Key Takeaways

“Multi-dip staking” refers to earning multiple yields on a single staked principal (such as ETH)
Liquid Staking Derivatives (LSDs) are the oldest and most mature form of multi-dip staking, using IOUs to allow for “double-dipping” in DeFi of staked assets and increase capital efficiency. LSD market is primarily for ETH because of historical reasons
ETH LSD Market is likely a long-run oligopoly, since there is a centralizing force of liquidity (provider trust), and a decentralizing force of security (BFT concerns). Lido, with 32% of all staked ETH, is already at the upper limit of allowed centralization.
Post-Shanghai Upgrade, in the short run smaller LSD providers may face liquidity issues (as there is an exodus of rewards), but in the long run a full PoS Ethereum may 3–4x the LSD market and heavily benefit small to mid-sized players such as Rocket Pool and Frax. A new generation of players will also likely enter, and the endgame would likely be several players with 10–20% market share each
Eigenlayer has greatly catalyzed the “multi-dip staking” model, and creates a lend-a-validator model for middleware projects (sequencers, bridges, oracles etc.) that benefits all parties, i.e. middleware projects, DApps (increased security), validators, mom-and-pop stakers (more choice and more yield).
Eigenlayer’s Practical Risks include regarding validator “interference” (simultaneously staking on multiple projects), fractured liquidity pools for LSDs, and the risk of re-centralizing the Ethereum ecosystem.
Post-Eigenlayer, “multi-dip staking” will have entered the mainstream, and this unlocking of capital is ultimately beneficial to all parties within the Ethereum ecosystem. Similar ideas will likely also expand into other “base” blockchains, especially Cosmos (IBS proposal), catalyzing a narrative of modular blockchains.

Introduction

Since the Ethereum Merge in September 2022, Proof-of-Stake (PoS) has established itself as the mainstream consensus mechanism for all major L1 chains apart from Bitcoin. As repeatedly touted, PoS has several advantages over Bitcoin’s Proof-of-Work (PoW), such as being more environmentally friendly, having lower hardware barriers to entry, and the use of efficient economic incentives rather than brute-force computational superiority to guarantee transaction safety [1].

One thing that is often overlooked is that the act of “staking” itself is an action that has greatly transformed DeFi, with the yield incentive within staking becoming a springboard for innovation in DeFi mechanisms. Within this article, I will explore the emergence of what I call “multi-dip staking,” where a user is able to earn multiple yields on a single staked deposit. Specifically, I believe that this is a metanarrative that combines the narratives of the liquid staking derivatives (LSDs), the Shanghai upgrade, and the impact of Eigenlayer on the future of staking, and will have ramifications both in the Ethereum ecosystem and beyond.

Liquid Staking Derivatives

Liquid staking is one of the earliest and most important forms of staking-based DeFi innovations, and in many ways is a byproduct of Ethereum history. In late 2020, as part of Ethereum’s long-term transition from its initial Proof-of-Work consensus mechanism to its Proof-of-Stake consensus mechanism, the “Beacon Chain” went live, initially as an independent blockchain running PoS and segregated from mainnet transactions to ensure protocol safety [2]. Validators could deposit a minimum of 32 ETH to secure the Beacon Chain, and were promised validator rewards to do so. However, validators are unable to withdraw their staked deposits and rewards in native ETH until the transition to Proof-of-Stake transition is complete after the Shanghai upgrade (which we will discuss later) [3]. So all this capital was just sitting there, stuck like a lame duck. Surely someone needed to do something about it.

Enter Liquid Staking. The basic premise is simple: when you stake a token (such as ETH) through a liquid staking service (such as Lido), that liquid staking service will return you an IOU token, guaranteeing you that you can redeem both your principal and any accrued “interest” through staking yields. These IOUs are known as “liquid staking derivatives,” or LSD. The big advantage of LSDs was that it allowed for a far greater capital efficiency, as these IOUs, such as Lido’s stETH, could be traded on secondary markets and used in DeFi applications, in the same way that one would use the native token. Of course, being temporarily unredeemable, all these IOUs were underwritten by trust in the Ethereum community, trusting that it will eventually the PoS transition will be complete and validators will be able to unstake their deposits. Viewed from this lens, one could argue that LSDs are essentially bonds, maturing with interest after the Shanghai upgrade.

Crucially, LSDs allowed stakers to both have their cake and eat it: one could reap the rewards of staking on the Beacon Chain, as well as have (at least somewhat) the liquidity of the underlying capital, ready to be reallocated and recollateralized. This process is also known as rehypothecation in TradFi, where the same underlying asset (say 10 ETH) is used as a collateral for many different uses, as you can use the same 10 ETH to stake and reap rewards on the Beacon Chain, as well as use the stETH LSD as collateral for a DeFi loan [4]. In other words, you’re double dipping that same principal of 10 ETH.

