The Restaking Ecosystem and What Comes Next

Introduction

When people start their journey in crypto, they tend to buy Bitcoin or Ethereum (the largest cryptocurrencies by market cap at $1.4T and $247B capitalization respectively) to gain some exposure. As they dive deeper into crypto ecosystems and get exposed to DeFi it quickly becomes apparent that there is much more to do than just hold crypto assets – it’s also possible to earn yield – the lifeblood of the crypto economy.

Those who farm it early stand to make solid gains… it’s intoxicatingly addictive in high doses… and it powers interchain travel (of liquidity)… it’s basically the spice melange of crypto.

“Moooar yield!”

Although the current options to earn yield are already very diverse, there is an ever-expanding number of creative new ways being developed for users & investors to earn additional yield.

In this article, we first break down the current ways ETH holders can earn yield (skip this part if you’re not a beginner) and then dive into the next evolution of crypto yields & security - restaking. Specifically, we cover what it is, the risks we should be aware of, current ecosystem participants and we finish with what we’d like to see next in the restaking ecosystem to make it reach its full potential.

1. Current Staking Options

To stake native ETH and earn yield, one must hold 32ETH and run their own node. At current prices this requires a holder to have ~$112,000 and as such is relatively cost-prohibitive. Additionally, more accessible yield options emerged early on, enabling anyone to stake ETH in any amount with:

1. Centralised exchanges (Binance, Coinbase)
2. Staking pools (Staked.us, Figment); or
3. Liquid staking protocols (Lido, Rocketpool, Frax)

With a centralised exchange, you stake your ETH directly with a Binance or Coinbase, and they take a cut of rewards. This makes staking extremely easy for beginners to start earning yield, but users assets are exposed to centralization and exchange risk – if the exchanges goes down then so does your Eth. If they choose to withhold your assets there’s not much you can do.

With a staking pool, you stake your ETH with a node operator that handles the technical backend and takes a fee in return. This is geared towards institutional clients, and while largely still centralised, it does come with a host of security features.

With a liquid staking protocol, you stake your ETH into a Lido or a Rocketpool smart contract. They then aggregate the ETH and spin up a validator, earn yield on it, and pass it along to users for a small fee. Unlike the previous options, when the ETH is deposited into their contracts, the user receives a liquid staking token (LST) in return. This token represents a claim on their ETH and its staking yield.

Current staking options for crypto assets like ETH

This liquid staking token can then be used in other DeFi activities, such as using it for collateral to borrow in protocols like Aave, etc. This allows users to increase their capital efficiency and earn yields and returns from various sources. When it was first launched, liquid staking was a key innovation and has since become a core DeFi building block (or ‘primitive’). As of today it accounts for 30% of Eth staked and is the most popular form of staking, edging out Eth staked on centralized exchanges Coinbase and Binance.

Source: https://dune.com/hildobby/eth2-staking

In the years after LSTs emerged, there has been a natural progression to more complex ways of earning yield and utilizing staking concepts.

This is where the concept of re-staking comes in.

2. Restaking & EigenLayer

Restaking allows users the ability to take their staked ETH or LSTs and stake them again in new pools that offer additional yields. These pools secure other protocols, projects and networks (like rollups, data availability layers and oracles) and be offered additional yield by them. This is an evolution of current staking.

With restaking, ETH holders can now earn more yield by supporting a myriad of emerging projects, products & protocols. At the heart of restaking is EigenLayer – the infrastructure protocol driving this novel innovation and allowing Ethereum’s Proof of Stake network to be flexibly used by other projects & protocols looking to launch quickly and securely.

Whenever new protocols launch, they have to establish their own networks for validation, often secured with their own tokens which haven’t had the time to accrue large amounts of value yet and could be vulnerable to 51% attacks. These newer protocols face challenges in achieving security and decentralization due to fewer resources and smaller communities.

EigenLayer helps solve this by taking the large amount of value that secures the Ethereum network and applying some of it to these new networks, for a fee (in the form of the new protocol paying additional yield).

