EigenLayer launched two years ago with a thesis that has aged unevenly. The promise — let staked ETH be rehypothecated to secure new applications, expanding the addressable market for Ethereum's economic security — generated enormous pre-launch attention, multi-billion-dollar TVL within months, and a sprawling ecosystem of "actively validated services" (AVSs) that were going to consume that security.
The TVL is still there. The operational footprint of services genuinely consuming restaked security is much narrower than the original vision implied. The protocol is now in the middle of a quiet repositioning around a smaller set of services that have a credible chance of growing into something durable.
What Was Supposed to Happen
The original EigenLayer thesis ran like this. Many crypto applications need security guarantees. Bootstrapping a fresh proof-of-stake validator set is hard, expensive, and slow. EigenLayer would let those applications "rent" Ethereum's security by paying restakers to validate their service in addition to the Ethereum chain.
In the maximalist version of the thesis, hundreds of services would emerge across data availability, oracle networks, bridge security, MEV management, and other infrastructure categories. Each would compete for restaker attention by offering yield. Restakers would optimize across services. The total stake securing the EigenLayer ecosystem would compound, and EigenLayer would become a foundational coordination layer for crypto infrastructure.
Two years in, the maximalist version has not materialized.
What Actually Happened
A handful of services launched and achieved meaningful operational scale. EigenDA, the data availability service built by EigenLayer's parent organization, has gained adoption among rollups looking for cheaper data publishing than Ethereum mainnet. A few oracle and bridge-security services have gone live with material stake securing them.
The total set of operationally meaningful AVSs is small — well under twenty as of early 2026. Many of the services that announced plans to launch on EigenLayer either delayed indefinitely, scaled back ambitions, or quietly pivoted to alternative security models. Several launched but failed to attract sufficient operator participation to be considered economically secured.
The mismatch between TVL and operational utilization is the crux of the issue. Restakers locked ETH into EigenLayer expecting yield from a vibrant ecosystem of services. The yield largely did not materialize at the levels that would justify the lock-up. The opportunity cost of restaked ETH — both in terms of foregone alternative yields and in terms of slashing risk for poorly-managed services — turned out to be higher than initial models predicted.
What's Changed in the Repositioning
EigenLayer's team and ecosystem have shifted strategy in several visible ways.
Slashing rollout. For most of the protocol's first year, slashing was not enabled in production. Without slashing, the security guarantees AVSs could offer were notional rather than real. Slashing went live in stages over the past 12 months, and is now active for a meaningful subset of services. This is necessary for EigenLayer's long-term thesis, but it has also surfaced operational questions — about how operators handle slashing risk, how AVSs design slashing conditions safely, and how restakers price the risk into their yield expectations.
Operator concentration. The protocol has tried to mitigate operational risk by encouraging operator diversity, but the practical reality is that a small number of well-funded operators run the bulk of restaking infrastructure for serious AVSs. This concentrates operational risk and creates correlated failure modes across services that share the same operator set.
Curation layers. A small number of "restaking curators" have emerged that take EigenLayer-deposited ETH and decide which AVSs to allocate it to, similar to how Yearn or other DeFi yield optimizers route capital. These curators add a useful abstraction layer for restakers who don't want to manage AVS selection themselves, but they also add another layer of intermediation and another set of risks.
Realistic positioning. EigenLayer's public messaging has moderated significantly from the early days. The team now talks about restaking as one tool among many for bootstrapping security, rather than as a foundational primitive that will reshape all of crypto infrastructure. That moderation is healthy and probably accurate.
The Real Use Case
What EigenLayer has demonstrated is that restaked security is a viable model for a specific category of services: those that need a large, distributed validator set quickly, can tolerate the operational complexity of EigenLayer's stack, and have a clear economic logic for paying restakers a meaningful yield.
EigenDA fits this profile cleanly. Data availability is a real, commercially valuable service; it benefits from a large validator set; and it has paying customers (rollups) who generate revenue that can fund restaker yield. A handful of bridge and oracle services fit similarly.
What EigenLayer has not demonstrated is that restaked security generalizes to dozens or hundreds of services. The cost structure, complexity, and slashing risk make EigenLayer a poor fit for many service types that early discussions assumed would naturally land there.
The honest read at two years is that EigenLayer is a useful piece of crypto infrastructure for a narrow set of services. That's a meaningful achievement. It is also a substantially smaller outcome than the original vision implied.
What's Next
The protocol's growth trajectory from here depends on a few questions.
Whether EigenDA continues to gain rollup adoption and generates enough fees to support a meaningfully scaled operator set will determine whether EigenLayer has at least one anchor service that justifies the protocol's existence at scale.
Whether other AVS categories — particularly those related to MEV management, liquid restaking integration, and cross-chain coordination — can find economically viable structures will determine whether restaking becomes a broad infrastructure category or remains a niche tool.
Whether the liquid restaking token ecosystem (Ether.fi, Renzo, Puffer, and others) consolidates or fragments will affect both the user experience of restaking and the political economy of how restaking yields get distributed.
The thesis hasn't been disproven. It's been narrowed. The next phase of EigenLayer is less about whether restaking is a good idea and more about whether the protocol can scale a small set of real services into something durable.
The TVL says yes; the operational reality says it's still a question.