Different Implementations of Liquid Staking. Adapted from: https://twitter.com/0xYugiAI/status/1636787968160731136

There are a great variety of ways to engage in liquid staking, with both centralized exchanges, such as Coinbase, and decentralized providers such as Lido and Rocket Pool. Furthermore, there is a diversity in the ways that different providers implement liquid staking. In particular, there are three main methods for staking rewards [5]:

Analysis of different implementations of Liquid Staking Derivatives

ETH2 Liquid Staking and the Shanghai Upgrade

LSD Market Analysis as of April 2, 2023. Source: Dune Analytics

The liquid staking derivatives market for Ethereum 2.0 is the most established and active LSD market, due to the historical reasons outlined above. The current market for Ethereum LSDs is highly concentrated, with Lido dominating, having approximately 75% market share. Nonetheless, as shown in the diagram, there has been an increase in diversification of liquid staking providers, with Coinbase and Rocket Pool having 15.6% and 5.8% respectively [6].

Fundamentally, the LSD market is shaped by two countervailing forces: a centralizing force of liquidity, and the decentralizing force of security. As is the case in the majority of DeFi protocols, liquidity is critical to a LSD project’s success. With increased liquidity comes increased adoption, as more DeFi projects trust the LSD and will be willing to use it as a collateral. This creates a self-reinforcing feedback loop, as the LSD’s increased adoption in DeFi will allow it to onboard more users, which increases the LSD’s liquidity and in turn expands its adoption in DeFi even more. This flywheel effect is arguably the most important reason why Lido has retained its dominance within the LSD market. The rapid rise of Coinbase as a liquid staking provider can also be explained through this framework, as it is able to leverage its unparalleled amounts of liquidity on its exchange to quickly kickstart this liquidity flywheel.

There is, however, also the decentralizing force of security. Most blockchains are built to be Byzantine Fault Tolerant, guaranteed to hold so long as a bad actor does not control more than 1/3 of the validators [7]. In a PoS system, the danger zone is if a single staking provider controls more than 33% of the staked ETH on the blockchain, as a single hack or bug for that provider could compromise the security of transactions on the entire system. Crucially, Lido is already close to hitting that upper threshold. As per data from Dune Analytics, around 43% of all staked ETH is current staked through liquid staking pools (such as Lido, Coinbase, Rocket Pool etc.), and of this, Lido commands 74.2% market share. This means that Lido controls almost 32% of all staked ETH, veering dangerously close to the point of centralization where there could be a systemic security risk. As such, there is a huge incentive for the Ethereum community to move away from Lido to actively use and promote alternate liquid staking providers in order to maintain sufficient decentralization and thus guarantee the underlying security.

A key transformative event within the ETH2 LSD sector is the Shanghai upgrade, scheduled for April 2023. As mentioned before, the Shanghai upgrade will allow the previously issued IOUs (i.e. liquid staking derivatives such as stETH) to finally be redeemable from the Beacon Chain, thereby potentially upending the liquid staking market. In the short run, there may be immense unstaking and selling pressure, as validators are finally able to withdraw their ETH after two and a half years of waiting. About 1 million ETH in staking rewards will be eligible to be withdrawn on the day-of the Shanghai upgrade, representing roughly 10% of daily trading volume [8]. As users swarm to withdraw staking rewards, this could create a negative spiral and pose liquidity challenges for smaller liquid staking providers.

Staking Ratios of Major L1 Chains. Source: Original Content. Data Source: Staking Rewards, accessed 2 April 2023

In the long-run however, the Shanghai upgrade could turbocharge Ethereum’s liquid staking market. Currently, Ethereum only has around 15% of its total volume staked, compared to other major PoS chains that have 60–70% staked [9]. One reason for this may be because validators are currently unable to withdraw their stakes from the Beacon Chain. Thus, after the Shanghai upgrade, it is very plausible that the staking ratio of Ethereum could reach triple or quadruple to 50–60%, roughly on parity with its other PoS peers. With the enabling of staking withdrawals, the Shanghai upgrade could also incentivize retail participants into the sector, and thus increasing the proportion of ETH that is staked through liquid staking pools (both centralized and decentralized services) such as Lido, Coinbase, and Rocket Pool.

As mentioned before, because there is a great security incentive for LSD market to decentralize, this increase in LSD demand post-Shanghai upgrade may greatly benefit midsize players such as Rocket Pool and Frax, while also potentially giving birth to a whole generation of liquid staking protocols that may implement LSD models in vastly different ways than the three commonplace methods mentioned above. In the long run, however, because of the centralizing effect of liquidity, I expect that the LSD market will achieve equilibrium as an oligopoly, with several LSD projects each commanding between 10–20% of the staked ETH.

