Imagine wanting to create a digital currency that behaves with the predictability of the dollar but without relying on the vast machinery of banks, vaults, treasuries, or governments. You want to conjure stability from pure code, to convince a global marketplace that tokens with no intrinsic backing can nonetheless hold their peg, and that people will treat them as equivalent to the most powerful fiat currencies in the world. That is the promise and the paradox of algorithmic stablecoins, one of decentralized finance’s most ambitious experiments and also one of its greatest cautionary tales. At their most compelling, they offered a glimpse of a self-sustaining digital money that could circulate freely without dependence on legacy systems. At their most fragile, they revealed how quickly elegant theory can collapse under the weight of market panic, human psychology, and the unforgiving mathematics of incentives.

To see why these designs captured imaginations, it helps to understand what they were meant to improve upon. Overcollateralized stablecoins like DAI demonstrated that a decentralized currency could maintain a peg without a central bank, but they did it by leaning heavily on collateral buffers. To mint $100 worth of DAI, a user had to lock up $150 or more of other cryptocurrencies. If the value of the collateral fell, the system would liquidate positions to preserve solvency. The architecture was straightforward and conservative, more like a cautious lender than an ambitious monetary experiment. And crucially, it worked. Through wild bull runs, brutal drawdowns, and multiple market shocks, DAI has maintained its peg with remarkable resilience. Yet this reliability comes at a steep cost in capital efficiency. Locking up more value than one creates limits scale, especially in a financial culture where capital is expected to move quickly.

The search for efficiency, for a design that could deliver the same stability without so much collateral tied up, led naturally to more daring experiments. Algorithmic stablecoins promised stability as an emergent property of supply, demand, and incentive design. They did not ask participants to over-collateralize; instead, they offered mechanisms that would expand or contract supply automatically in response to deviations from the peg. Where central banks rely on monetary policy tools and reserves to defend currencies, these systems attempted to achieve the same effects through algorithms alone. It was as if the bootstrap metaphor of DeFi—finance pulling itself up by its own code—was taken literally.

The mechanics varied, but they shared a common principle: when the stablecoin traded above one dollar, the system would expand supply; when it traded below, the system would contract it. TerraUSD, the most notorious example, achieved this with a dual-token model. Its stablecoin UST was kept in line with the help of its sister token LUNA. Whenever UST drifted above its peg, arbitrageurs could burn LUNA to mint more UST and capture a profit, thereby increasing supply and pushing the price down. Whenever UST traded below its peg, holders could burn UST to mint LUNA, decreasing UST supply and theoretically restoring price equilibrium. On paper, the design resembled a perpetual balancing act, with the market incentivized to do the heavy lifting of stabilization. In practice, the system required ever-growing confidence and demand to keep functioning.

Other algorithmic systems experimented with different variations. Rebasing coins like Ampleforth changed the number of tokens in every wallet directly, increasing balances when prices rose and shrinking them when prices fell. The value of one’s holdings in dollar terms remained roughly stable, but the experience of seeing balances fluctuate was psychologically jarring. Multi-token systems like the designs inspired by Basis proposed using bonds and shares to create a complex interplay of incentives, where some participants would absorb supply in bad times in exchange for future rewards in good times. Fractional models like Frax sought compromise, mixing collateral with algorithmic adjustment, gradually reducing reliance on reserves as the system matured.