Research & Insights

Why We Saw a Depeg Across Yield Stablecoins in 2025

2025 has been a brutal year for yield-bearing stablecoins. In April, Synthetix's sUSD dropped to 68 cents. In October, Ethena's USDe briefly hit 65 cents on Binance during a market crash, triggering $8.3 billion in outflows. In November, Stream Finance's xUSD collapsed to 26 cents after an external fund manager lost $93 million in user funds.

These aren't isolated incidents. They're symptoms of the same underlying disease: yield stablecoins built on "trust me" foundations.

The Yield Stablecoin Promise

The pitch is appealing. Hold a stablecoin, earn yield. Your dollar stays a dollar while generating returns that traditional savings accounts can't match. For users tired of watching inflation erode their purchasing power, yield stablecoins seem like the obvious choice.

But yield doesn't come from nowhere. Someone has to generate it. And how that yield is generated—and what happens when the strategy fails—determines whether your stablecoin stays stable.

What Went Wrong

Each of these depegs had different immediate causes, but they share a common architecture: users deposited funds, managers deployed those funds into strategies, and users had limited ability to verify what was actually happening with their money.

Stream Finance's xUSD is the clearest example. The protocol relied on external fund managers to generate yield. When one of those managers lost $93 million, users discovered that xUSD's backing assets totaled only $170 million while borrowing had reached $530 million—a leverage ratio exceeding 4x. On-chain analyst Cbb0fe had warned about this days before the collapse, but most users had no idea the protocol was running with such thin margins.

The depeg didn't stay contained. Because xUSD served as collateral in other protocols, its collapse cascaded outward. Elixir's deUSD had lent $68 million in USDC to Stream—representing 65% of its backing. When xUSD fell, deUSD wobbled too. This is the nature of interconnected systems built on opaque foundations: problems spread faster than information.

Synthetix's sUSD depegged for more transparent reasons—a governance decision gone wrong. SIP-420 reduced the collateralization ratio from 500% to 200% and introduced a shared staking pool. The result was an oversupply of sUSD and reduced individual incentives to maintain the peg. The stablecoin dropped to 68 cents before partially recovering.

Even Ethena's USDe—arguably the biggest and most successful of the yield stablecoins—was not immune from depegs in 2025. During October's market crash—reportedly the largest liquidation cascade in crypto history—USDe dropped to 65 cents on Binance. Ethena argued this wasn't a "real" depeg, just a Binance-specific oracle issue caused by thin order books. They also pointed out that redemptions continued processing and reserves remained overcollateralized.

Nevertheless, Ethena’s market cap dropped from $14.7 billion to $6.4 billion as users fled for safety following the depeg. Whether or not the protocol was technically solvent, confidence evaporated. And in stablecoins, confidence is everything.

The Common Thread

Strip away the specifics and you find the same structural weakness in each case: users couldn't verify what was happening with their deposits until something went wrong.

Stream Finance users didn't know their funds were leveraged 4x until after the collapse. Ethena users couldn't independently confirm that order book issues were exchange-specific rather than protocol-wide—they had to trust Ethena's explanation. Synthetix users voted for a parameter change without fully modeling its consequences.

This is the "trust me" model. Deposit your money. Trust that managers are competent. Trust that leverage is reasonable. Trust that collateral exists. Trust that when things go wrong, someone will fix it.

Trust works until it doesn't. And when it breaks, it breaks suddenly.

Why Automatic Enforcement Matters

The alternative to trust is verification. Not "we'll publish an audit" verification or "our dashboard shows reserves" verification—real-time, on-chain, automatically-enforced verification.

This is what distinguishes protocol architectures that survive crises from those that create them. The question isn't whether managers are trustworthy. Most probably are. The question is: what happens automatically when positions deteriorate?

In the yield stablecoins that depegged this year, the answer was largely "nothing automatic." When Stream Finance's manager lost funds, the protocol suspended withdrawals while engaging lawyers. When Ethena faced stress, users had to trust the team's claims about solvency. The protocols relied on manual intervention during crises—exactly when manual intervention is slowest and least reliable.

Contrast this with systems where enforcement is built into the protocol itself. Where every manager must post collateral that can be automatically seized if positions become unhealthy. Where leverage ratios are enforced by code, not policy. Where anyone can trigger a liquidation when health thresholds are breached, without waiting for governance votes or team announcements.

In such systems, the question "is my deposit safe?" has a verifiable answer. Not "the team says so" but "here's the on-chain collateral backing it, here's the health ratio, here's the liquidation threshold."

The Rehypothecation Problem

The xUSD collapse revealed another danger: collateral reuse across protocols. When xUSD served as backing for other stablecoins, its failure cascaded outward. This is rehypothecation—the same collateral supporting multiple claims.

In traditional finance, rehypothecation is regulated and disclosed. In DeFi, it often happens invisibly. Protocol A accepts token X as collateral. Protocol B accepts protocol A's receipt token as collateral. Protocol C accepts protocol B's token. The leverage multiplies, but no single dashboard shows the full picture.

When the base layer wobbles, every layer above it shakes. And because these connections are often undisclosed, users in protocol C may have no idea they're exposed to protocol A's risks.

Solving this requires transparency at every layer: clear accounting of what backs what, with automatic mechanisms to unwind positions before cascades spread.

What Users Actually Need

After watching three major yield stablecoins depeg in a single year, the requirements become clear:

Collateral you can verify. Not "we have reserves" but "here's the on-chain address, here's the oracle price, here's the math." Every dollar of yield should trace to a specific source with capital at risk.
Automatic liquidations. When positions become unhealthy, they should close automatically—not when governance convenes, not when the team decides, but immediately, triggered by anyone who notices.
Transparent leverage. If a protocol is running 4x leverage, users should see that before depositing, not after collapse. On-chain positions should be queryable by anyone.
Bounded risk. When something does go wrong—and eventually something will—the damage should be contained. Bad debt should be socialized transparently, not hidden until the next crisis.

A Different Architecture

This is why Flux exists. Not as another yield promise, but as infrastructure where yield is verified rather than trusted.

Every manager who borrows from a Flux vault posts collateral—their own capital at risk. Every position is tracked on-chain, visible to anyone. When a manager's health ratio drops below threshold, anyone can liquidate them. No governance delay. No team intervention needed. The protocol enforces solvency automatically.

When vaults face liquidity pressure, Flux's Auto-Deallocating mechanism can force-close healthy positions to free up withdrawals—without penalizing managers, but without leaving depositors trapped either.

None of this prevents all losses. Markets move, strategies fail, bad debt happens. But when it happens in Flux, users can see exactly what occurred, verify the math themselves, and watch automatic mechanisms respond. No waiting for team statements. No trusting that reserves exist.

The yield stablecoin depegs of 2025 weren't caused by bad actors or unprecedented market conditions. They were caused by architectures that substituted trust for verification. The protocols that survive the next crisis will be those that learned this lesson: safety comes from mechanisms, not promises.