Enter Eigenlayer

The liquid staking derivative scene presented above, however, is only the tip of the iceberg for “multi-dip staking”. One of the most important recent developments within the ecosystem is the rise of “restaking” as a concept, a narrative popularized by the debut of Eigenlayer’s whitepaper in February 2023. In its whitepaper, Eigenlayer describes itself as a “restaking collective” that allows validators to “opt in to validating new software modules built on top of the Ethereum ecosystem” [10]. But what does this mean, specifically? What problem does it actually solve within the Ethereum ecosystem?

Essentially, all DApp smart contracts running on the Ethereum blockchain are secured by Ethereum validators, which provides incredibly strong security guarantees. But as DApps have become more and more complex, they’ve begun to require many other moving pieces, including bridges, oracles, and other forms of middleware, which can’t live on the Ethereum blockchain, and so are not secured by Ethereum validators. Many of these middleware projects employ their own blockchains (such as Chainlink) run their own blockchain with their own validators, or on centralized servers — in short, methods that do not have the security rigor guaranteed by the Ethereum validators [11].

Indeed, these oracles, bridges, and other pieces of middleware often have a “validator bootstrap problem,” where setting up validators for this project is expensive and time-consuming, not necessarily trustworthy, and prone to attack [12]. Eigenlayer aims to solve this issue through allowing projects to define a set of validation rules and use these to “borrow” Ethereum’s validators, as an alternative to bootstrapping their own and thus provide these projects with a stronger security guarantee.

The Different Parts of Eigenlayer’s Ecosystem Explained. Source: Original Content

As shown above, Eigenlayer’s ecosystem has several moving parts and stakeholders, and thus may be difficult to understand at first glance. So let’s break this down into each of the different stakeholders (pun intended) in the system, and explain why Eigenlayer’s design benefits them.

Middleware Projects

Middleware projects are arguably the most important part of Eigenlayer’s ecosystem. These are usually infrastructure projects that are 1–2 levels removed from the end-user, and include rollup sequencers, bridges, oracles, and RPC nodes [11]. Instead of bootstrapping the validators themselves (which we have discussed before is problematic), middleware projects simply needs to define two things: the task and the bounty. For the task, the middleware projects needs to define the validation and slashing rules, and for the bounty the project needs to define how validators will be rewarded. Intuitively, think of this as a job posting: you define what you need to do in that position (eg. flip burgers for an hour), and your compensations (eg. $15 minimum wage salary). And with other people flipping your burgers for you, you can move on to the more productive things in life.

Validators

Validators are the ones flipping the burgers for these middleware projects. But they’re picky too; after all, it’s a free market out here. With all these multiple middleware “job-postings,” validators are free to pick and choose any combination of middleware projects and rewards to validate, buffet style. As different validators may have different specialities and risk preferences, this will result in different validators securing different middleware projects and choosing different “restaking” combinations, thus offering differing validator yields. All these yields will also be in addition to the base yield of the Ethereum base layer.

DApps

User-facing DApps, such as decentralized exchanges, lending platforms etc. will still be built on the Ethereum base layer. However, within the Eigenlayer ecosystem they will interface with the middleware projects built on top of Eigenlayer (which are secured by the re-staking validators). Compared with the status quo, using an Eigenlayer-secured middleware will allow for the top-level user-facing DApp to also be more secure.

Users

In addition to interacting with user-facing DApps (that in turn use Eigenlayer), users can also directly interact with the validators through LSDs. Specifically, each validator may choose to release their own LSD, with a different yield resulting from their different pick-and-choose project combinations. In turn, users have a diversity of yield strategies to choose from, and create a “multi-dip” staking strategy which they can tailor according to their own risk profile.

Thus, in theory Eigenlayer hopes to provide a win-win solution to all parties involved. But what are some of the implications and corollaries of such as system? Firstly, with different validators restaking to different projects, they will have different yield APYs, as established above. As each validator may release its own LSD, this may result in a myriad of LSDs floating out in the market. Now recall that we’ve established that for LSDs projects to thrive, they need a liquidity moat, as it is only through sufficient liquidity that DeFi projects will be able to trust these LSDs. If every validator releases its own little xETH, yETH, zETH LSD, this may result in a very fractured liquidity pool that handicaps LSDs usefulness in rehypothecation of DeFi.

