The Loop Contagion

Backing a stablecoin with another stablecoin that's backed by the first stablecoin isn't revolutionary finance - it's just expensive performance art.

Stop us if you've heard this one before.

Last month we published that exact line in a piece dissecting Stream Finance's recursive mechanics.

Then a little time off… and all Hell broke loose.

Stream Finance minted synthetic dollars while Elixir minted the collateral backing those dollars - same capital allegedly spinning $14 million in tokens from under $2 million in real assets.

JPMorgan vocabulary wrapped around Terra mathematics: "market-neutral strategies," "delta-neutral positions," "institutional-grade returns."_

October 28th brought receipts instead of reforms, as Schlagonia exposed the recursive minting loop.

CBB warned of 4.1x leverage, calling it degen gambling - $170 million backing $530 million in liabilities.

November 3rd brought consequences: $93 million vanished, xUSD crashed 77% from $1.00 to $0.26, $285 million in contagion unwinding across DeFi's lending markets.

Hyperithm - managing sophisticated vaults - suddenly discovered $75 million in Stream's leveraged mHYPER positions plus $2.6 million trapped in an xUSD lending vault on Euler Plasma, which they confirmed cannot be withdrawn due to xUSD market illiquidity.

What looked like yield strategy became crisis management.

Smart money doesn't abandon opportunity - it exits when the math stops working.

If a system collapses the moment someone asks how it works, was it ever stable - or just waiting for the right question?

Stream Finance didn't show up in a hoodie promising moon missions.

Founded by Diogenes Casares through Klyra Labs in Buenos Aires.

Stream spoke the language of trading desks: "market-neutral fund strategies," "stat arbitrage strategies," "hedged top of book market making."

Casares brought credentials - former activist DAO investor who made millions by pushing underperforming startups to return money to their investors.

Now he'd build one that would become the poster child for everything he once criticized.

But Stream wasn’t the whole story. Every loop needs a counterparty - and that’s where Elixir stepped in.

Elixir Network positioned deUSD as institutional-grade rails for DeFi, backed not by cash reserves but by stETH deployed in delta-neutral strategies and exposure to T-bills through MakerDAO’s USDS protocol.

Both protocols wore legitimacy like armor, targeting the same sophisticated investors who'd watched Luna collapse and sworn they'd learned the lesson.

Stream advertised high yields - 18% APY on USDC vaults, 12% on ETH - claiming these returns came from market-neutral strategies such as lending arbitrages, dynamically hedged market-making, and incentive farming.

Elixir's deUSD offered stability as a fully collateralized, yield-bearing synthetic dollar, serving as the primary rails for institutional funds to access DeFi without changing their original asset exposure.

Public coverage later revealed that deUSD’s backing was not maintained as a strict 1:1 reserve of U.S. dollars, and some analysts argue that Elixir did not rely on centralized real‑world asset reserves.

On‑chain contracts show that Stream’s own StreamVault contract has special permission to mint new tokens as yield rounds roll - meaning Stream retains direct minting authority via yield accounting.

Both promised transparency - Stream's proof of reserves perpetually "coming soon."

Nobody reading the fine print noticed they were being sold a perpetual motion machine dressed in a tailored suit.

What exactly was spinning inside that machine?

The Loop Recapped

Our original investigation detailed how Stream and Elixir built their recursive mechanism.

The short version, based on Schlag’s research: USDC became deUSD, deUSD became leverage, leverage minted xUSD, xUSD borrowed more USDC.

Four steps that allegedly turned $1.9 million into $14.5 million in synthetic tokens.

The mechanics worked like this: Stream converted user deposits to deUSD through Elixir's minter, bridged to Avalanche or World Chain, borrowed stablecoins against deUSD collateral, swapped back to USDC, then used that borrowed capital to mint more xUSD - their own token.

Elixir completed the circle - lending millions against xUSD that Stream had just minted.

Schlag’s analysis revealed $70 million USDC supplied to hidden Morpho markets, $65 million borrowed, Elixir as sole depositor.

The loop closed: borrow USDC, mint xUSD, borrow against xUSD, mint more deUSD.

