Our NewsletterInvestor RelationsOur NewsletterInvestor RelationsOur NewsletterInvestor RelationsOur NewsletterInvestor RelationsOur NewsletterInvestor RelationsOur NewsletterInvestor Relations
Necessary cookies help make a website usable by enabling basic functions like page navigation and access to secure areas of the website. The website cannot function properly without these cookies.
Preference cookies enable a website to remember information that changes the way the website behaves or looks, like your preferred language or the region that you are in.
Statistic cookies help website owners to understand how visitors interact with websites by collecting and reporting information anonymously.
Marketing cookies are used to track visitors across websites. The intention is to display ads that are relevant and engaging for the individual user and thereby more valuable for publishers and third party advertisers.
Image The xUSD collapse: leverage eating itself

The xUSD collapse: leverage eating itself

Timer4 min read

  • Finance

Stream Finance just gave us a $280 million lesson in why DeFi still hasn't learned from its own mistakes. A $93 million loss by an external fund manager turned into a full-blown systemic crisis, and if you're surprised, you haven't been paying attention.

What actually happened

The facts are straightforward. Stream Finance ran what it called "market neutral strategies" promising 18% yields on $382 million in assets. An external fund manager lost $93 million over the weekend. By Monday morning, xUSD,  their stablecoin that was supposed to hold $1, was trading at $0.26. Withdrawals froze. $160 million in user funds locked. Another DeFi implosion.

But the $93 million hole wasn't the real problem. The real problem was what Stream had built around it.

The loop

Stream's core strategy was recursive leverage, which they dressed up as "yield arbitrage." Here's how it worked: deposit xUSD as collateral, borrow stablecoins like USDT0, use those stablecoins to mint more xUSD, deposit that xUSD as collateral, borrow more, repeat. One dollar could become several dollars of xUSD through enough loops.

Days before the collapse, on-chain analyst under the pseudonym Cbb0fe flagged that Stream's supporting assets totaled around $170 million while borrowing hit $530 million. That's over 4x leverage through their looping strategy. Stream became the largest holder of its own token, not because of organic demand, but because they'd engineered a circular dependency where the collateral backing the asset was the asset itself.

This is the kind of structure that looks genius in a bull market and catastrophic the moment anything breaks. And something broke.

The leverage loop was bad enough. The oracle design made it exponentially worse.

On Euler markets running on Plasma, $107 million of xUSD sat as collateral with the oracle hardcoded at $1.27. Not reflecting market prices. Not responding to reality. Just a number in the code that said "this is worth $1.27" while the market was screaming that it wasn't.

When xUSD started depegging, positions couldn't liquidate. The oracle refused to acknowledge what was happening. Lenders have been stuck with bad debt because the system was designed to ignore price discovery. Meanwhile, borrowers faced 75% APY that only deepens the hole, and liquidations sat frozen waiting for bankruptcy proceedings to resolve what functioning risk systems should have handled automatically.

This might end up being the largest single loss for DeFi lenders from a non-exploit event. And it was entirely by design. Someone chose to hardcode those oracles. Someone decided that pretending volatility doesn't exist was safer than dealing with it. That person was wrong, and now thousands of users are paying for it.

There's a detail here that matters: Stream's "missing transparency report." When you're managing $382 million and promising 18% yields, your users aren't customers but creditors taking your counterparty risk. They deserve to know what backs their deposits.

The uncomfortable question: did Stream's team actually understand their own risk? When you're running recursive loops across multiple chains with bespoke agreements and external managers, can you even map your dependencies? Or had the structure become too complex for anyone to fully comprehend? 

DeFi is excessively,, and paradoxically, concentrated and thus, systemic. One protocol's excessive leverage becomes everyone's problem. Borrow rates spike. Liquidations trigger elsewhere. Liquidity evaporates. Each unwind forces another. The timing didn't help: coming right after the $100 million Balancer exploit and broader market weakness, confidence was already fragile.

Patterns we keep ignoring

This script is familiar: Terra/Luna, Celsius, Three Arrows Capital. Unsustainable yields, recursive leverage, opacity about risk, catastrophic unwind. We've seen this movie before, and yet here we are again, watching the same plot unfold with different characters.

DeFi has sophisticated primitives but immature risk management. We can build complex financial instruments on-chain, but we haven't built the culture of prudence that prevents disasters like this.

This isn’t bad luck. This is a disaster engineered through negligence and growth-at-all-costs mentality.

Every bull market follows this pattern: leverage builds, yields compress, someone reaches for more risk, something breaks, contagion spreads, deleveraging cascades. We shouldn't be surprised. The question is whether we'll learn from it.

Based on history, it’s fair to show scepticism. But there are reasons for cautious optimism. The pain from events like this changes behavior, slowly but surely. More importantly, the arrival of mature institutional players with actual risk frameworks and regulatory oversight should eventually force these practices out. Traditional finance learned these lessons through centuries of crises. DeFi is compressing that education into years, and institutional capital won't tolerate the kind of opacity and reckless leverage that just imploded Stream.

The cowboys running recursive loops with hardcoded oracles will either adapt or get crowded out. Not because they suddenly develop prudence, but because capital flows toward sustainable risk-adjusted returns, not 18% yields backed by circular dependencies. That process takes time, and we'll see more failures along the way. But each xUSD makes it harder for the next one to raise funds.

The question is whether retail gets wiped out first, or whether the industry matures fast enough to protect them. Right now, it's a race.

Written by
Jérémy Le Bescont Author Picture
Jeremy Le Bescont
Published on07 Nov 2025

Welcome to CoinShares

Personal data

0102

When you visit CoinShares website, cookies enhance your experience. They help us to show you more relevant content. Some cookies are necessary for the site to work and will always be active. Blocking some types of cookies may impact your experience of the website and the services which we offer on our website.

We use cookies on our site to optimize our services. Learn more about our EU cookie policy or US cookie policy.

  • Necessary
    Question circle icon
  • Preferences
    Question circle icon
  • Statistical
    Question circle icon
  • Marketing
    Question circle icon