Ethereum Restaking and EigenLayer Explained: How It Works, What It Pays, and What Could Go Wrong

Most financial innovations in crypto arrive with grand claims about changing the world and then spend several years quietly figuring out whether they actually work. Ethereum restaking, as implemented through EigenLayer, is one of the few that arrived with a genuinely novel idea and then proceeded to accumulate $19.7 billion in total value locked before most people had fully understood what it was.

The concept is, at its core, elegant. Staked ETH already secures the Ethereum network. What if that same staked ETH could simultaneously secure other protocols, earning additional yield in return? What if the cryptoeconomic security that took years and billions to build on Ethereum could be extended to new applications without each of them having to rebuild it from scratch?

That is the proposition. EigenLayer is the implementation. And with a 93.9% share of the restaking market and over 4.6 million ETH committed, the capital markets have indicated that the proposition is, at minimum, interesting enough to deploy substantial money against.

This piece explains what restaking actually is, how EigenLayer's Actively Validated Services model works, what the real yield numbers look like, who the key players in the liquid restaking ecosystem are, and where the risks are that the TVL headline does not advertise.

What Restaking Is and Why It Exists

To understand restaking, it helps to briefly understand why Ethereum staking works in the first place.

Ethereum uses proof-of-stake consensus. Validators lock ETH as collateral and participate in block validation. If they behave honestly, they earn staking rewards. If they behave dishonestly, such as by attempting to validate conflicting transactions, they lose a portion of their stake through a mechanism called slashing. The collateral at risk is what makes the system trustworthy. Validators have a strong financial incentive not to cheat because cheating is expensive.

This security model is powerful, but it is also constrained to Ethereum itself. Every other protocol that wants to operate with cryptoeconomic security has historically had to build its own validator set, issue its own token, and convince enough holders to stake it. That is expensive, slow, and produces fragmented security across hundreds of protocols rather than a deep, consolidated pool.

EigenLayer's insight was that Ethereum validators could opt into securing additional protocols using their already-staked ETH, subject to additional slashing conditions specific to those protocols. The validators earn extra yield. The new protocols get cryptoeconomic security without bootstrapping their own. In theory, everyone benefits.

The additional protocols that validators secure through EigenLayer are called Actively Validated Services, or AVSs. The AVS model is the core of how EigenLayer works.

How AVSs Work

An AVS is any system that requires a set of distributed validators to perform some off-chain computation or verification task and needs those validators' honesty to be guaranteed by economic incentive. The category is broad by design.

EigenDA is the flagship AVS and currently receives the majority of restaked ETH security. It is a data availability layer that helps Ethereum rollups store transaction data more cheaply than posting it directly to the base chain. Rather than paying Ethereum's full data posting costs, rollups can use EigenDA's restaking-secured system at significantly lower cost, with the security provided by the restaked ETH backing EigenDA validators.

EigenAI and EigenCompute went live on mainnet in late 2025. EigenAI provides verifiable AI inference, while EigenCompute handles off-chain execution verification. Both represent what analysts are calling the Vertical AVS trend in 2026, where AVSs specialise in specific validation types rather than offering generalised security services. AI verification is particularly active in this space, as the need to verify that an AI computation was performed correctly and honestly maps well onto the AVS model.

Other notable AVSs include AltLayer for rollup infrastructure, Brevis and Lagrange for cross-chain messaging and data, and WitnessChain for DePIN coordination. Each defines its own slashing conditions, its own fee structure, and its own requirements for the operators validating it.

By early 2026, over 1,900 active operators are running services through EigenLayer, all of whom have opted into the additional slashing conditions associated with the AVSs they validate. The current base restaking reward sits at approximately 3.87% annually on top of standard ETH staking yield, though the number varies with AVS demand and fee revenue. Ethereum solo staking currently offers a base yield of approximately 2.8% to 3.2%, making the combined restaking return meaningful for validators who are comfortable with the added risk.

The Liquid Restaking Ecosystem

Native restaking, where a validator directly opts their staked ETH into EigenLayer, requires running your own node and managing your own keys, plus a minimum of 32 ETH. This is not where most retail capital enters the ecosystem.

Liquid restaking tokens (LRTs) solve the access problem. A liquid restaking protocol accepts deposits of ETH or liquid staking tokens like stETH, restakes them through EigenLayer on the depositor's behalf, and issues a token representing the restaked position. That token can then be used in DeFi, posted as collateral, or traded, while the underlying ETH continues earning both base staking rewards and restaking rewards.

EtherFi leads the liquid restaking market with approximately $5.6 billion in TVL, which makes it the largest LRT protocol by a substantial margin. Kelp DAO, Renzo, and Puffer Finance follow, with the top liquid restaking protocols collectively accounting for over two-thirds of EigenLayer's total deposits.

The LRT layer adds a layer of convenience but also adds a layer of complexity and risk. The retail user depositing into an LRT protocol is not directly choosing which AVSs their ETH secures. That decision belongs to the LRT protocol's governance and operators. The yield shows up in the token's staking rewards. The risk is buried in smart contract relationships that most holders will not read, because reading smart contract relationships is how people spend their Friday evenings in a very specific subset of the crypto community.

