What Is Restaking? The Yield, the Slashing, and the Part Nobody Prices

The problem it solves

A new decentralised service — an oracle, a bridge, a data availability layer — needs validators with something at stake, or its guarantees are worth nothing. Historically that meant launching a token, convincing people to buy and stake it, and hoping the resulting security budget was large enough to deter attacks that might profit more.

That bootstrap is brutal, and it produced a decade of small networks whose token market cap was lower than the value they secured — an arrangement that works right up until someone does the arithmetic.

Restaking sidesteps it. Rather than creating new staked capital, rent existing capital. Billions are already staked on Ethereum doing exactly one job; let that same capital take on a second job, with real penalties attached, and pay for the service. New protocols get meaningful security on day one. It is a real improvement over the alternative.

How it works

Three moving parts:

1. Base stake. Capital staked on the underlying network — directly, or through a liquid staking token. It already has slashing conditions: go offline, or sign conflicting attestations, and you lose part of the principal.
2. Opt-in. Through a restaking protocol, the staker commits that capital to one or more services, accepting each service's own slashing conditions on top of the base ones.
3. Payment. The services pay for the security they rent, typically in their own token or in fees. That payment is the restaking yield.

The services are usually called AVSs — Actively Validated Services — and the distinction that matters is that each one defines its own rules for what constitutes misbehaviour and how much gets slashed. Securing five services means five independent sets of conditions applying to the same capital.

EigenLayer established the category on Ethereum; Symbiotic and others have since taken varying approaches to collateral types and permissioning, and Babylon applies a related idea to Bitcoin. The differences are meaningful for a builder; for a depositor, the shape is the same.

Liquid restaking tokens

Restaking directly means locked capital and per-service decisions. Liquid restaking tokens package the whole thing: deposit, receive a token representing the position, keep it liquid, use it elsewhere in DeFi.

Convenient — and it moves one decision out of your hands. The issuing protocol chooses which services to secure. That choice is the entire risk profile of your position, and it is made by someone whose incentive is to show a competitive headline yield, which means securing more services rather than fewer. You generally cannot see the current set at a glance, cannot veto an addition, and cannot opt out of one service while staying in the rest.

There is also the familiar mechanic from liquid staking: an LRT can trade below the value of what backs it. It is redeemable only through an exit queue, so in a stress event the market price is what you can actually get, and it decouples exactly when you would want to leave. LRTs used as collateral inherit that — a discount can trigger liquidations in an entirely separate protocol.

Why the risk stacks

The core issue is that restaking risks are correlated. They are frequently presented as a list of small, independent probabilities, which makes the total look manageable. They are not independent.

A restaked position through an LRT carries, at minimum:

• Base network slashing.
• Slashing from every service secured — n sets of conditions, not one.
• Restaking protocol contract risk.
• LRT issuer contract risk.
• LRT market discount risk.
• Liquidation risk, if the LRT is posted as collateral anywhere.

Now consider what a genuine stress event looks like: a widely-used service is exploited, restaked capital is slashed, the LRT's backing falls, the LRT trades at a discount, positions collateralised by it are liquidated into thin liquidity, and the discount widens further. That is one cause arriving through six doors — the exact scenario the "independent risks" framing prices at approximately zero.

None of this makes restaking unreasonable. It makes it a leveraged position on the correctness of several systems at once, and it should be sized like one rather than like a savings account.

What happened to the yields

Restaking's growth was driven substantially by points — non-transferable scores implying a future airdrop. People were not chasing the restaking yield; they were chasing an unpriced expectation, which is why deposits far exceeded what the actual service fees justified.

As those programmes have resolved, the picture has become more honest. The real yield is whatever AVSs actually pay for security, and that is a function of genuine demand — a much smaller number than the points era implied. Anyone evaluating restaking now should separate the two cleanly: what does this pay from service fees, and what am I hoping for on top? The first is a return. The second is a lottery ticket, and worth what lottery tickets are worth.

On APY figures generally, and why a headline number often is not what it appears, see APY vs APR in crypto.