Musical chairs with the same USDC pile, marketed as two separate dollar-pegged assets.

October 27th saw three complete rounds. Those rounds raised some red flags.

So, who else was looking?

The Red Flags Were There

October 28th, CBB - a crypto KOL with enough clout to move markets - issued warnings that should have emptied vaults:

"If you have money on Morpho or Euler, withdraw from vaults with exposure to mHYPER and xUSD. This is max opacity finance. The level of leverage from xUSD is insanity."

Early warning signs were obvious - analysts later pointed out that Stream’s xUSD and Elixir’s deUSD models were structurally fragile.

Analysts eventually called out Stream’s house of mirrors - leverage stacked on leverage, recursive loops disguised as “strategy,” and zero proof-of-reserves to back any of it. The whole thing was primed to go sideways the moment reality showed up.

But were there earlier warning signs?

May 2025: A Chainlink oracle on Avalanche briefly misreported deUSD’s price at $1.03, triggering over $500K in liquidations on Euler Finance for users with deUSD debt.

August 2025: Inverse Finance voted to sunset deUSD markets over "risk and operational concerns."

DeFi protocols refusing your collateral isn't volatility - it's reputation dying in real time.

Then there is Stream's "Coming Soon" syndrome.

On its transparency page, Stream Finance declared: “Coming soon! We are currently integrating with third parties to offer proof of reserves to improve our transparency.”

Analysts noted that Stream’s promise of third‑party verification and “proof of funds” remained in perpetual hiatus, even as the protocol claimed big backing and high yields.

DefiLlama showed different TVL numbers than Stream claimed.

On October 30th, Stream clarified total assets deployed across strategies ($520M), distinguishing this from user deposits ($160M). This figure became the reference for reported “proof of assets.”

Community analysis: ~$170 million in real backing versus ~$530 million in liabilities, with a 4x leverage ratio.

By October 31st, the chatter had grown loud enough that we started asking the obvious question: was Stream’s math a House of Cards?

November 3rd: $93 million gone. No Nostradamus needed, it was painfully obvious how unsustainable it was by then.

Warnings are cheap - but what happens when the math finally collapses?

The Contagion Was Real

Our original investigation documented the circular dependencies.

YieldFi's yUSD borrowing from Morpho, serving as collateral for mHYPER, which lent to xUSD. Every protocol's solvency depending on every other.

When one link broke, 24% of yUSD's market cap vanished in 24 hours.

October 28th, Hyperithm made their exit official.

Removed ALL yUSD exposure. Removed ALL xUSD exposure. Held $10 million unleveraged as GP commitment - proof they weren't running scared, just repositioning based on data. Even provided wallet addresses for independent verification.

Sophisticated institutions don’t exit because they panic - they exit because they’ve stress‑tested what happens if liquidity dries up or borrowing costs spike.

Hyperithm’s strategic unwind hinted at just that.

Because Morpho is permissionless, much of the recursive loop risk was hidden in plain sight.

Retail depositors three degrees removed from the recursive loop had no idea their "safe" yields depended on tokens backed by 40 cents on the dollar - if that.

The vaults themselves layered risk: one vault lends, another supplies, and tokens cycle in a way that magnifies systemic fragility.

With just $170 million actually backing $530 million in borrowing - a 4.1× leverage nightmare - the “real” collateral was paper-thin.

BlockEden even estimated the collateral to be lower than ten cents on the dollar.

If redemptions surged or market conditions shifted, that circular leverage could unravel fast.

Hyperithm seemed to recognize this risk - they disclosed pulling down exposure and shared wallet addresses publicly to show how they disengaged.

But knowing isn't the same as preventing - so what did the architects of this system have to say when dominoes started tipping?

The Defenses

Stream's October 30th explanation of their recursive looping strategy:

"Recursive looping is when a protocol loops its own asset to capture a spread in interest rates. This increased yield then goes back to users either in the form of yield or our insurance fund. This allows users to earn more while taking on more risk. "

What Stream didn't disclose: the "asset" being looped was one they controlled supply of. They were both minting AND borrowing their own token.

