Disclaimer: This article is for informational purposes only and is not legal, business, investment. Do not base investment decisions on it, nor consider it as accounting, legal, or tax guidance. Mention of specific assets or securities are examples and not endorsements. The author’s views may not reflect those of any affiliations and are subject to change without update.
This article has been co-written with Redacted Labs, stemming from our extensive and thorough research on the LSD ecosystem.
Key Takeaways :
-
Not less than 100 LSD protocols have emerged in recent years, offering a simple and easy access to LSD and especially for Ethereum. Even if they all look very similar, you only need to focus on few of them to create profitable strategies
-
Prisma, Gravita, Lybra and Silo are definitely the best way to leverage your ETH-variant. On the other hand Pendle, Merkl and Stella serve as the optimal finishing touches for a refined strategy.
-
Restaking is still not mature, but you should keep and eye on the protocols build on top of Eigenlayer
-
The synergy between these protocols seems boundless. However, we still need more composability, more interactivity to reach the DEFI’s real potential
You may have encountered numerous discussions and articles about LSD across various platforms. However, this piece takes a unique angle, moving beyond mere descriptions and readily available data.
Here, the focus is on exploring yields and crafting strategies.
After a swift analysis of 105 protocols, we've conducted an in-depth examination of a select few to uncover the most lucrative strategies. These strategies are tailored for both asset managers and retail investors. To further aid your investment journey, I have WL protocols that I believe will dominate the LSD ecosystem and have provided more sophisticated insights into them
-
Summary :
-
Overview of the LSD landscape
-
Unlock the composability of ETH staking
-
Market size
-
DEFi’s features and the utilisation cost
-
On-chain robustness
-
The Silo’s underestimate model
-
-
Leverage your LSDFi
-
LSDFi leverage
-
LSDFi leverage and farming
-
-
What is next ? Restaking
-
Essential data for efficiently leverage your LSD
-
Find the parameter with the greatest impact on yield
-
LTV impact on the native yield
-
Why LTV and native yield matter when you implement and additional yield strategy
-
-
There's still a long way to go ; final thoughts
-
1. Overview of the LSD landscape
What is better than a mapview ? Take a glance bellow ;
These protocols have been divided into 4 parts :
-
LSD providers : infrastructures that provide liquid solutions for Ethereum staking. Among them, you have : CEXs such as Coinbase, Binance, OKX, Bitcoin Suisse; Enterprise-grade protocols like Kiln, RockX, Liquid Collective, Figment, Nethermind, Allnodes, Everstake, Stakefish ; Decentralized Liquid Staking like Lido, Rocket Pool, Frax, Stader Labs, Ankr, StakeWise, Swell, Bound Finance, SharedStake, Stafi, NodeDAO, Ether.Fi, Hord, Bifrost and pStake (soon Diva and LanterFi)
-
LSDFi : these protocols are focused on how to use LSD as collateral (exactly how Liquity initialised it), adding a layer of guaranteed return on an asset, triggering the first stage of the composability. Among them, you have : Gravita, Lybra, Prisma, Raft, Agility, Tai Money, Davos, Curve, Alacrity, Silo and soon Liquidty v2, Ethena, Tenet and Unstable Protocol
-
LSD booster : they bring another utility for LSDFi by : allowing the possibility to borrow against them with Silo, Aave, MakerDAO, Alchemix, Cat-In-A-Box, Radiant, Spark, Sonne, Abracadabra, Lodestar and soon Timeswap, Ion, Morpho Blue, Raman, Fluid, and Synthetix v3 ; trading / leveraging yield with Pendle (and so, Penpie, Equilibria and Stella), Timeless, Flashstake, Spectra, Numoen & Asymetrix in a kinda way, and Hourglass ; providing liquidity on DEXs with Merkl, Aura, Maverick, Gamma, Convex, Bunni, Beefy, Liquis ; aggregation of ETH-variant like Yearn Finance, Origin, unshETH, IndexCoop, Zero Liquid, LSDx, OctoLSD, Asymmetry and soon Tapio ; providing turnkey solutions like Sommelier, DEFi Saver, InstaDapp, Parallax and Opium ; offering “free” leverage with f(x) Protocol, Mori and Adscendo
-
Restaking : Eigenlayer creates a new market on top of Ethereum staking, adding new functionality and income to the basic staking yield. The native features of Eigenlayer have encourage other protocol to build another layer, on top of Eigenlayer. Among them, you have : Restake, Astrid Finance, Supermeta, Restaking Cloud, Genesis, Puffer, Stader Labs, and Swell
That’s a lot.
