This post is about economies of scale implications that Eigenlayer has for the infrastructure level of Web3. Read for ELI5 value prop provided to both stakers and validators through introducing the concept of shared security in the ecosystem.

Feel free to skip this section, but perhaps a helpful refresher: Let’s rewind a bit back to how Ethereum actually works. When transactions are sent to Ethereum, they are effectively sequenced into blocks that are then attested to by the validator pool. When you stake 32 ETH and become a validator, you essentially get rewarded for proposing and attesting to blocks. You get slashed if you propose uncooperative behavior that could threaten the security of the chain. If you think about Ethereum as a decentralized database, what you are doing when you attempt to execute transactions as a user on Ethereum is proposing a modification to the state of Ethereum, or the state of the database and what it contains.

Given there is no central authority stating the veracity of these state changes, the PoS (Proof of Stake) system relies on a decentralized committee of ‘validators’ whose job it is to attest that the transactions coming in are correct. The more transactions validators approve at a time, the faster your transactions will be approved. Blocks on Ethereum = bundles of transactions. Side note: The more transactions a validator approves, the more computationally expensive it would be for a validator to participate. This is why blocks on Ethereum are relatively smaller than for example, blocks on Solana, and why transactions can be slower. The trade-off made here is one of security. I.e. if you have a relatively large decentralized committee of validators, this increases the probability your system is safe because more parties are validating and less likely to collude), you need to have relatively small blocks of data.

One way of thinking about the economics of ETH, is how much you would get by staking your ETH on Ethereum, which currently gains you 4% APY. The reason why this is significant is that stakers on Ethereum are actively contributing to providing and securing Ethereum security. Staking your ETH means you are providing security to Ethereum, a service for which you get paid in staking rewards. Naturally, the more value is within the Ethereum ecosystem, the more valuable the service you are providing is.

Provides valuable security-providing infrastructure blocks. The key thing here is that for any transaction or data that doesn’t already exist on Ethereum, dApps need to find another way to validate transaction or data. In this sense, each dApp which does this must re-invent and build the wheel that suits their own tech stack each time. By providing the building blocks of security, Eigenlayer opens the door for more infrastructure-level innovation by reducing the work needed to generate your own security protocol. An example of this is an oracle like Chainlink which provides off-chain data to Ethereum. They introduced a staking protocol based on its own token, in order to validate the data they are gathering off-chain. Think about the resources this requires! Not every dApp has hundreds of millions of dollars in funding to do this. By greatly decreasing the difficulty of accessing security, Eigenlayer could increase infrastructure innovation and competition for the ecosystem as a whole.

By allowing the re-staking of your ETH, i.e. the ETH that you’ve staked on Ethereum which is returning you 4% APY, can also be used to stake additional services (which would leverage Eigenlayer’s middleware as mentioned above) for which you can earn additional yield.

This model can be extended to both chains and dApps. More on this to come.

How can validator sets be attacked? If you follow a model similar to Ethereum, then a malicious attacker (or group) must first own more ETH than the rest of the honest validators and use that to outvote the block attestations made by those validators. Given that blocks are executed only after being attested to by validators, this could be a potential attack vector. As of Jan 2023, that’s nearly 22.4 billion. The problem here for the malicious staker/validator(s) is that they are likely to lose all of the ETH that they used to outvote the honest validators because there are systems of recourse such as slashing or forking within Ethereum. So unless the attack would earn you more ETH than what you are putting down, doesn’t make sense for you to do it.

The problem this presents for smaller applications is that they would likely have much smaller capital pools for staking, which means it would be economically much easier for an attacker to succeed and be profitable.

The promise of shared security is that by using the Ethereum staking pool to secure other applications which have much lower total locked value, the cost and feasibility of attacking the protocol go way up because you would have to attack the entire capital pool of Ethereum re-stakers.

Scaling Ethereum/Web3 so far has meant separating data availability (DA) and execution layers as a means to reduce the amount of computing that any system has to handle. Modularizing chains = separating data availability and execution. We have seen the rise of L2 execution chains on Ethereum (Optimistic rollups, ZKEVMs) using Ethereum as the DA layer, as well as new chains that look to either be data availability (Celestia, EigenDA), execution layers, or both (Fuel).

Eigenlayer is contributing to infrastructure modularity in the same vein. Security models are a core part of building blockchain infrastructure, whether that’s DA or execution layers. Modularity = Introducing security as a component you can effectively break out of your core tech stack, and replacing it with an Eigenlayer solution.

Ethereum is generally recognized as having the best-in-class security model for blockchains. Whenever you decouple from this, security challenges arise. Thus, whether for execution or data availability, protocols must build their own new security models one way or the other.

By providing Security as a Service to stakers and dApps, Eigenlayer presents possibilities for modularizing more components of how we conduct transactions or collect data on-chain. Some possibilities they have outlined in their whitepaper:

• Hyperscale DA layer

• Decentralized sequencers

• Light-node bridges

• Fast-mode bridges for Rollups

• Oracles

• Opt-in Event-Drive Activation

• Opt-in MEV Management

• Settlement Chains with Ultra-Low Latency

• Single Slot Finality

I will note many of the above applications exist in one way or another, many of them at very high valuations. Personally, I am very interested in seeing how Eigenlayer could potentially level the playing field for further infrastructure innovation.

The barriers to creating security infrastructure mean that economies of scale have accrued to dApps which have already created their own infrastructure, oftentimes funded by unicorn valuations at the outset.

Protocols and applications such as L1 or L2 chains, bridges, or oracles, can only receive a comparable amount of funding as their previous competitors if they are able to create compelling technology innovation to receive the funds that are needed at the outset to build. This naturally means that user-facing developments such as UX/UI are often not the core focuses of these organizations. If Web3 is to truly reach a broader audience, teams that have a more commercial or user-facing focus or simply provide value propositions unrelated to core infrastructure innovations should also be able to compete on those grounds alone, without also having to be cutting-edge researchers. The Coinbase launch of Base chain is a great example. A dApp that is unencumbered by the developmental overhead of creating their own chain are able to focus on their developer-facing and user-facing strategy and offerings. Jesse Pollock expounded on the Coinbase vision here: To make crypto application-centric rather than chain-centric for users and developers and optimizing for a better user experience, “it’s still too hard to use crypto” [for mainstream audiences].

Eigenlayer is launching EigenDA as a proof of concept and also as a building layer. Clarification: the shared security model of Eigenlayer is not constrained to just dApps.

Kudos to: Jacob, Brandon, and Yuan for sanity-checking understanding. Shoutout to 0xJim for providing feedback on structure and flushing out modularity/UX claims.