The LUNA / UST Collapse: A Death Spiral by Design

On 25 April 2022, LUNA traded at $97.

On 13 May 2022, it traded at $0.0001.

Eighteen days. A decline of 99.9999%. Note that the chart uses a logarithmic scale — on a linear axis the entire curve would flatten onto the x-axis in the final days and show nothing at all.

This case belongs in any risk-education material because it demonstrates a category of risk fundamentally different from ordinary losses: not a decline, but a zero.

How the system was supposed to work

The Terra ecosystem had two core assets:

• UST, an “algorithmic stablecoin” designed to hold a $1 peg.
• LUNA, the ecosystem’s native token, freely floating.

Unlike USDT or USDC, UST was not backed by dollar reserves. It held its peg through a mint-and-burn arbitrage mechanism:

• When UST traded above $1 → anyone could burn $1 worth of LUNA to mint 1 UST and sell it for a profit. UST supply rises, price falls back toward $1.
• When UST traded below $1 → anyone could burn 1 UST to receive $1 worth of LUNA. UST supply falls, price rises back toward $1.

In calm conditions this worked. Arbitrageurs did capture the spread, and the peg did hold.

But note the assumption embedded in the second case: it requires LUNA to carry enough market value to absorb every possible UST redemption.

Where the 20% came from

Demand for UST came largely from a lending protocol called Anchor Protocol, which offered roughly 20% annualised yield on UST deposits.

In traditional finance at that time, near-risk-free dollar yields were under 1%. Twenty percent is a number that demands an explanation.

And the explanation was not primarily interest paid by borrowers — borrowing demand fell far short of covering deposit yields. The shortfall was subsidised from project reserves.

Which means: that yield was a marketing expense, not the output of economic activity.

How long a subsidy lasts depends on how long the reserve lasts. And the moment the yield falls, the primary source of UST demand disappears with it.

9 May: the loop runs backwards

When large UST sales pushed the price below $1, the arbitrage mechanism engaged as designed: people burned UST and minted LUNA.

The problem was scale.

UST’s market value was roughly $18 billion. Absorbing redemptions at that scale required minting an enormous quantity of new LUNA. So:

UST depegs → mass redemptions → vast LUNA minting → LUNA diluted, price collapses → LUNA’s market value can no longer back UST → deeper depeg → more redemptions → …

This is the death spiral. It was not an exploit or an external attack. It was the mechanism functioning as specified, under stress.

How large was the dilution? LUNA’s circulating supply went from roughly 350 million tokens before the collapse to over 6 trillion within days — an expansion of roughly twenty-thousand-fold.

The price did not “fall” to $0.0001. It was printed there.

What transfers

One: every yield must have an explainable source

Faced with a yield well above the market benchmark, the only question that matters is: who is paying this, and why are they willing to?

If the answer is “borrower interest” or “real fee revenue,” it may be sustainable. If the answer is “the project’s subsidy” or “incoming deposits,” it is a countdown — and you are betting on not being last.

Two: circular dependency is not collateral

UST’s value depended on LUNA’s market value. LUNA’s market value depended on confidence in UST.

That is a closed loop, not collateral. Real collateral must be exogenous — its value cannot depend on the thing it backs. When confidence goes, both ends of a loop fail simultaneously rather than supporting each other.

Three: some risks are not price risk but existential risk

This point matters especially to anyone doing technical analysis.

Chart reading, stop-losses, position sizing — the entire toolkit assumes something: the instrument continues to exist, prices move continuously, and you have an opportunity to exit along the way.

LUNA’s collapse destroyed that premise. It fell fast enough that stops were largely meaningless — a stop at $50 might fill at $10, or might not fill at all because exchanges were congested or trading was halted.

For an instrument that can go to zero, technical analysis does not address the main risk. Judging whether it can go to zero is analysis of mechanism, counterparty, regulation and fraud — an entirely different discipline.

A closing note

Estimated losses exceeded $40 billion. The event triggered a chain reaction across the crypto sector; several leveraged institutions failed over the following months.

This case is written up not to judge any asset class, but because it is the clearest recent teaching example of one idea: when a system’s stability depends on confidence in it, and that confidence depends on its stability, collapse is not a question of “if” but “when.”

This is a historical review written for risk education. It is not investment advice, does not recommend any asset, and makes no inference about any current market conditions.