Concerning LSD providers, the protocols I've highlighted stand out as the most robust in the current landscape and, in my opinion, are best positioned to endure in the dynamic and challenging realm of DeFi ;
The DEFi’s integrations have been took from Defillama ; most of them are included but not at 100%. For exemple, Opium also integrates sETH2 and this is not refer by Defillama. The amount of liquidity includes only majors DEXs ; Uniswap / Balancer / Curve
stETH is by far the most integrated ETH-variant in DEFi. They offer one of the best yield and it’s the most liquid LSD with more than 510M, not a surprise that the Lido’s growth is hard to stop
rETH has also a deep liquidity with many composability accros DEFi, but offers the worst yield
swETH and ETHx have a relevant product as well, a quite good yield and need more DEFi integration to be largely used
cbETH is the only one CEX’s LSD selected because of its number of integration. They have a high market share while having a low on-chain liquidity
sfrxETH seems to be the main competitor of Lido in term of yield + liquidity + integration score
I have been surprise by the quality of StakeWise and their sETH2. The Uniswap pool has 71M and the yield is really interesting. I look forward to see other DEFi’s integration before consider it interesting for yield strategies.
The other ETH-variant are less interesting if we look at these criteria. We have seen recently entreprise-grade staking solutions like Liquid Collective & Figment gaining market share ; how will they succeed to integrate their LSD into the DEFI ?
Lido dominance is already very high and StakeWise and Frax seems to be the one capable to compete with it, because yieldors look at the yield first, and then yield again : only the yield => we will understand later in the article why this native yield is really important. Diva aims to answer to this question of centralisation, but still at early stage for now.
Now we have selected the stronger LSD providers, giving us a first layer guarantee layer of yield, let’s see how we can leverage it.
2. Unlock the composability of ETH staking
LSD providers offer a “risk-free” yield according to some seachers, between 3-4% during the bear market. This yield will inevitably growth during the bullrun.
Now that the ETH staking are liquid and reference as a sure value, they can be use as collateral. CDP protocols give the best institutional’s grade solution : the ability to borrow a stablecoin with a yield-bearing collateral while being able to precisely estimate the cost of the operation.
I have listed 14 protocols, but I think that those are less interesting to used :
Raft has an interesting turnkey solution to leverage, but their new RSR product similar to the DSK or Maker is not clear enough in term of scalability capabilities and the flow of the yield between the RSR and the CDP fee pool. Additionally, their interest rate are quite high compare to competitors
Agility boost the yield of their aUSD with LPing’s yield. Smart, but the TVL is going to 0 because of a high yield’s redistribution toward AGI holders and an agressive native token’s liquidity mining.
Tai Money has probably the most interesting liquidation model (with the Llama) as the functioning protects the borrowers while being attractive for liquidators ; a kind of soft-liquidation system : the borrower can have a “bad debt” but should repay it asap. Despite of that and the ongoing LM, their min CR is quite high, limiting the profitability of the leverage. We should keep an eye on Tai Money, I have seen a couple of SM using it.
This is the moment to talk about Curve Llama. Interesting liquidation model which is done by band, that protects the borrowers. The bidirectionnality of the pool (similar to Timeswap model) ensures the functioning of this soft-liquidation model. Quite high fees and UI not really helpful.
Davos has a high min CR and their mechanism is not well explained. I can say the same for Alacrity.
I’m not looking to devaluate these protocols ; I sincerely hope that they will improve their model in a way that I will wan to use it.
Only 3 of the 14 protocols have been WL :
2.1 Market size
First, let’s talk about which are these 3 protocols. Gravita is a Liquity fork, Prisma a ‘half-fork’ of Liquity while Lybra create its own model.
You have a good overview, especially in terme of scalability ; Lybra & Prisma seems to have an advantage ; they have a large TVL and their stablecoin peg is maintained
2.2 DEFi’s features and the utilisation cost :
Lybra doesn’t have a recovery mode, this is important to highlight it because that impacts a lot the borrowing power. Prisma’s recovery mode is at CR = 150% while Gravita is 140%. So, if you build a strategy using Prisma and Gravita, it would be wise to set up a CR above the minimum required during the recovery mode.
No stability = less security ? Absolutely not. Lybra adopted an elegant model combining Liquidator and Keepers hand-in-hand work to ensure liquidation which is partial as long as the CR of the bad debt don’t decreases below 125%. They’re also the only one that allow to use flash loan.
Lybra is more confortable to use because your trove can’t be redeem due to others users’ activity in the protocol.
The key advantage of Gravita is the long-term approach they have, similar to Liquity.
