The AMM Renaissance: How MEV Auctions and Dynamic Fees Prevent LVR
✨Arrakis Finance
0x929f
December 12th, 2024

TL;DR AMMs can use MEV auctions to tax arbitrageurs and collect fees to mitigate MEV and LVR. We explore how cutting-edge solutions like Unichain, Semantic Layer, and Arrakis are using similar mechanisms to prevent LVR today.

Key Takeaways:

  • Arbitrage is the largest form of MEV extraction. Arbitrage is an unnecessary cost to DeFi’s liquidity providers and a threat to the sustainability of onchain liquidity.

  • Several solutions are exploring ways to collect fees and tax arbitrageurs to ease the cost of MEV extraction.

  • While MEV awareness is a relatively new concept, it’s one of DeFi’s most promising verticals, offering a path to a sustainable onchain economy.

  • Arrakis is set to focus on MEV protection and recapture to make token issuers and LPs more profitable when they deposit liquidity onchain.

AMMs are the backbone of DeFi. But as more and more value gets leaked to the MEV supply chain, the whole ecosystem faces a problem: LPing on AMMs is unsustainable because MEV searchers extract value from LPs—making the already difficult financial endeavor of maintaining profitability from LPing nearly impossible. Arbitrage-related MEV extraction negatively impacts both token issuers and yield-seeking retail LPs.

In our previous deep dive on the AMM Renaissance, we looked at how offchain systems like Order Flow Auctions can reduce LVR for LPs. For part two of this series, we focus on how solutions adjust their fees, collect MEV taxes, and adopt other mechanisms to mitigate LVR and make LPing more sustainable. Rather than directly reducing LVR, these designs focus on minimizing its impact by charging a fee or reclaiming the value extracted by MEV searchers.

Many solutions effectively introduce an “arb tax” to protect LPs and we detail their various approaches below. As with our previous feature, our research leads us to an optimistic conclusion: Next-generation AMMs are taking novel approaches to end DeFi’s MEV crisis and that’s a positive sign for the onchain future.

Mitigating LVR

Offchain systems like oracles can help directly reduce LVR for LPs by minimizing arbitrage opportunities. The alternative approach to address the issue is to mitigate LVR to lessen its impact. Many new AMM designs take this approach.

Dynamic Fees

Dynamic fees protect LPs against arbitrage-related MEV. But to understand how, it’s worth explaining how static fees can protect users against arbitrage. Let’s say BTC is trading at $101,000. As the LP, you are quoting $100,000 via an AMM with a 1% fee. If the searcher extracts $1K in MEV, you do not suffer from arbitrage because the $1K fee you receive offsets the losses. But there is a catch: the static fee is only protective against arbs that are equal to or less than the size of the fee (plus gas). If the arb is higher than the fee, the searcher may extract a larger sum than the fee. When it’s low, the fee may unnecessarily deter order flow.

As atiselsts.eth frames it in Dynamic Fees for Automated Market Makers: Liquidity, Volatility, and Collected Fees, dynamic fee models put a price on volatility, where fees increase as prices deviate from the updated CEX price. They can also help identify toxic flow. HOT, the MEV-aware AMM built by Arrakis and Valantis Labs, implements a dynamic fee model for this reason. Dynamic fees can be classified into one of two categories: uniform and source-dependent.

Price evolutions on an AMM with constant fees vs. an AMM with dynamic fees. With HOT, Arrakis uses dynamic fees that increase with time until the price resets with a new solver update. This means arbitraging becomes less profitable because prices rarely deviate from the “no-trade zone” (as shown in pale red in the lower diagram).
Price evolutions on an AMM with constant fees vs. an AMM with dynamic fees. With HOT, Arrakis uses dynamic fees that increase with time until the price resets with a new solver update. This means arbitraging becomes less profitable because prices rarely deviate from the “no-trade zone” (as shown in pale red in the lower diagram).

Uniform dynamic fees charge the same price to arbitrageurs and retail. This fee can be calculated by finding the pair’s historical price volatility as a time-weighted average and adding a dynamic fee to the base fee (as implemented in Trader Joe and Algebra). Alternatively, a fee can be calculated by looking at the volatility across many pools (as implemented in Ambient Finance).

We can think of uniform dynamic fees in the context of ride apps like Uber. The user must pay a base fee for their ride but the app may charge them a premium if they want a bigger car or they’re ordering at peak time. It doesn’t matter who makes the order—the app does not differentiate between a power user or someone who orders a ride once a year.

Source-dependent dynamic fees aim to identify toxic flow so that arbitrageurs are taxed while retail is not impacted. The following variables can be analyzed to identify arbitrageurs:

  • Transaction source (by looking at a wallet’s trade sizes, total MEV extracted, and other variables)

  • Protocol source, where solvers route to a protocol that minimizes MEV and the protocol is rewarded with lower fees (as seen in CoW Swap)

  • Pattern matching, where trades that closely track previous arbitrages are assigned a higher fee (as seen in Project Guidestar)

If we imagine Uber implementing a source-dependent dynamic fee model, users would pay a different fee based on their app rating. So if they had previously canceled lots of orders or misbehaved in the car, they’d be charged a premium, and if they had been a polite customer, their fee would come in cheaper. The system looks at the user’s history and charges them a fee based on their behavior.