EigenLayer has proposed an operator safety score and AVS risk rating system to help users understand the risk profile of their restaking positions. As of mid-2026, these tools are still developing rather than fully deployed. This matters, and the gap between its importance and its current state of readiness is worth noting.

The Competitive Landscape

EigenLayer commands 93.9% of the restaking market, which is a dominant position by any measure. The $15.258 billion it held in TVL as of early 2026, against total Ethereum restaking market TVL of $16.257 billion, reflects both first-mover advantage and the network effects that come from having the most AVS integrations and the deepest operator ecosystem.

Symbiotic and Karak are the primary competitors. Both offer native restaking for Ethereum, and Karak has expanded to support restaking of assets beyond ETH, including LSTs and certain stablecoins. Neither has come close to challenging EigenLayer's market share, but the presence of competitors at least establishes that the restaking primitive is not permanently locked to a single implementation.

Lido, operating in the adjacent liquid staking category, maintains $18.32 billion in TVL across Ethereum, Solana, and smaller networks. Its continued dominance in simple liquid staking, as opposed to restaking, reflects the market's ongoing preference for battle-tested products over more complex ones, a preference that exists in tension with the higher yields restaking can offer.

What Could Go Wrong: The Risks Worth Understanding

EigenLayer has attracted enormous capital, which means the risk profile of what could go wrong has also grown considerably. Several structural risks deserve honest examination.

• Slashing complexity and correlated failures. Each AVS defines its own slashing conditions. An operator validating multiple AVSs is exposed to slashing risk from each of them. If multiple AVS networks experience simultaneous validation failures, whether through technical bugs, attacks, or network partitions, an operator could face penalties across multiple protocols at once. The compound risk of running many AVSs has no clean historical precedent in Ethereum staking, and the protocol's ability to fairly adjudicate complex multi-network slashing events is largely untested at scale.
• Governance of slashing disputes. Currently, slashing disputes in EigenLayer are managed through human veto committees rather than fully automated enforcement. This is a pragmatic approach given the early stage of the ecosystem, but it is also a centralisation point that becomes more significant as the value at risk grows. Whether this mechanism can scale to handle a much larger AVS ecosystem without becoming a single point of failure is an open question.
• LRT concentration risk. The concentration of restaked ETH in a small number of liquid restaking protocols means that governance decisions made by those protocols, such as which AVSs to opt into, have outsized effects on the aggregate risk profile of restaked Ethereum. A poor AVS selection decision by a large LRT protocol followed by significant slashing would impact all holders of that LRT, including retail investors who did not closely track the underlying AVS exposure.
• AVS economic sustainability. The business models of individual AVSs need to generate sufficient fee revenue to pay restakers competitively over the long term. If AVS fees do not grow to match or exceed token emissions used to subsidise early restaking rewards, the economics of the ecosystem depend on continued subsidy rather than genuine utility. This is a known transition risk in the protocol's development roadmap, not a speculation.
• Systemic Ethereum risk. Restaking increases the amount of ETH subject to slashing conditions. If a significant slashing event occurred across a widely adopted AVS, the resulting ETH liquidations could create downward pressure on the asset's price, affecting not just restakers but the broader Ethereum ecosystem. At current TVL levels, this is a theoretical concern. At higher TVL levels, it becomes a structural one.

What to Watch

EigenLayer's operator safety and AVS risk ratings. When these tools become fully operational, they will provide the first standardised framework for retail participants to understand what they are actually buying into through LRT products. The deployment timeline matters.

AVS revenue maturation. The shift from token emissions driving restaking rewards toward AVS fee revenue doing so is the critical test of whether EigenLayer's economics are sustainable. Data availability fees from EigenDA and compute fees from EigenAI and EigenCompute are the leading indicators to watch.

Competing restaking architectures. Symbiotic's multi-asset restaking model and Karak's broader collateral approach represent bets that the future of shared security involves assets beyond ETH. If either gains meaningful traction, the competitive dynamics of the sector shift.

Institutional adoption of restaking. Currently, concerns about slashing complexity and multi-token reward accounting are significant barriers to institutional participation in native restaking. If EigenLayer simplifies these sufficiently, the institutional capital sitting in standard Ethereum staking represents a substantial potential inflow.

The Bottom Line

EigenLayer has built something genuinely new and genuinely valuable. The ability to extend Ethereum's cryptoeconomic security to new protocols without requiring each of them to bootstrap their own validator set solves a real problem, and the capital that has flowed in reflects that.

The risks are also genuine and not yet fully stress-tested. The slashing mechanics are complex. The LRT layer introduces opacity between retail depositors and the actual risks their capital faces. The AVS economic models are still proving themselves. These are solvable problems, but they need to be solved, not wished away by pointing at the TVL chart.

At $19.7 billion in TVL and growing, restaking is no longer a niche experiment within Ethereum's staking ecosystem. It is a meaningful piece of infrastructure whose success or failure will affect the broader Ethereum network in ways that go beyond yield percentages. That makes understanding it properly somewhat more important than understanding which meme coin is trending this week. Only somewhat, though.