Lybra has to my pov, some advantage over its competitors but still have a higher min CR, impacting a lot the profitability of leverage / looping strategies. They could reduce it more, but it would weaken the stability of the protocol.
On the other hand, Prisma just launched an aggressive liquidity mining which help them to become the dominant LSDFi protocol.
2.3 On-chain robustness
What I means by “robustness” is the ability of the stablecoin to be use accros DEFi
Even if you have the best feature ever to offer a stablecoin back by ETH-variant, your stablecoin will never be use because of the lack of DEFI integration or poor on-chain liquidity
The liquidity mentioned is the one of the major pool. The price impact have been calculated with the native pool, meta-aggregator and Odos. The DEFI integration are those listed by Defillama and the one I know
eUSD and mkUSD are the stronger stablecoins. Let’s see how they achieve that ;
First, eUSD 20M of liquidity is in a single pool : eUSD/3CRV. The pool is a bit unbalanced (50% eUSD) but deep enough to limit slippage. Also, Lybra has a Stability Fund Bounty Mechanism while Gravita and Prisma mainly rely on DEX arbitrage and dynamic redeem fees to maintain peg.
mkUSD recently succeed to maintain peg because of the huge liquidity mining, all the Curve’s pools are boosted. Attractive use-cases for the stablecoin = stronger peg resilience
The DEFi integrations also help to maintain peg !
Anon, will you mint a depeg stablecoin with rather no DEFI integration ? If you answer yes, this is because you bet of the fact that the stablecoin will depeg even more, sad for the protocol and your collateral …
Currently, this is a big problem for these stablecoin as they don’t offer lot of possibility to use it afterwards. $GRAI has many pools, but without deep liquidity. eUSD offers a better yield on Convex and can also be use in Pendle -- eUSD is yield-bearing at 5,50% which help to find a better yield DEFi.
The insufficient liquidity and the lack of composability weaken these stablecoin which are minted and swapped for another one, most of the time the USDC and DAI.
Gravita has interesting features. I’m pretty confident on this protocol and the as the $GRAI is not that much composable, it’s quite dangerous to swap for another stablecoin it at a depeg value, and then swap it back later when the peg will be reach again.
2.4 The Silo’s underestimate model
Silo created an innovated model allowing it to be categorized as a CDP protocol and Money Market at the same time.
Silo Finance highlights three main pillars in its protocol design :
-
Isolated risk : Silos can only borrow each other's bridge assets, keeping the risk isolated to a single Silo.
-
Shared liquidity : to concentrate liquidity, each tokenized asset gets a single Silo. With the bridge asset linking all Silos, liquidity moves fluidly throughout the protocol, allowing any collateral token to borrow another
-
Permisionless : Silo is designed to support any tokenized asset on the chains on which it operates
To make it simple ; the CDP protocol is on L1, where you can borrow XAI and the CDP protocol is on Arbitrum where you can borrow USDC instead of XAI.
Silo unlocks an unprecedented level of customization as you have the possibility to chose which asset you want to put as collateral and which one you want to borrow. And, this is really powerful for LSD strategies.
Theu actually incentive the pools with LM and soon with the 1M $ARB they have for the STIP 1.
To make it simpler to explain, I have frozen (on the date of the 31th october) the yield of the protocol to explain the different strategies you can use.
On Ethereum :
Only the cbETH’s Silo has a deep enough liquidity for large trade. The big advantage of this leverage is that you’re not exposed to liquidation due to price movement (if we rely on Chainlink).
On Arbitrum :
Limitation :
the borrowing APR are calculated with 204k / 234k borrowed. For the USDC, the borrowing APR will inevitably be lower as you may not borrow with a 2,8% max DD only
I have chosen to don’t include the yield-bearing yield of the cbETH in the result to make the computation simpler further in the article
Compare to Lybra, Prisma and Gravita, Silo have a key advantage by giving the possibility to borrow ETH vs ETH-variant, or the opposite. Only Lodestar and Abracadabra can be consider as the direct competitors, regarding these feature. I think we can also quote Cat-in-a-box here ; they had a relevant idea with their boxETH-ETHvariant pools with a an interest rate base on your LTV (as Tai Money does, but dynamically) but in my point of view, not efficient enough.
With Silo, the max CR to borrow stablecoin is 117,65% on Ethereum and 125% on Arbitrum ; quite competitive as well. So, you may consider this protocol if you want to leverage your LSD and farm the stablecoin borrowed. Consider that $XAI is currently depeg at -0,59% (31/10/23).
4 protocols stand out when constructing the second layer of yield generation. While none is superior to the others, there's likely one that aligns more closely with your strategic approach.