What do the ETH security numbers look like? Conceptually, restaking utilizes a portion of the security from the underlying ETH, and applies it to another protocol, thereby helping it bootstrap and “inherit” security from Ethereum – a $427b network with 930,000 validators globally and 25% staked – an incredibly decentralized and economically secure system. According to this February 2024 paper by Nuzzi, Waters & Andrade attacking Ethereum would cost billions and be relatively insurmountable, making it ‘secure’.

There are four key participants in the restaking ecosystem:

• Restakers / Users – Those who wish to stake their ETH or LST to earn additional yield
• Liquid Restaking Providers – User interface that abstracts the complexity of managing a node and selecting different protocols to secure. Users deposit ETH in and let Liquid Restaking Providers manage who and what to allocate to
• Operators – Validators that secure tasks on new protocols
• Actively Validated Services – The protocols that are being secured through restaking

The flow looks a little like this:

3. The Restaking Ecosystem Today

EigenLayer is relatively new, so where are we today in the setup of restaking?

Active Validated Services (AVSs): AVSs are any system that would like to utilize re-staked ETH to bootstrap its network – it could be a rollup, a data availability layer, an oracle, a co-processor or even a simple encrypted mempool. Relying on re-staked ETH to achieve network validation and security bypasses them having to raise their own token to do so.

As of writing, in Holesky Testnet of EigenLayer, permissionless AVS’s are not live yet. At this point, only the in-house and first AVS - EigenDA (a data availability layer) is live on testnet as Stage 2 mainnet approaches sometime in Q22024. Stage 3 will introduce AVSs beyond EigenDA, which will go to testnet and mainnet in the second half of 2024.

Currently, there are about 10 AVSs scheduled for the coming months before it goes permissionless. These are split across Rollup Infra (AltLayer, Lagrange), Sequencers (Espresso), other chains (Ethos for Cosmos, Near), L2/Rollup Aggregators (Omni, Hyperlane) and others (Silence for privacy, Aethos for compliance). The full list can be found here.

Operators: Operators secure AVS “tasks” (a unit of work to be done for an AVS, this could be validating a transaction and finalising a block, providing data availability guarantees, guaranteeing a co-processor output, or verifying rollup state etc) on EigenLayer by restaking their ETH on the new protocol. They register with EigenLayer, enable ETH stakers to delegate their staked assets, and then opt-in to provide a range of services to AVSs to enhance the overall security and functionality of their networks. The full list is here, with the most active ones being Figment, P2P, Chorus One, and Kiln.

Liquid Restaking Providers: Of course, it wouldn’t be crypto if there wasn’t another innovation to free up liquidity – enter “Liquid Restaking Tokens” or LRTs. These are a riff on liquid staking tokens (LSTs). In the same way users stake ETH into LST protocols like Lido and use the derivative they receive (stETH) to engage in further DeFi activities, LRT users can now stake their ETH or LSTs into a liquid restaking protocol and use the derivative token they receive to engage in further DeFi activities. LRTs abstract all complexity from the end-user and act as this interface layer between the user and EigenLayer Operators. Their sole goal is to maximize the return for the user while minimizing risk exposure.

Full ecosystem: https://www.eigenlayer.xyz/ecosystem?category=AVS%2CRollup%2COperator

4. Why is the emergence of Restaking groundbreaking and what are the key concerns?

Let’s start with the main positives -

• Inheriting Ethereum’s security - It’s taken years to get to Ethereum’s critical mass. Just putting this into numbers, it would have cost an estimated $34.39 billion to 34% attack the Ethereum network at the start of 2024 — when ETH was priced at ~$2,300 — and it would take until June 14, 2024, for the attacker to successfully gain the required control over the network. Inheriting even a fraction of this decentralised and thorough security at a fraction of the cost and effort is a no-brainer. It’s making ETH the base asset for programmable trust.
• Using existing infrastructure will spur innovation - A big concern and cost for any middleware or infrastructure solution is to ensure a decentralised validator set. It’s certainly a big undertaking. With EigenLayer, that can be abstracted away and focusing on building the best tech can be the prime goal.
• LRTs make this significantly easier for retail adoption - LRTs provide a seamless integration between restaking and DeFi, allowing users to deposit Eth or LSTs, and not have to worry about which operators to partner with and which AVSs make sense to support. This is evident with the billions of TVL that they have managed to bring onboard.

