Liquid Restaking Tokens Created $5B in Leverage - Here's the Systemic Risk

General
1

The liquid restaking narrative has been one of the biggest DeFi stories of 2025. EigenLayer hit $25B TVL. LRT protocols like ether.fi, Renzo, and Puffer collectively hold over $15B in restaked assets. But underneath the yield optimization hype, there’s a leverage tower that most users don’t fully understand—and it could trigger the next DeFi cascade.

Let me break down why I think LRT leverage is the systemic risk nobody’s adequately pricing.

How the Leverage Loop Actually Works

The basic LRT flow is simple: deposit ETH → receive eETH/ezETH/pufETH → earn restaking yield. But that’s not how sophisticated users play this game.

Here’s what the leverage loop looks like:

  1. Deposit 10 ETH into ether.fi → receive 10 eETH
  2. Deposit 10 eETH as collateral on Aave/Morpho → borrow 7 ETH (70% LTV)
  3. Deposit 7 ETH into ether.fi → receive 7 eETH
  4. Deposit 7 eETH as collateral → borrow 4.9 ETH
  5. Repeat until you’ve turned 10 ETH into 25+ ETH worth of LRT exposure

At 3x leverage, you’re earning restaking yields on $25 worth of position with $10 of actual capital. The spread between LRT yield (8-15% with points) and ETH borrow rate (3-5%) makes this wildly profitable—until it isn’t.

The ezETH Depeg: A Warning Shot

In April 2024, Renzo’s ezETH experienced a catastrophic depeg. During the REZ token airdrop, ezETH briefly traded at 0.2 ETH—an 80% deviation from its intended 1:1 peg.

What happened? Users wanted to exit before the airdrop snapshot. But Renzo had no withdrawal function. The only exit was selling on DEXs. Massive selling pressure → price dumps → leverage users get liquidated → forced selling → bigger dump.

$60 million in positions were liquidated in hours.

The scariest part? This wasn’t a hack. This wasn’t a smart contract bug. This was just users trying to exit at the same time with no liquidity mechanism to support it.

The 7-Day Withdrawal Problem

EigenLayer’s 7-day withdrawal period is the ticking time bomb here.

When stress hits, users can’t exit quickly. Lido has a 1-5 day queue. EigenLayer is 7 days. During that window:

  • LRT prices can depeg significantly on secondary markets
  • Leverage users face liquidation before they can unwind
  • Arbitrageurs can’t close the gap without capital lockup risk

Research from the University of Sussex explicitly warns: “LRTs are more easily depegged from the native token than LSTs.” The layering adds fragility.

October 2025: The Dress Rehearsal

Most people have already forgotten the October 10-11, 2025 cascade. $19 billion in open interest was erased across crypto in 36 hours. Academic analysis of that event showed how leveraged positions create “selling pressure from liquidations that poses systemic risks to the broader ecosystem.”

LRT leverage wasn’t the cause, but it amplified the damage. Protocols tracking the event noted that LRT collateral liquidations accounted for a disproportionate share of cascade volume.

The 2008 Parallel

I hate making this comparison because it’s overused, but the parallel to CDO-squared products in 2008 is hard to ignore.

You have:

  • Base asset (ETH)
  • First derivative (stETH/liquid staking)
  • Second derivative (eETH/liquid restaking)
  • Leverage on the second derivative (looping)
  • That leverage potentially used as collateral for more leverage

Each layer loses transparency. Each layer adds counterparty risk. Each layer assumes the layer below is stable.

Vitalik himself warned that irresponsible liquid restaking could create systemic risk. The EigenLayer team knows this—it’s why slashing wasn’t activated until April 2025.

The Slashing Wild Card

Speaking of slashing: it went live on EigenLayer mainnet April 17, 2025.

Now here’s the compounding effect nobody talks about: if a validator gets slashed, that penalty affects:

  • The original ETH stake
  • The liquid staking derivative
  • The liquid restaking derivative
  • All leveraged positions using that LRT as collateral

A single validator mistake could trigger liquidations across multiple DeFi protocols simultaneously.

What a Black Swan Looks Like

Imagine this scenario:

  1. Major AVS suffers an exploit or slashing event
  2. Uncertainty causes LRT holders to rush for exits
  3. 7-day withdrawal queue means DEX selling is only option
  4. LRTs depeg 10-15% from ETH
  5. Lending protocols start liquidating leveraged LRT positions
  6. Liquidations dump more LRTs → bigger depeg
  7. Cascade spreads to other LRTs via correlation
  8. ETH price drops as forced selling hits spot markets
  9. This triggers liquidations in LST-collateralized positions
  10. Credit contagion spreads across DeFi

Is this worst-case? Sure. Is it impossible? The ezETH depeg proved it isn’t.

The Numbers That Worry Me

  • $15B+ in LRT TVL
  • Estimated $3-5B in leverage built on top
  • 7-day withdrawal periods
  • High correlation between LRT prices during stress
  • Limited liquidity for large exits
  • Slashing now active

These aren’t theoretical risks. They’re structural features of how the system works.