I am eager to see new competitors enter the scene such as Morpho Blue, Synthetix v3, Timeswap, Ethena, Liquity v2, Fluid and Ion.
In the same way that Silo does, Raman Finance brings and interesting use-case by creating a stablecoin backed by yield-bearing stablecoins (eUSD) ; similar to Blueberry and Flux Finance. But, the question remains ; unlocking efficiency or adding a superfluous layer of yield compare to the risks involved ?
Having established Layer 1 and 2, let's delve into the third layer of yield.
3. Leverage your LSDFi exposure
Now we have our LSD in collateral and we borrow stablecoin or any ETH-variant, how to leverage it more ?
3.1 LSDFi leverage
The first use cases I want to talk about are eUSD / mkUSD / GRAI / XAI. I did some simulations to compare the final yield ;
I intentionally performed the calculations prior to Prisma's liquidity mining event to ensure a more meaningful comparison
eUSD ; disadvantaged by its high min_CR ; you can improve your yield with Equilibria/Penpie
XAI ; disadvantaged by its slippage on the main pool Curve ; you can improve your yield with Yearn Finance (+1,50% on the final yield)
mkUSD ; if you open a leverage with a CR above 150% (recovery mode), with a poor native yield the strategy is hardly profitable (now compensate with the LM)
GRAI ; disadvantaged by its slippage and low liquidity + currently depeg, present a risk when you will swap it back
Globally, you notice that this is not really profitable to use stablecoin of the yield bearing stablecoin. The main reasons are the lack of DEFi integration and the high slippage.
eUSD is an exception. You can also easily swap it to another stablecoin and use it in a more profitable strategy (mkUSD also today)
3.2 LSDFi leverage and farming
Among the plethora of protocols mentioned in the beginning of the article, only few can be used for large size. Here they’re :
Pendle (Equilibria & Penpie as well, according to your risk tolerance) ; even without locking any $PENDLE, you can have a decent yield by LPing on ETH-variant pools. They have 124M TVL in LSD, enough to widely LP. Other competitors are quite far, but keep an eye on Timeless v2 anon.
F(x) Protocol is the most serious protocol of its kind. You can “freely” leverage your exposure to ETH by using stETH as collateral. Not suitable for large trade today.
Merkl is to my pov the best solution to high yield with large amount of $. They integrate various protocols and their UI/UX is simpler than Bunni. Basically, Merkl is an “incentive market place” where you can easily find boosted rewards on DEXs. Take a glance below to see how Merkl fits in well with the strategies outlined above :
The features of the current best protocols (borrowing stable vs ETH-variant) don’t allow to create interesting strategy. Most of the time, the low LTV and poor native yield is responsable of that (by choosing a high max DD). Hard to compete with the Sommelier high yield vault.
However, when you have the possibility to borrow ETH vs ETH-variant, the dynamics change significantly as you can use a higher LTV.
Aura pool can compete with Merlk, but I will not be surprise to see in the future the Aura’s pools directly in Merkl. Pendle also offers high yield, but you need to consider prices impacts and the exposure to the underlying assets
We had talked about the “layer-3” of yield, but some protocols go even further. Stella brings a unique opportunity to leverage your Pendle LP and even the Penpie LP ! Their one-click leverage strategies are made possible because the liquidity providing is a bit incentive. This is a really smart approach to allow cheap looping, and Stella still has a huge margin or maneuver to boostrap liquidity and provide other high yield vault. Definitely interesting to use at the end of a yield loop.
4. What is next ? Restaking
It will be a quick overview of the restaking market, the functioning and technical feature explanations are out of the scope of this article. The idea here is to see where the restaking will come to fit into this market with an overlay of yield.
Eigenlayer is the pioneer of restaking, allowing you to use your staking for another services. Basically, you multiply the utility of the staking and receive additional fees for that.
Howerver, by using Eigenlayer your assets are locked again and you don’t have to much flexibility on the validators/operators and AVS choice.
Enter to the Eigenglayer liquid locker world. 8 protocols (probably more) are working on that. They have lot of similarities but each one distinguishes itself by its flexibility they give to the users. While Astrid Finance offers un turnkey solution where you don’t have to chose operators and benefit from a better yield thanks to rebalancing / auto-compounding, Genesis seems to be more focus on the freedom to choose the operators.
(Edit : Genesis has sunset since)
Restaking is going to make evolve the LSDFi landscape in this way :
The yield will be considerably increase. If the basic staking yield increase at 5-6% during the bullrun, we will easily find 30% yield on ETH.
You have to be aware of the inherent risk of restaking. This is a tough task which can compromise the security of Ethereum and provoke centralization through Eigenlayer. Since I have begun my researches for this article, I have already seen 4 restaking protocols shutting down, and Astrid Finance has already been hack even if the hacker has decided to return funds.
