidity pools. A failure somewhere connected can force selling that pressures pegs across the system, even for coins that weren't directly involved.

Why regulators care so much

As stablecoins started to look more like money people actually depend on, governments paid closer attention. The focus has mostly been on reserves: what issuers can hold, how often they must prove it, and who supervises them. In the European Union, the MiCA regulation has applied specific rules to stablecoin issuers since 2024, and the United States passed a federal stablecoin law, the GENIUS Act, in 2025. Rules vary by country and are still developing, so check the current situation where you live.

The logic is simple. The more a stablecoin behaves like money, the more a broken peg hurts people who never chose to take a crypto bet.

What to check before you trust a stablecoin

1. Identify the design

   Fiat-backed, crypto-backed or algorithmic? That tells you which kind of risk you're taking on.

2. Read the reserve reports

   For fiat-backed coins: what's in the reserves, how often are they checked, by whom, and where are they held? Lots of cash and short-term government debt is generally safer than riskier assets.

3. Check the collateral rules

   For crypto-backed coins: how over-collateralized is the system, which assets are accepted, and how has it behaved in past crashes?

4. Understand redemption

   Can ordinary users redeem directly, or only big partners? If you can't redeem, you're relying on being able to sell to someone else.

5. Don't put everything in one coin

   If you hold meaningful amounts, spreading across more than one well-established stablecoin limits the damage if one has a problem.

Stablecoins vs a bank account

A dollar stablecoin can feel like a digital bank balance, but it isn't one. In most countries, bank deposits come with government deposit insurance up to a limit. Stablecoins generally don't. If an issuer fails, how and when you get paid depends on its legal structure and local rules. Some platforms offer interest on stablecoin deposits; that interest comes from lending your coins out or other activities, which adds extra risk on top of the stablecoin itself.

This article is general education about how things work. It is not financial, investment or tax advice, and it doesn't consider your circumstances. Do your own research and talk to a licensed professional before making decisions with money.