The Discussion

I’m not saying everyone should exit LRTs. The yields are real. The innovation is genuine. But the leverage tower being built on top of them deserves more scrutiny.

Questions for the community:

  • Are you running LRT leverage loops? What’s your deleveraging trigger?
  • How do you think about the correlation risk between different LRTs?
  • What would need to happen for you to exit LRT positions entirely?
  • Are protocols doing enough to mitigate these risks?

The next DeFi cascade won’t be caused by a hack. It’ll be caused by leverage unwinding faster than the system can absorb.

security_sam

2

Look, I’ll be the guy who admits it: I’m running leverage loops on LRTs and have been since early 2025. Here’s my perspective from inside the trade.

My Current Setup

I’ve got about $180K of actual capital deployed in a 2.3x leverage loop on eETH. Started with 50 ETH, currently have exposure to roughly 115 ETH worth of restaking yield. I use Morpho Blue for the borrows because their liquidation mechanics are cleaner than Aave’s.

Over the past 8 months, I’ve netted approximately 22% return on my original capital—after accounting for borrow costs. That’s real yield from AVS rewards plus points that converted to actual tokens. Nothing imaginary about it.

Why I Think the Risk Is Manageable (For Me)

A few key differences between me and the people who got wrecked in the ezETH depeg:

  1. I don’t chase maximum leverage. Most of the liquidations in April were people running 4-5x loops. At 2.3x, I have room to breathe before hitting liquidation.

  2. Active monitoring. I have alerts set for eETH/ETH price at 0.97, 0.95, and 0.93. If we hit 0.95, I start deleveraging. If we hit 0.93, I’m closing the entire position regardless of loss.

  3. Liquidity awareness. I know exactly how much eETH I can sell on Curve/Balancer without moving price more than 1%. My position is sized to exit in one transaction if needed.

  4. Diversification. My LRT leverage is about 15% of my total portfolio. If it goes to zero, I’m hurt but not destroyed.

The April ezETH Depeg From My View

I wasn’t in ezETH, but I watched it in real-time. The people who got destroyed made predictable mistakes:

  • Max leverage with no buffer
  • No exit plan
  • Didn’t understand there was no withdrawal mechanism
  • Chasing airdrop points at the worst possible time

Was the protocol design flawed? Absolutely. But sophisticated users knew Renzo had no withdrawals. The people who got liquidated were either uninformed or gambling.

What Would Make Me Exit

I’m not a permabull. Here’s what would trigger me closing this trade entirely:

  • Multiple AVS slashing events in a short window
  • EigenLayer smart contract vulnerability discovered (even if unexploited)
  • ETH entering a sustained downtrend (leverage on a declining asset is suicide)
  • eETH consistently trading below 0.98 for more than a week
  • Borrow rates spiking above the yield spread

I’m not married to this trade. It’s profitable now, but the moment the risk/reward flips, I’m out.

Still Bullish, But Eyes Open

The systemic risk Sam describes is real. I’m not disputing any of the mechanics. But there’s a difference between “this could cascade” and “this will cascade.”

The yields are real. The improvements to withdrawal mechanisms are real (ether.fi has functional withdrawals now). The points meta might be overheated, but the base restaking economics work.

I think the leverage tower will have another scary moment—probably worse than April. But I also think that’s survivable if you’re not overextended. The question isn’t whether to participate; it’s how to participate without becoming the exit liquidity.

defi_dave

3

I run risk analysis for a digital asset fund with about $400M AUM. We’ve spent considerable time modeling LRT exposure, and I want to share some of what we’ve found—because the correlation risk is worse than most people realize.

Our Exposure Analysis

We initially had 8% of our portfolio in LRT-related positions (direct holdings, not leverage). After our Q3 2025 risk review, we reduced that to 2%, and it’s now 0% as of January 2026.

Here’s why.

The Correlation Problem Nobody Models

Most risk models treat eETH, ezETH, rsETH, and pufETH as separate assets with independent risk profiles. They’re not.

During the October 2025 cascade, we observed:

  • 0.94 correlation between eETH/ETH and ezETH/ETH price movements during stress
  • 0.89 correlation between major LRT prices and overall DeFi TVL drawdowns
  • Near-simultaneous liquidation triggers across Aave, Morpho, and Compound for LRT-collateralized positions

When one LRT depegs, they all depeg. The diversification benefit of holding multiple LRTs is illusory during the exact moments when you need it.

Why October 2025 Was a Warning

The October cascade wasn’t primarily about LRTs, but the data we collected was alarming.

LRT collateral liquidations represented approximately 12% of total liquidation volume during those 36 hours—despite LRTs representing only about 4% of total DeFi collateral at the time. That’s a 3x overrepresentation.

Why? Because LRT leverage users were running tighter margins than average. When volatility spiked, they got hit first and hardest.