The backing was circular. Leverage wasn't 4x on stable assets - it was 4x on already-leveraged synthetic tokens.

The insurance fund? Opaque and non‑segregated - effectively just surplus profits. Observers allege the team retained roughly 60% of deposits here, only disclosing it after scrutiny forced a public explanation.

DCF God, who disclosed a stake in Klyra (Stream's parent company), defended the mechanics:

"If there's a token paying X APR and you can loop at 0.5X APR, you should do that all day.

Then came the caveat: "I've never advised anyone to deposit in xusd (nor would I)."

His personal strategy: stopped looping at 5x leverage because he was "too lazy to deal with the bs." For Stream running the same strategy at scale with user funds: "That's their business - it's the most degen farm."

Philosophy: "If you're gonna be degen, might as well be full degen."

0xlawlol (Stream founder) in a now-deleted Oct. 29 post on Twitter, stated that Stream maintains a more than “$10 million insurance fund,” and that all positions outside the main wallet are “fully liquid immediately.” The founder also promised a transparency report on Stream’s strategies.

The coming soon was oxlawlol would ghost Twitter altogether.

"Coming soon" became the refrain, and safety theater doesn't survive stress tests - so what happened when the real test arrived?

The Inevitable Collapse

November 3rd: The Announcement.

Stream discloses $93 million loss via "external fund manager." Suspends all deposits and withdrawals. Announces Perkins Coie LLP engagement for investigation. Comments were conveniently disabled on announcement.

xUSD enters freefall: $1 to $0.50 within hours per PeckShield. Further collapse to $0.26 - a 77% loss. Some reports showed lows of $0.10.

Mass panic, aggressive swaps to USDC across DEXs as remaining liquidity evaporated.

The Catalyst Theory

November 3rd: Balancer suffers $128 million exploit. Around the same time, the Balancer hack spread panic through the market. Users began to pull their funds from Stream Finance, which created a bank run. The system could not hold under this stress.

Timing tells the story: Balancer hacked November 3rd, Stream discloses $93 million in losses November 3rd.

November 4th: The Exposure Revealed

Yields and More publishes a comprehensive creditor analysis.

Total Direct Debt: ~$285 million

Major creditors (As referenced by YAM):

TelosC: $123.64 million

Elixir: $68 million

MEV Capital: $25.42 million

Varlamore: $19.17 million

Re7 Labs: $14.26 million

Enclabs: 2.56 million

Mithras: 2.3 million

TiD: 0.38 million

Trevee: $14.7 million (detailed in November 7th post-mortem, recovered by November 14th)

YAM's analysis noted: "This is not an extensive list; there likely are more stables/vaults affected."

Protocols froze markets for xUSD, deUSD, and related assets.

Euler Finance issued a detailed FAQ making one thing clear: curator vaults ate the losses, not Euler DAO's own markets - those had "zero exposure."

The curators chose the oracles, curators chose the parameters. xUSD ran on eOracle reading contract prices. USDX/sUSDX used RedStone.

Liquidations worked exactly as coded - except when collateral turned worthless and oracles kept pretending everything was fine, nobody wanted to step in and take the hit.

Euler scrubbed affected vaults from their UI so new depositors wouldn't accidentally walk into the crime scene.

Emergency containment prevented wider contagion, but the damage spread anyway - particularly to the protocol that lent Stream the most.

November 6th: Elixir Collapses

Elixir shuts down deUSD completely. Claims Stream holds 90% of deUSD supply (~$75 million).

"Stream has decided not to repay or close positions."

Elixir processed redemptions for 80% of holders as deUSD plunged more than 97% in 24 hours.

November 7th: Trevee Post-Mortem

Publishes detailed exposure breakdown: $14.7 million across multiple products. But Trevee moved differently than other victims - immediately partnering with Telos to freeze Euler markets and segregate Stream's collateral. Within a week, they'd announced 100% recovery for splUSD holders through aggressive creditor action.

Trevee's post-mortem included a curious detail: 'Stream's losses originate either from the 10/10 drawdown or later.' The phrasing suggests uncertainty about when Stream first went underwater.

