Is Restaking Really Risk-Free? Real 2026 Cases
In the first year of restaking, the dominant pitch was “you don’t change your ETH staking principal, you just stack a few AVS yields on top” — basically a free lunch. By 2026 the bills are arriving: slashes, LRT depegs, Operator operational failures, contract bugs. None of these wiped out the category, but together they removed the words “no risk” from the conversation.
This piece is not a framework. It walks through three event types that have actually happened, with rough loss estimates so you can hold a ruler up against your own position.
Quick vocabulary check before the case studies
So the rest of the article reads fast:
- Restaking: re-pledging already-staked ETH (or its LST form) to an AVS, accepting the AVS’s extra slash conditions.
- AVS (Active Validation Service): the service that consumes that restaked security — EigenDA, Hyperlane, MACH.
- Operator: the entity running the AVS software on your behalf.
- LRT (Liquid Restaking Token): a tokenized restaking position, e.g. ezETH, weETH, rsETH.
- Slash: principal is partially burned for breaking protocol rules.
A wider picture is in the EigenLayer primer and in the underlying ETH staking risks. Here we only look at “what has gone wrong”.
Risk type 1: AVS slashes actually fire
Before 2024 EigenLayer had not turned on full slashing yet, and a popular take was “it’ll never really cut funds”. 2026 broke that assumption — real slashes happened on mainnet.
A representative case: a fast-finality AVS required its Operators to sign L2 blocks within 6 seconds. One Operator ran a version mismatch between the node and the AVS client, missed 30 minutes of signatures, and triggered a “continuous unavailability” slash worth 2% of delegated assets.
The direct victims were every restaker delegated to that Operator — they shared that 2%. Suppose you had 1 ETH delegated:
| Item | Value |
|---|---|
| Delegation | 1 ETH |
| Slash ratio | 2% |
| Actual loss | 0.02 ETH |
| Approx. USD value | ~100 USD |
| Months of AVS yield to recover | 5-8 months |
The point is not the absolute number, it is that one event wipes out most of a year’s extra return. That is the biggest experiential gap versus plain LSTs.
Risk type 2: LRTs depeg in the secondary market
Most retail users hold LRTs, so depegs are far more frequent than slashes. Between late 2025 and early 2026 there were at least three episodes of mainstream LRTs trading below 95% of NAV, with the worst case hitting around 90.2%.
The triggers are not exotic:
- Redemption queue congestion — Ethereum’s withdrawal window already adds days, restaking’s unbond queue stretches the path further than stETH.
- AVS-related panic contagion — once the market suspects an AVS used by some LRT is in trouble, DEX sellers move first.
- Leverage loop unwinds — many users had LRTs in Pendle / Morpho / Aave loops, the price dip triggered chained liquidations.
For an unleveraged holder the depeg is a paper loss — barring a real protocol incident, redemption eventually returns 1:1 ETH. For a leveraged holder the same event can wipe out 30-50% of principal in one go.
Risk type 3: Operators do not “run away”, but they do fail
This category was barely discussed in 2024. By 2026 it has happened multiple times. Operators cannot actually take your funds — they don’t have that authority — but their operational quality directly drives your loss:
- Datacenter outage past threshold → missed AVS tasks → slashes.
- Wrong software upgrade → signature format mismatch → some AVS triggers a “false proof” slash.
- Too many AVS without risk control → a bug in one AVS bleeds through and reduces payouts from the well-behaved ones.
There is a public 2026 case: a top-10 Operator skipped upgrading an older AVS client during a team handover, accumulated 1,200 missed signatures in two weeks, and was deducted 0.6% of delegated assets. Every delegator absorbed that loss passively, having done nothing wrong.
The takeaway forming in the community — Operator concentration risk is the same shape as exchange concentration risk. Too much weight on one is the problem.
Risk type 4: smart contracts and governance
Restaking adds two more contract layers — EigenLayer core + AVS contract + LRT contract. No catastrophic core bug fired in 2024-2026, but:
- One LRT redemption logic bug delayed withdrawals by 12 hours and triggered one of the depegs above.
- One AVS governance vote passed by a single large holder bumped the Operator cut from 5% to 12%; delegators only had “exit or accept” as choices.
- Several LRT protocols silently added AVS to their plug-in list without clear disclosure, and some users only discovered post-hoc that they were exposed to an AVS they would not have opted into.
None of these were overnight zeros, but together they form a slow erosion — you think the risk is bounded, while governance is quietly rewriting your exposure.
A rough loss distribution table
Aggregated across 2024-2026, here is what a median restaker has absorbed:
| Risk type | Frequency | Per-event loss | Annualized expected loss |
|---|---|---|---|
| AVS slash | Occasional | 1-3% of delegation | 0.05-0.2% |
| LRT depeg, no leverage (time cost only) | Once every few months | 0 eventually | ~0 |
| LRT depeg, leveraged liquidation | 1-2 industry-level / year | 20-50% of position | Depends on leverage |
| Operator operational incident | Occasional | 0.5-1% of delegation | 0.05-0.1% |
| Smart contract / governance event | Rare but real | Variable | Hard to quantify |
An unleveraged user’s annualized expected loss is around 0.15-0.4%, which should be subtracted from the 4-5% blended yield calculated in the AVS status piece to get a real risk-adjusted return.
A checklist for the everyday restaker
If you don’t plan to become a full-time on-chain risk analyst, this is the event-tested minimum action set for 2026:
- Stop treating restaking as pure passive yield — check AVS changes and Operator performance monthly.
- Refuse to put LRTs into 5x or higher leverage loops — leverage is the actual amplifier in 90% of disaster cases.
- Diversify Operators, don’t park everything with the top-ranked one.
- Keep at least 30% of your position in native stETH or ETH as a fallback when restaking goes wrong — see Lido vs Rocket Pool for the boring side of the allocation.
- Watch whether LRT redemption channels are clear — redeemability matters far more than APY.
Read the bill before you stay at the table
Restaking in 2026 is no longer the “free stacking” story. It is a real trade with extra layers of risk and a small extra layer of yield. The trade can still be worth it, but only if you are willing to read AVS docs, monitor Operator distribution histories, and watch LRT prices. If you just want quiet ETH yield, the plain staking guide path is genuinely the better fit.
Take the bill first, then decide whether to keep your seat — that is the right 2026 attitude.