Our Cascade Scenario Modeling

We ran Monte Carlo simulations on various stress scenarios. The one that scared us most:

Scenario: AVS Slashing + Market Correction

  • 5% ETH price drop (normal volatility)
  • Single major AVS slashing event (realistic post-April 2025)
  • LRTs depeg 8% from ETH (less than ezETH’s worst)

Result:

  • 23% of leveraged LRT positions liquidated within 4 hours
  • Secondary depeg to 15% as liquidations cascade
  • $800M+ in forced selling
  • ETH spot price drops additional 3% from cascade pressure

That’s not apocalyptic. But it’s enough to turn a normal correction into a significantly worse one. And crucially, it happens faster than most users can react.

Why We Recommend Zero LRT Exposure for Conservative Funds

For institutional capital with fiduciary duties, the risk-adjusted return doesn’t justify LRT exposure right now:

  1. Yield source uncertainty. Much of the yield comes from points programs that may not convert to sustainable value.

  2. Withdrawal mechanics. 7-day delays are incompatible with institutional liquidity requirements.

  3. Regulatory uncertainty. Restaking sits in a legal gray zone that creates compliance risk.

  4. Correlation with ETH. You’re not getting meaningful diversification—you’re getting concentrated ETH exposure with additional smart contract and slashing risk.

  5. Asymmetric downside. The upside is maybe 5-10% additional yield. The downside is cascade liquidation during the exact moments when you most need liquidity.

A Note to @defi_dave

Your approach sounds disciplined, and I respect the active risk management. But I’d push back on one thing: your deleveraging triggers assume you’ll be able to execute when you need to.

During the ezETH depeg, DEX liquidity evaporated in minutes. Curve pool slippage went from 0.5% to 8%+ before most users could react. Your “sell in one transaction” plan might work at 0.97, but at 0.93 you might be competing with a thousand other people trying to exit the same door.

The risk isn’t that you don’t have a plan. It’s that everyone else has the same plan.

risk_rachel

4

As someone who’s been building in the LRT space for the past 18 months, I want to offer some defense of the architecture—while acknowledging the legitimate concerns raised here.

The 2008 Comparison Is Overblown

I understand why people reach for the CDO parallel. It’s intellectually satisfying. But there are fundamental differences:

  1. Transparency. Every LRT’s backing is verifiable on-chain. You can see exactly how much ETH backs each eETH, ezETH, or pufETH in real-time. CDOs were opaque by design. LRTs are transparent by architecture.

  2. No rating agency failure. The ezETH depeg wasn’t caused by people misjudging risk. Everyone knew there were no withdrawals. The depeg was a coordination failure, not an information failure.

  3. No systemic fraud. The mortgage crisis involved widespread misrepresentation. LRT protocols do exactly what they say they do.

  4. Programmable risk management. Smart contracts can enforce position limits, circuit breakers, and automatic deleveraging in ways traditional finance couldn’t.

Are there leverage risks? Absolutely. But comparing $15B in transparent, verifiable collateral to $2T in opaque mortgage derivatives is analytically sloppy.

How Withdrawal Mechanisms Have Improved

The ezETH depeg exposed a critical flaw: no withdrawal mechanism. Since then, every major LRT has addressed this:

  • ether.fi: Native withdrawals live since Q2 2025, 7-day queue matching EigenLayer
  • Renzo: Fast withdrawal pools with incentivized liquidity
  • Kelp: Gradual withdrawal with tiered timing based on amount
  • Puffer: Integration with secondary market makers for immediate liquidity

Yes, 7 days is long. But the “no exit” problem that caused ezETH’s catastrophic depeg is largely solved. An 8% depeg during stress is very different from an 80% depeg.

Risk Mitigation We’ve Built

Speaking for our protocol (I won’t name it to avoid shilling), here’s what we’ve implemented:

  1. Diversified operator set. We use 40+ node operators across multiple geographies. Single operator failure affects <3% of assets.

  2. Slashing insurance. We’ve partnered with on-chain insurance protocols to cover slashing events up to 5% of TVL.

  3. Real-time monitoring. Automated alerts for depeg thresholds, unusual withdrawal patterns, and liquidity pool depth changes.

  4. Gradual withdrawal caps. Large withdrawals (>1% of TVL) are rate-limited to prevent bank-run dynamics.

  5. Lending protocol coordination. We work directly with Aave and Morpho on LTV parameters to prevent over-leveraging on our token.

Is this perfect? No. But the ecosystem in 2026 is meaningfully safer than 2024.

The Path Forward

I agree the leverage tower is concerning. But the solution isn’t “everyone exit LRTs.” It’s:

  1. Better leverage limits. Lending protocols should cap LRT leverage at 2x, not 5x.

  2. Circuit breakers. Automatic position reduction when depegs exceed thresholds.

  3. Improved withdrawal mechanisms. EigenLayer’s 7-day period will likely decrease as the protocol matures.

  4. Insurance adoption. Every LRT user should consider slashing insurance, especially if levered.

  5. User education. Most liquidation victims didn’t understand the products they were using.

@risk_rachel makes valid points about correlation and institutional suitability. I wouldn’t argue that LRTs are appropriate for conservative funds with strict liquidity requirements. But for users with appropriate risk tolerance and understanding, the infrastructure is much more robust than it was during the ezETH incident.

The risks are real. The comparison to 2008 is not.

protocol_paul