Assuming 10/10 means October 10th- then Stream operated for nearly a month knowing they were insolvent while most protocols were deleveraging.

October 10th marked the largest liquidation event in crypto history - $19 billion erased in 24 hours after Trump announced 100% tariffs on China.

While smart money was unwinding leverage. Stream went the other direction: increased leverage, borrowed across more chains, accepted fresh deposits. That's not risk management going wrong. That's something else entirely.

The Aftermath was brutal.

StableWatch data showed yield-bearing stablecoins suffered their largest outflow since UST's collapse - $1 billion total, with $411 million from xUSD alone.

DeFi curator TVL: $10.3 billion to $7.5 billion.

Largest capital loss event since Terra/UST in May 2022.

Perkins Coie's investigation continues. Stream's external fund manager remains unnamed.

Recovery estimates range from 10-30 cents on the dollar, but with $285 million in competing creditor claims against $79 million in estimated remaining assets, mathematics suggest most depositors face near-total losses.

When the music stops and everyone's holding the same worthless collateral, who gets paid first?

Round and round and round it goes.

Where it stops, $285 million in creditors are still trying to figure out.

Stream and Elixir built functional protocols. Smart contracts executed perfectly. The loop worked exactly as designed.

We warned about this house of cards - turned out we just needed to wait a week for gravity to catch up.

Terra failed loudly with algorithmic promises. Stream failed quietly with institutional vocabulary.

Protocol-controlled supply allegedly at 60%. Recursive minting that turned $1.9 million into $14 million in synthetic tokens. Proof of reserves perpetually "coming soon."

Publishing proof of reserves "coming soon" while operating insolvent isn't transparency - it's a countdown timer on fraud.

Perkins Coie's investigation continues.

Stream's external fund manager remains unnamed.

Most depositors face near-total losses while competing for scraps from $79 million in remaining assets.

Hyperithm read the on-chain data, stress-tested the collateral loops, and removed $10 million before the collapse.

Most depositors didn't have that option because they didn't know they needed it.

The contagion isn't finished either. Trevee recovered fully for some users through aggressive creditor action. Others still wait for partial recovery as Stream's legal unwinding drags on.

More bodies may surface as protocols conduct their own forensic accounting - discovering exposures they didn't know they had, or worse, exposures they hoped nobody would ask about.

This collapse exposed what's been obvious to anyone paying attention: DeFi needs risk rating agencies that actually understand recursive leverage and circular collateral.

Not curators chasing fees. Not protocols marking their own homework. Independent analysts who can decode what "market-neutral strategies" actually means when someone's minting their own collateral.

DeFi promised trustless systems where code replaced intermediaries and transparency replaced faith.

Stream and Elixir proved you don't need to eliminate trust - you just need to eliminate disclosure until sophisticated money is already repositioned.

When stability requires nobody asking questions and institutions document their exit with transaction hashes while retail reads Twitter threads about "coming soon," who exactly was this infrastructure built for?

REKT는 익명 작성자들에 의한 공공 플랫폼이며, REKT에 작성된 관점이나 내용에 대해서 그 어떤 책임도 지지 않습니다.

기부 (ETH / ERC20): 0x3C5c2F4bCeC51a36494682f91Dbc6cA7c63B514C

disclaimer:

REKT는 당사 웹 사이트의 익명의 작성자 또는 REKT에 의해 게시되거나 관련된 서비스에서 게시되는 콘텐츠에 대해 어떠한 책임도 지지 않습니다. 당사는 익명 작성자들의 행동 및 게시물에 대한 규칙을 제공하지만, 익명의 작성자가 웹 사이트 또는 서비스에 게시, 전송 혹은 공유한 내용을 통제하거나 책임지지 않으며, 귀하가 웹 사이트 또는 서비스에서 직면할 수 있는 불쾌함, 부적절함, 음란함, 불법 또는 기타 해로운 콘텐츠에 대해서도 책임을 지지 않습니다. REKT는 당사 웹 사이트 또는 서비스 사용자의 온라인 또는 오프라인 행위에 대한 책임을 지지 않습니다.