Secondly, there is the question of what is the “interference factor” here. In other words, is there an upper limit to how many “jobs” a validator can “work”? It is unclear what the constraint is here, and how each “job” the validator works may interfere with other jobs. The real life analogy here is that I only have so much working time and energy in a day, so can’t possibly work infinite jobs. Knowing this interference factor is incredibly important, as it directly affects the distribution and overlap of validators across different projects, which in turn underwrites the diversity of yield strategies that each validator gives, and in turn affects how fractured the resulting LSD market may become.

Thirdly, there is the question of systemic risk, and if Eigenlayer poses a risk of recentralization. Specifically, Eigenlayer itself, with its access to validators stakes (as it needs to apply rewards/slashing rules), may represent a single point of failure, and if exploited a significant amount of ETH may be stolen [13]. This becomes a more and more significant issue once a significant percentage of validators on Ethereum opt in to secure Eigenlayer projects, as it snowballs out to be a recentralization of Ethereum itself, becoming antithetical to the initial purpose of the project and network. To combat these systemic risks, Eigenlayer proposes an EigenDAO and a manned committee of overwatchers, comprised of experienced Ethereum, Eigenlayer, infrastructure, and application developers, that will review and resolve cases of systemic risk [13]. But the feasibility, transparency, and effectiveness of such an institution remains to be seen.

The bottom line is that Eigenlayer is a revolutionary project that solves a real need, and promises a new way of conceiving the different roles that participants play within the Ethereum ecosystem. However, because it is such a revolutionary leap, the full implications of this project cannot possibly be theorized in the abstract and predicted beforehand. Only time will tell its vision of a win-win scenario for all stakeholders within its ecosystem is actually feasible, and if it is robust enough against adversarial and malign exploits.

The Horizon Beyond Ethereum

So far, we’ve concentrated our discussion of “multi-dip staking” on the Ethereum ecosystem. But of course, Ethereum isn’t the only PoS chain out there, and multi-dip staking is absolutely possible on other PoS chains. As we’ve hinted at before, there already exist LSDs for alternate PoS chains, such as Lido’s stMATIC, stSOL, stKSM and stDOT. But it is dubious if the LSD actually increases value for the token itself in this cases. Arguably, Ethereum LSDs success was due to its special historical circumstances (of the Beacon Chain not allowing withdrawals), as well as its robust DeFi ecosystem that unleashes the full potential of capital reallocation that LSDs allow for. This is not necessarily true for all PoS chains, and as such, releasing an LSD for such a chain may not actually increase the underlying token’s value [14].

A more interesting question is if a restaking idea such as Eigenlayer may hold up within other ecosystems. I believe that it can, given certain network traits. First, the underlying network’s validators are sufficiently decentralized. For a lot of PoS chains, their validators are highly centralized and controlled by a single entity. This means that having a restaking mechanism such as Eigenlayer will not increase decentralization and security of infrastructure protocols, and such becomes meaningless to implement. Secondly, there needs to be sufficient infrastructure demand on this network. In other words, an ecosystem needs to have enough bridges, oracles, sequencers, and other middleware projects such that it warrants the necessity of having an Eigenlayer-style restaking mechanism.

Given these two prerequisites, I believe that the ecosystem that best satisfies these is the Cosmos Network. Firstly, the Cosmos Hub has sufficiently decentralized validators, and thereby can provide a relatively strong security guarantees for other projects (or in Cosmos’ case, other chains built using the Cosmos SDK). Secondly, there is sufficient demand for such a system on Cosmos, as many chains do indeed face a validator bootstrap problem, and there are many infrastructure projects built on top of the Cosmos SDK.

Indeed, Cosmos v2.0’s whitepaper vision of an Inter-Blockchain Security framework is very similar to Eigenlayer’s vision for Ethereum, as the IBS framework essentially allows for Cosmos chains to “borrow” validators from the Cosmos Hub to solve the validator bootstrap problem, and end users can stake ATOM tokens to earn yields on multiple Cosmos ecosystem chains. And even though Cosmos v2.0’s whitepaper was ultimately vetoed by its community, the IBS framework has recently been greenlit in a landslide vote [15], and so we can expect this idea of “multi-dip staking” to also gradually emerge within the Cosmos ecosystem.

Overall, throughout this essay we’ve explored in-depth the idea of “multi-dip staking,” tracing through Liquid Staking Derivatives on the Ethereum network to the Shanghai upgrade, to Eigenlayer’s significance in reforming liquid staking, and finally to a discussion on opportunities for “multi-dip staking” beyond Ethereum. Through this, I hope that I’ve shown how Proof-of-Stake itself can unlock a myriad of new DeFi opportunities, complete with a new set of incentives, risks, and opportunities. It looks like there is indeed a lot at stake, after all.

🐦 @0xfishylosopher

📅 8 April 2023