Reflections on Restaking

“If only you could see yourself in my eyes” — Lost by Dermot Kennedy

We spend a lot of time studying restaking protocols at Reverie. It’s an exciting category for us to invest in because of how fuzzy everything is (opportunity lies in murky markets) and just how much there is going on (dozens of projects are launching in the restaking space over the next 12 months). 

As part of our work on restaking, we’ve stumbled onto some observations on how we see things playing out for restaking marketplaces over the next few years.  

Much of this stuff is emergent, so what’s true today may not be true tomorrow. Nonetheless, we wanted to share some of our preliminary observations on the commercial dynamics for restaking marketplaces.

LRT’s as a point of leverage 

Today, LRT’s like Etherfi/Renzo hold a powerful position in the restaking supply-chain: since they are close to both the supply-side (stakers) and the demand-size (AVS’s), they are in the privileged position of being on both sides of the transaction. If you play this out, this gives LRT’s the ability to (i) dictate their rake, and (ii) influence the rake of the underlying marketplace (e.g., EigenLayer, Symbiotic). Given their powerful position, you would expect to see restaking marketplaces launch first-party LRT’s to rein-in the power of third-party LRT’s.*

AVS’s/restakers as a point of leverage

The world’s best marketplaces share two features: a fragmented supply-side and fragmented demand-side. To form intuition around this point, it helps to look at the opposite situation where either/both sides of the marketplace are concentrated. 

Imagine a simple marketplace that deals in apples, where the largest seller of apples controls 50%+ of the apple supply. In this scenario, if the marketplace operator decided to increase her marketplace rake from 5 to 10%, the large apple seller could threaten to take his business elsewhere. 

Similarly on the demand-side, if the largest buyer of apples controls 50%+ of the apple demand, she can threaten to use another marketplace (or potentially go direct to the apple supplier) if the marketplace operator increases the marketplace rake. 

Turning back to restaking marketplaces, if the end-state market structure for restaking marketplaces is concentrated on either the AVS side (the top 10% of AVS’s make up 50%+ of the revenue) or restaker side (the top 10% of restakers make up 50%+ of the deposits), then the natural consequence is the marketplace has a reduced ability to extract a rake for itself (and should thus command a lower valuation).

While there’s not enough data to be rigorous here, our intuition is the power law will apply here: large AVS’s will be the majority of gross payment volume, giving them bargaining power over the rake the marketplace will eventually want to charge. 

Fight for exclusive AVS’s

From the point of view of each restaking marketplace, every opportunity to do something a competing restaking marketplace cannot do is worth taking. The easiest way to differentiate as a restaking marketplace is to offer restakers access to exclusive AVS’s — either first-party ones like EigenDA or third-party ones through exclusive partnerships. Conceptually, this is similar to how Sony develops games exclusive to the PS5 to drive hardware sales. 

Due to these dynamics, we expect to see restaking marketplaces do things like launch more first-party AVS’s and/or enter into exclusivity arrangements with third-party AVS’s. In short, expect to see a land-grab for AVS’s over the coming months.

AVS subsidies

AVS’s need to pay operators/restakers for the service being provided, which, practically speaking, means AVS’s need to be ready to pay in their native token, ETH/USDC, or potentially points/future airdrop. That said, since most AVS’s to date have been early-stage startups without a token, a large balance sheet, or a well-designed points program/airdrop, signing up operators/restakers has proven to be a cumbersome process (most of the EigenLayer partnerships have been bespoke contractual agreements that are negotiated in private). Simply put, this is a situation where the customer wants to buy a service, is probably good for the money, but doesn’t have the money yet. 

To grease the wheels of business here, it’s very likely restaking marketplace will “front” the initiate payments to operators/restakers, either through their native token, balance sheet assets, or potentially by issuing “cloud credits” for AVS’s to spend with operators/restakers. In return for fronting the money, you would expect AVS’s to promise an airdrop/token allocation to the restaking marketplace. Alternatively, the restaking marketplace can front this money to the AVS to convince it to go with you instead of a competing restaking marketplace. 

In short, we expect to see restaking marketplaces to play hardball with each other by subsidizing AVS spend over the next 12-24 months. And similar to the market dynamics seen with Uber/Lyft, the restaking marketplace with the most dollars/tokens to spend may end up as the winning marketplace. 

White-glove onboarding

Going from “I want to spin up an AVS” to actually going into production is more difficult than it looks, particularly for small teams that don’t have much bandwidth for R&D. Examples of things teams need to figure out include how much security should I buy, how long should I buy it for, how much should I pay operators/restakers, what should I slash for, and how much should I slash? 

Best-practices will emerge eventually, but for the time being, restaking marketplaces will need to hand-hold AVS teams through these questions (it’s worth noting that EigenLayer doesn’t have payments or slashing yet). 

To that end, we would expect to see the winning restaking marketplaces look a little like enterprise sales businesses, which provide white-glove integrations/services help to their customers to onboard them onto their product. 

Graduating from the marketplace

One fascinating dynamic that might emerge is one where the most successful AVS’s end up “graduating” from the restaking marketplace, and instead, use their own token/revenue to purchase security. 

Today, the pitch for restaking is most relevant to smaller projects that (i) don’t have the time/money/brand/connections to recruit a validator set, and (ii) don’t have a highly-valued token which secures the network. But as projects get larger, there is a world where the natural next step for them is to leave the restaking marketplace in favor of recruiting their own validator set and securing it with their own — now more highly valued — token. 

