Productivity-Based Incentives on EigenLayer

X: @jaeyongp1, @Benjamin918_

Thanks to @ishaan0x and EigenLayer members for their review and discussions.

TL;DR: A new mental model to understand EigenLayer is to view it as a productivity-based incentives platform where operators can take a more active role, building on the following assumptions:

1. An Actively Validated Service (AVS) on EigenLayer is not inherently a validation network.
2. As such, operators are not bound to perform validation tasks and can instead complete individual tasks. This allows them to receive rewards proportionate to their work, freeing them from the bounds of passive yield

EigenLayer Recap: Universal Trust Marketplace

The core framework used to understand EigenLayer is the idea of decoupling trust and innovation, via a decentralized trust marketplace where AVSs can purchase trust from Stakers and tasks from Operators.

Under this framework, AVSs on EigenLayer take the form of a consensus network, where Ethereum’s cryptoeconomic security is bootstrapped to validate their networks. As in the case of sidechains, DA layers, VMs, Oracle networks, and bridges, Operators supply trust to AVSs via validation work for off-chain compute, hence the “Validated” in its name.

New mental model: Productivity-Based Incentives

We derive a new mental model shifting away from viewing EigenLayer solely as a marketplace for validation services, and viewing it as a productivity-based incentives platform. In this model, operators can be individually rewarded for any arbitrary task, guaranteed to perform honestly via the underlying cryptoeconomic security of EigenLayer.

The productivity-based model is not by itself a new idea. The Lagrange ZK Prover Network rewards operators proportionate to the number and type of zk proofs they generate. Operators can generate different types of proofs based on latency and hardware requirements. The productivity of the operator is proportional to the computational power of the software they run to generate proofs.

We extend this idea by examining two observations

1. Any AVS on EigenLayer is not by definition a validation network. An AVS on EigenLayer can consist of a diverse set of operators that can each perform any arbitrary task, including validation. However, developers today limit the scope of an AVS to a validation network, necessitating additional operator node software as a prerequisite to deploying to EigenLayer (as highlighted in Building an AVS from EigenLayer docs).
2. Accordingly, the operator’s task is not inherently bounded by validation. In other words, instead of performing identical validation tasks, operators can take a more active role as a solver. This allows them to be rewarded in proportion to their individual value creation, as opposed to passive participation in validation networks.

Operators participating in this framework become active yield generating agents where restakers underwrite the risk of the individual agents that they delegate to. The legitimacy of the operator is established via the slashing conditions that are fully onchain. Similarly, the rewards for the stakers are verified and distributed onchain.

To illustrate the new mental model in action, we look at the following use cases.

Underwriting risk for arbitrary yield generation: Restakers underwrite the risk for operators, where the delegations allow operators to access cost-free capital from collateral baskets to generate yield. Operators are incentivized to generate a minimum benchmark yield, keeping the entirety of the excess yield as a reward for participating in the system. Each operator acts independently and is responsible for sourcing restaker buy-in via strategy creation and due diligence. Given the highly profitable structure for operators, talent can be sourced to extract yield from hard-to-reach sources like MEV, arbitrage, and complex farming.

Liquidation protection: Onchain DeFiSaver, where operators are independently incentivized to rebalance debt positions. Fees can be charged by active operators for rebalancing transactions. Objective faults can be structured to penalize operators if users are liquidated or if operators use admin powers for malicious activity.

Aligning incentives

Currently, EigenLayer secures $10 billion+ in TVL. There have been growing concerns about whether the passive yield for validating AVSs will be sufficient for stakers and operators. AVSs that generate active yield will be the solution to the oversupply.

By segregating the activity of each operator and allowing them to pursue any arbitrary task, EigenLayer can attract new types of operators that are incentivized to capture active yield. Highly performant operators in each specialized domain will generate sustainable yield that far outperforms passive yield. In turn, restakers can enjoy a wide range of yield from a diversified counterparty.

The future of AVSs

The future of AVSs is unbound by a rigid set of credible commitments. Any task, regardless of the execution method, can be secured via EigenLayer. The challenge is to maintain the agency over design at the AVS level, giving builders the freedom to innovate on the stack.