Restaking Decentralization: An Analysis on EigenLayer's Infrastructure

EigenLayer’s mainnet, live for over a year now, currently stands as the largest restaking infrastructure in the market. With over $12 billion in total value locked (TVL)—representing more than 4.5 million ETH—EigenLayer’s trust marketplace has successfully attracted 29 Autonomous Verifiable Services (AVSs) to its ecosystem.

However, a closer look at the platform’s data reveals that not all AVSs benefit equally from capital allocation. While leading AVSs such as EigenDA, ARPA Network, and Lagrange have millions of ETH staked to secure their protocols, others struggle to accumulate even a few thousand ETH.

EigenDA, MEV-commit, and Vision AVS Metrics (Source)

This raises an important question: Who decides where capital is allocated, or in other words, who really holds the keys to EigenLayer’s shared security infrastructure?

In this article, we’ll take a deep dive into EigenLayer’s architecture and analyze how capital is distributed across the protocol.

TL;DR:

• EigenLayer has $12B TVL and 29 active AVSs, but capital allocation favors top AVSs such as EigenDA.
• LRT protocols control ~80% of restaked assets, centralizing decision-making.
• 10 entities hold ~80% of restaked assets, and 7 node operators handle ~75%, raising centralization concerns.
• EigenLayer needs decentralized governance, balanced power dynamics, and permissionless onboarding to ensure long-term success.

EigenLayer Capital Allocation Architecture

Users looking to participate in EigenLayer can deposit their assets in two ways:

1. Directly through EigenLayer (manual restaking)
2. Via Liquid Restaking Protocols (LRTs)

When restaking directly on EigenLayer, users delegate their assets to operators, who are responsible for securing AVSs. Node operators are then transferred rewards for handling the restaked assets, and these rewards are distributed back to restakers after deducting operator fees.

In contrast, LRTs abstract the operator selection process, making restaking more accessible but also more centralized. Instead of manually selecting an operator, users deposit their assets into an LRT protocol, which issues a derivative token representing their restaked position. The LRT protocol itself then selects and allocates capital to a curated set of operators based on criteria such as reputation, slashing risk, or expected yield.

While this model simplifies the experience for users, it reduces individual control over where funds are allocated, which in our opinion centralizes the ecosystem and raises concerns about governance and power concentration within a few dominant LRT protocols.

EigenLayer Restaking vs LRT-Based Restaking Architecture Difference

The Current Panorama of Restaking in EigenLayer

EigenLayer’s restaking ecosystem is now dominated by LRT protocols, which control ~80% of the total restaked assets on the platform.

This shift indicates a growing preference for abstracted restaking, where users rely on LRTs instead of managing operator selection themselves. However, it also highlights an increasing level of decision-making centralization within a small number of key players.

Key Highlights of EigenLayer’s Capital Distribution:

• 10 entities control 80% of the total restaked assets on EigenLayer.
• The top 7 LRTs represent 60% of all EigenLayer restaked assets.
• Only 7 node operators capture 75% of all LRT-based restaked assets.

This level of capital concentration raises questions about network decentralization and the influence that a few key players have over EigenLayer’s future:

Top 10 EigenLayer Restaking Curators

EigenLayer Node Operators

Is EigenLayer Becoming Too Centralized?

LRTs have become the dominant force in EigenLayer’s restaking ecosystem due to their ease of use and passive earning model. However, by outsourcing operator selection, users are effectively ceding control to these protocols. This shift could jeopardize EigenLayer’s decentralization, leading to centralized governance and potential collusion risks.

While node operators show slightly better capital dispersion, the fact that just 10 operators control the majority of stake creates an oligopolistic environment where only a handful of players decide which AVSs survive. This centralized influence could limit competition and innovation within EigenLayer’s trust marketplace.

LRTs and Their Influence Over Node Selection

• Ether.fi’s top 7 node operators control 88% of its restaked assets, making it more centralized in capital distribution compared to the 75% market-wide concentration across all LRTs. However, Ether.fi remains the most decentralized LRT at the node operator level, despite being highly centralized on a curation level.
• Puffer appears to have consolidated most of its restaked assets into a single node operator, making it one of the most centralized LRTs.
• Swell, on the other hand, maintains a more balanced distribution of capital across node operators, unlike other LRTs that are highly concentrated in their validator selection.

Despite the massive potential of restaking, ensuring decentralization within EigenLayer is crucial for the long-term success of the protocol. The ecosystem needs to actively promote more decentralized governance models and incentivize permissionless onboarding for all stakeholders, including operators, restakers, AVSs, and LRT protocols - which have become such an important piece in the puzzle.

In our opinion, only by introducing the rule of asset or vault curator EigenLayer will fulfill its vision of a truly decentralized trust marketplace while maintaining its competitive.