Conceptually, this would be similar to dating marketplace dynamics (e.g., Hinge, Tinder), where the most successful customers end up churning out of the marketplace. For the marketplace operator, however, this churn is bad news since you’re losing a customer (this is one reason for why dating marketplaces trade at lower valuations/multiples than marketplaces with repeat usage/low churn).

One-stop-shop for crypto SaaS

To set the table for this observation, let’s dive into a bit of software history: cloud providers like AWS made it easy for developers to access everything they need to spin up an app or web service (e.g., hosting, storage, and compute). By significantly reducing the cost and time required to develop software, a new category of web services emerged that was far more specialized in its service offering. The combination of first-party cloud services and a large range of “microservices” that were offered within the platform made the cloud providers a one-stop shop for everything you need besides your core business logic.

Restaking marketplaces like EigenLayer intend to create a similar crop of microservices for Web3.**  For example, before EigenLayer, crypto microservices could either keep their offchain components completely centralized (and pass on this risk to their customers), or incur the cost of bootstrapping a set of operators and economic stake to purchase security. 

Restaking marketplaces potentially break this trade-off for microservices — if it works as intended, you’ll be able to prioritize security without compromising on cost and speed to market.

Let’s say you’re developing a cheap, high performance zk-rollup. If you go to a restaking marketplace like EigenLayer, you will have multiple options for core services such as DA and bridging to enable easy onboarding. And through this process, you will see dozens of other AVS microservices that you will be able to integrate with.*** 

The more microservices a restaking marketplace offers, the better the experience for customers — instead of evaluating the features and security of services across dozens of independent vendors, apps will be able to buy all the services they need from one restaking marketplace. Come for service X, stay for service Y and Z.        

Certain AVS’s will have network effects (e.g., preconfs) 

To date, restaking use-cases have largely focused on exporting Ethereum’s validators and economic stake. But there’s another category of “inward” focused restaking use-cases that could actually add features to Ethereum’s consensus without needing to change the protocol. 

The idea is pretty simple — you allow validators to opt into making additional commitments around the blocks they propose in exchange for payment, and keep them accountable via slashing if they don’t uphold these commitments. We suspect that there are only a handful of commitment types with enough demand to garner high levels of participation****, but the amount of value flowing through these commitments has the potential to be massive.

Unlike “external” restaking use cases, the effectiveness of this category of use-cases is directly tied to validator participation. That is, even if you’re willing to pay to be included in a block, it’s not very useful if only 1 out of every 10 validators has opted into upholding that commitment. 

But if every validator has opted into a given commitment, the guarantee behind it would be equal to the guarantees provided by the Ethereum protocol itself (i.e., valid blocks). By this logic, we can expect this category to have powerful network effects, as the users of the AVS benefit from each marginal validator that opts into the commitment marketplace.

While this AVS category is still emergent, the logical distribution channel to facilitate these use-cases would be through Ethereum client sidecars and plug-ins (e.g., Reth). And similar to proposer-builder separation, it seems likely that proposers will outsource this work to specialized actors in exchange for revenue share. 

What’s less clear is the shape that these AVS’s will take. Though it’s possible that one entity could create a generalizable marketplace for any kind of commitment, we suspect it’s more likely that we see a few players emerge that specialize based on the source of demand (e.g., L2’s for interop vs L1 DeFi driven demand).     

In conclusion

To students of business strategy, the commercial dynamics of restaking marketplaces are a treasure-trove of content to sink your teeth into. And as you can probably tell from the above content, we’re having a lot of fun getting into the weeds of these projects. 

If you’re building a restaking marketplace, AVS, or something related to it, Reverie would love to chat with you. 

* * * Footnotes * * * 

* To expand on a nuance, we think it makes more sense to build first-party infrastructure that allows third-parties tobuild LRTs within the protocol rather than build a definitive first-party LRT. 

** One analogy we’ve found helpful is to view L1’s like Ethereum as the physical cloud and protocols like EigenLayer as the virtualization of these physical data centers. Similar to building data centers all around the world, it’s nearly impossible to replicate the amount of operators and economic stake that secure Ethereum. Some have tried to do so on smaller scales (e.g., Cosmos), but similar to cloud vs on-premise, renting security from an existing network will always win on cost and speed to market. Restaking marketplaces like EigenLayer aim to enable the emergence of a similar class of microservices for crypto to accelerate the development like the cloud did for Web2.       

*** We’re already seeing the effects of this unlock, as there are already ~40 microservices in development on just EigenLayer. For example, if you’re a rollup developer, you can integrate with oracles to support your DeFi ecosystem, opt-in privacy services to shield user transactions, offchain coprocessors to provide my apps with superpowers, policy engines to ensure compliance, and security services to protect users. The result is that your rollup can offer a far more feature complete development platform than anything on the market today. Importantly, it all works as intended, you’ll be able to acquire these features at a fraction of the cost and time it would have taken to develop them in-house

****  Examples could include L2’s wanting pre-confirmation of inclusion to speed up their time to finality, a DEX auctioning rights to the first transaction in the next block, a lending protocol auctioning liquidation rights (i.e., transaction placement directly behind the next oracle update), or searcher/builders that want to purchase whole blocks from future proposers.