The restaking landscape is changing fast ; Swell recently announce they will integrate swETH into Eigenlayer
5. Essential data for efficiently leverage your LSD
After reading all of that, you may notice that I have highlighted many parameters which affect your yield ; mint/burn fees, slippage, interest rate, LTV max, underlying yield (LSD), secondary yield (boosters)
How to know when it’s profitable to leverage instead of simply use LSD boosters protocols ?
No worry. I have done some simulations that will give you the answer.
Let’s go step by step
5.1 Find the parameter with the greatest impact on yield
On a CDP protocol, when you provide a LSD as collateral to borrow an asset, you will be charge by mint / redeem fees and interest rate. These fees will inevitable have an impact on your native yield : how each of these parameters impact the yield ? (the TLV taken here is 70%)
The graph below shows the evolution of the native yield :
The interest rate is the parameter you should pay most attention to.
5.2 LTV impact on the native yield
Well, now let’s look of what is the impact of the LTV on the native yield :
The results are totally normal : the more your borrow, the more you will pay fees.
5.3 Why LTV and native yield matter when you implement and additional yield strategy
Let’s say we want to leverage and use the borrowed yield to LPing for 14%. At what time it’s profitable to leverage instead of simply LPing ?
The goal here is to compare the final yield when we leverage VS when we simply LP.
Here is a simulation made with Desmos. You can move by yourself the parameter to see how quickly the Leverage_yield VS Simple_yield moves. Below the letter signification :
-
m = mint fee
-
r = redeem fee
-
i = interest rate
-
z = collateral yield
-
l = LTV
-
s = yield of the secondary strategy (e.g Pendle)
-
o = initial amount in $ (not useful for the simulation)
What we notice ?
The redeem and mint fees doesn’t impact a lot the final yield
The interest has a greater, but still lesser effect
The native yield is an important parameter …
…. couple with the LTV, they can make a leverage strategy really profitable
the secondary yield strategy plays a crucial role on whether or not to leverage your investment
Use case with Prisma : Should I provide in the liquidity pool or should I leverage my LSD ?
When you set up the parameter in Desmos, you clearly see that leverage offers a pretty higher yield (LM data, 03/11/2023) :
This is an evidence for this protocol as they just start liquidity mining. But, this is not alway as evident as this one
Break event point ? There are probably several, but when you have …
-
mint fees = redeem fees = 0,5%
-
interest rate =1%
-
native yield = 5%
-
LTV = 70%
-
secondary yield = 12%
… you will receive exactly the same final yield if you leverage or not.
Caveat : the simulations don’t include gas fees and any potential slippage
Leveraging can be extremely profitable and you can have a clear idea about the expected yield.
To encapsulate the key points, it's crucial to first assess if the native yield is sufficiently high; finding effective leverage strategies is challenging when the yield falls below 4.50%. Then, the LTV is second most important ; ETH/ETH-variant borrowing will clearly have a huge advantage, or stable/YB-stable. After the LTV, the interest rate can also affect your final yield, but you must have to privilege it before the mint / redeem fees.
6. There's still a long way to go ; final thoughts
This article help you to understand what are the greatest way to create leverage strategy with LSD. Each protocol, CDP or MM, has its own features and reason to have them to offer a suitable product for certains users.
Should I use a CDP or MM protocol ? CDP protocol seems more suitable for large size as you can control the cost of the leverage (require to monitor governance activities) ; interest rate are known and frozen but this is less efficient in term of borrowing capacity. On the other hand, MM protocols allow you have a greater borrowing power as this market can be design for ETH-ETH variant pools, but most of the time the interest rate is dynamic and affect your final yield afterwards
You want all of this feature in one protocol ? You can’t today …
I think we need a LSD hub to create efficient product. Ion, Tapio and unshETH paved the way of this idea but still need to go further.
Here are my thoughts :
A hub where all the ETH-variant are put together in a pool, represents by allETH which earn fees from the underlying yield + swap fees. A such model can improve the efficiency and reduce the cost of the trade between ETH-LSD.
This allETH can be used as collateral to either mint a stablecoin and take advantage of the fixed and low cost, either borrow ETH at a high LTV. To go even further, all the DEXs’ LPs could be used as collateral in a same way.
A hub for LSD, especially make for the LSD market ; unlocking numerous possibility of strategies. The same architecture can be considered for RWA-market.
I hope you enjoy this article. I have spent many hours for the researches, simulations and written. If it was useful for for you, please share it with your crypto-mates.
See you soon anon :)