Here are some of the things we need to watch out for -

• The big leverage effect or ‘Rehypothecation’ - Operators / Restakers will stake on multiple AVSs and there’s nothing stopping them from doing so. In fact, it is in their financial interest too. We must understand the ripple effects this can cause. In short, if they get slashed in one place that ETH is lost and the security decreases everywhere else, and this can have a domino effect and potentially be problematic. A workaround here is on attributable security, but this means each dollar is staked and protecting only one AVS, which comes with better security but lower yield.
• Assuming the subset is as honest as the majority - Ethereum as a whole is decentralised and secure, but is the very small subset of these operators who are protecting your AVS sharing the same characteristics? It’s hard to guarantee this, and ultimately lies on the AVS to ensure. While EigenLayer conveys the importance of decentralisation within each AVS, the security of any project leveraging a fraction of the validators could be in trouble.
• Risks of using native ETH vs LSTs to restake with an LRT – If you are staking an LST with an LRT (even more options now), which is then staking it with AVSs, you have 3 layers of smart contract risk per AVS you are restaking with. And after all, EigenLayer is also just a system of smart contracts on Ethereum.
• LRTs are basically ‘Risk Managers’ given they decide which AVS’ to secure with their ETH, but do they have the requisite skills to manage risk? While it’s great that LRTs abstract away everything from a retail user and allow them to get exposure to restaking, they must trust not only the LRT smart contracts, but their ability to work with the right operators and choose the best AVSs and manage risk accordingly. There is very little today to prove the risk management capabilities of any of the LRTs, and hopefully no single LRT goes only after the highest rewards without taking risk into consideration.

Another recent development has been partnerships between LRT Providers and AVSs to guarantee specific levels of security. For example, EtherFi committed $500m of Ethereum security each to Lagrange and Aethos and $600m for Omni Network. These kinds of deals are expected with the space heating up, but it is important to understand that security guarantees in absolute figures are harder to guarantee, given TVL depends fully on the USD* price of ETH, and that there is no lock-up for users who can unstake when they want, putting even more unreliability on that absolute figure.

EigenLayer and the Restaking Ecosystem are growing tremendously, despite technically being in “Testnet.” We’re expecting a lot of missing pieces and streamlining products to be added, along with a lot more innovation that can be unlocked with EigenLayer.

Here we’d like to mention a (definitely non-exhaustive) list of a few innovations and missing pieces that we think will be very interesting as EigenLayer hits mainnet and becomes permissionless.

5. Innovations and some of the missing pieces in Restaking we are excited about…

Restaking is in its infancy and there are several undeveloped aspects that will help it achieve its full potential by streamlining the various stakeholders and minimizing user and liquidity friction.

Here are some of the key innovations on the horizon:

MEV

Aside from inheriting Ethereum’s trust on both the decentralised security and economic security aspect, you also inherit the Ethereum inclusion trust, which means that because these operators (Validators) are Ethereum validators (aka proposers), and there are quite a few possibilities for innovation (and some required risk control) there.

This is being able to make changes to block inclusion, block ordering and block structuring, thereby ensuring sound MEV management entirely secured by EigenLayer. One example could be proposers adding new transactions in the block (because of 1559) potentially improving censorship resistance or having block proposers committing to passing arbitrage or liquidation gains back to let’s say a Uniswap pool. Also, for rollups, managing MEV includes having decentralised sequencers or even threshold encryption which could all be AVS’s! Below is an example from MEV Boost+.

Source: EigenLayer

It is important to note that some risks do come up here, especially as EigenLayer grows, such as EigenLayer operators potentially having exclusive order flow, or extract more cross domain MEV since the operators here are also builders and validators here. One good example is if an EigenLayer node responsible for updating the oracle is also the current proposer on Ethereum, they can build the block locally and earn the MEV associated with the oracle update.

Decentralised AI

Even though decentralised AI is still largely in build mode, the potential opportunities arising with the use of EigenLayer are intriguing. Not a surprise given Eigenlayer’s founder has a deep background in AI research. Recently, Ritual announced its partnership with EigenLayer, where EigenLayer will provide a higher levels of decentralization and security for Ritual’s AI operations than it might otherwise have achieved if it launched its own network. But what else can AI and EigenLayer enable?

Let’s start with an example from the recent Unchained podcast with EigenLayer’s founder. A simple Web3 wallet, which we all know is in dire need of an upgrade. Now imagine running an AI model that can allow you to execute transactions or bridge assets for you based on natural language. It’s just like calling your broker and asking him to buy shares of a company, and he says it’s done. The end user does not have to bother with what exchange was used or where it is being custodied. All this, secured by operators who sign off on the intent and transaction. A Natural Language Processing-driven transaction secured by Ethereum.

Another one could be using that security to guarantee an AI inference response using a co-processor. That inference protocol runs off-chain, and then the answer comes back, but how can you trust that? One way is using ZKMLs, which are great but still largely in development. A simpler one could just be a pool of security backing this answer. This crypto-economic security adds trust to an AI inference protocol, and in case the claim is not right, there is a pool of capital that can be distributed accordingly.

This is just the start, and we can expect to see many other innovations emerge. In fact, as we move ahead, we anticipate that many AI Projects that did not have EigenLayer as a part of their roadmap will realise that this shared trust model is a great way to start incorporating crypto economic security into their protocols.

Zk Proof Verification

Ethereum is a great general-purpose VM. You can, however, build ZK verifiers on top of this, but there is a cost for each operation, otherwise you could potentially spam the system. Cost of each zk verification depends, SNARKs generally being cheaper than STARKs and Mina’s IPA commitment scheme.

Given that this cost is quite high, having operators on EigenLayer participate in ZK proof verification offchain, and certify that proofs are correct onchain can potentially be more efficient and is something that has been highlighted in EigenLayer’s whitepaper as well.

The folks at Aligned Layer (who will be an AVS) are tackling this challenge, and they are building a layer for verification and aggregation on top of Eigenlayer that can be used with any proving system. In fact, this can also work very well with optimistic proofs, who are actively exploring fraud proofs to reduce the proof window ending up with better finality.

Optimising Payments between AVS’s and Operators

As AVSs become more permissionless and a plethora of them pop up, it’s important for them to understand how much they are paying operators to secure their network. Are they paying too little, too much or just the right amount? This requires some serious risk and security modeling, understanding the cost it would incur to corrupt a protocol, how much an attacker could profit, and making sure the former is always higher than the latter. Anzen from Hydrogen Labs is trying to solve this issue by building an economically secured oracle which brings required data on chain, and their Reserves Manager Smart Contract which dynamically adjusts fees paid to Operators to ensure they are within the safety factor.

Easy Integration between Operators and AVS’s

Node Operators will want to quickly integrate with the best AVS’s to ensure optimised returns. But integrating every AVS will have different CLIs to manage and different integration paths. What if there could be a single interface that would allow easy integration between the operators and AVS’s? While still early, that’s what the folks at Nethermind are whipping up, and we can clearly see this creating a further frictionless experience.

EigenCerts

Still in its development state, and linked to the attributable security model mentioned above, EigenCerts is built for AVS’s and allows signatures to be aggregated and posted to mainnet, detailing how each dollar restaked on an AVS is attributable to one operator. Once this comes live, it will be interesting to see how restakers would want to see this cert and ensure their stake isnt going across to too many AVS’s, even though that bears a lower yield. We can see some early stages of the code here.

Conclusion

EigenLayer heralds a transformative future for Web3, offering a scalable, secure foundation for Ethereum innovation and beyond. While some risks exist and high yields make taking that risk even more lucrative, we hope the key stakeholders here take risk management seriously and take a proactive approach towards mitigating these concerns. EigenLayer has the potential to disrupt the Web3 ecosystem and drive innovation at an unprecedented speed, and we are looking forward to what is yet to come.

Superscrypt is not invested in Eigenlayer or the ecosystem participants listed in this article.

statement:

1. This article is reproduced from [Superscrypt], the copyright belongs to the original author [Andy Chen, Pranav P, Jacob], if you have any objections to the reprint, please contact the Gate Learn team, and the team will handle it as soon as possible according to relevant procedures.

2. Disclaimer: The views and opinions expressed in this article represent only the author’s personal views and do not constitute any investment advice.

3. Other language versions of the article are translated by the Gate Learn team and are not mentioned in Gate.io, the translated article may not be reproduced, distributed or plagiarized.