Identifying arbitrageurs to implement dynamic fees requires access to advanced software and data. Such systems pose challenges around reliability and decentralization. While dynamic fee models are still new in the AMM space, they could be a useful way to protect LPs in the future.

MEV Auctions

Several solutions use auctions to sell the right to capture MEV. Rather than implementing a new fee model, these solutions charge searchers a tax. Auctions find a market price for captured MEV and use the proceeds from the high bid to repay LPs and token issuers.

Sorella’s Angstrom will use two types of auctions to protect traders from arbitrage and frontrunning. Angstrom will run arbitrage auctions where searchers bid for the right to arb pools at zero fees, with the auction proceeds redirected to LPs. A batch auction then establishes a fair clearing price for LPs from the users’ sell orders and buy orders. LPs only provide liquidity at the fair price, eliminating the risk of providing stale quotes.

Other models like am-AMM, proposed by Uniswap Labs and Paradigm, explore running auctions for the right to set trading fees in a pool. The emergence of such MEV auction designs offers a preview into how the next wave of AMM solutions may prioritize MEV recapture to protect users in the future.

Chain-Level Solutions: Unichain Unpacked

As L2 expands, new chain-level solutions represent the next evolution in MEV recapture. Unichain, the Superchain L2 launched by Uniswap, is the clearest example today.

Unichain will take two approaches to solve MEV at the rollup layer: priority ordering and ultra-fast block times.

Priority Ordering

Priority ordering redirects MEV to proposers because users pay a fee to express the urgency of their transaction and proposers give priority to those who pay the biggest fees. If a searcher can make 10 ETH from arbing a pool, they will spend more on gas than the minnow moving 0.1 ETH from Base to OP Mainnet. As a result, priority ordering creates auctions for MEV, where searchers bid with gas to win the right to extract MEV.

Applications can use priority ordering to recapture the bulk of MEV for users instead of sending it to the proposer. This is because a smart contract can charge its own fee with a multiplier on the sum paid to the proposer. If the contract asks the searcher to pay a fee 99x higher than the one they pay to the proposer, the application can claim most of the MEV. Dan Robinson and Dave White describe this fee as a MEV tax in the June 2024 piece Priority is all you need.

Faster Block Times

Unichain will use priority ordering to redistribute MEV to users. Additionally, Unichain aims to mitigate LVR by increasing block speeds. Most MEV gets leaked on Ethereum L1 in part because of its 12-second block times, which leave enough time for sophisticated actors to extract value. When blocks are produced faster, arbitrageurs have less opportunity to rebalance pools and LPs have lower adverse selection costs. Unichain will produce new “Flashblocks” every 250ms.

Unichain is implementing priority ordering and Flashblocks through Rollup-Boost, a platform co-designed by Flashbots, OP Labs, and Uniswap. Both solutions are Rollup Extensions within Rollup-Boost.

Unichain has set an early template for other AppChains to follow, internalizing MEV to offer users a better experience. It’s likely that other chains will build on Rollup-Boost in the future. Arrakis Pro will be a day one launch partner for Unichain as part of our commitment to double down on MEV recapture and support LPs.

Application Specific Sequencing

One specific type of priority ordering is Application Specific Sequencing (ASS), where applications can set their own sequencing rules. Applications can use ASS to minimize MEV extraction.

Arguably the most successful implementation of ASS to date is in CoW Swap, which uses batch auctions to order intents, protecting LPs from MEV. But a wave of nascent projects are experimenting with ASS models to usher in a new era of MEV protection on Ethereum.

Solutions like Semantic Layer provide a layer for applications to implement customized transaction execution, thereby reducing MEV extraction. Spire Labs is another example of a stack that lets apps adopt ASS, except it focuses on supporting based AppChains that offer composability with L1.

ASS is a nascent design space today with few applications using it. This is partly because applications want composability, but establishing customized sequencing rules without breaking composability is difficult. If Ethereum ships preconfirmations in the future, builders will be a step closer to solving these challenges.

Dynamic Fees and MEV Auctions on Arrakis

Arrakis is focusing on MEV protection and recapture as part of its roadmap. Arrakis is building onchain solutions to vertically integrate into the MEV supply chain, starting with HOT AMM. Built by Arrakis and Valantis Labs, HOT lets solvers unlock liquidity through signed quotes. Valantis is a groundbreaking modular DEX and HOT is one of its modules.

How Arrakis adopts a MEV-aware, intent-based design to empower LPs with HOT AMM. 
How Arrakis adopts a MEV-aware, intent-based design to empower LPs with HOT AMM. 

In the future, Arrakis will let other actors in the MEV supply chain participate in sealed-bid auctions to win access to liquidity.

We’re also exploring how innovations like Uniswap V4’s Hooks could be used to manage liquidity without exposing LPs to MEV. It’s clear to us that AMMs are entering a renaissance and a new wave of cutting-edge solutions are using innovative onchain mechanisms to minimize MEV and mitigate LVR. We’ll share more news on how Arrakis fits into this landscape very soon.

To learn more about how next-generation AMMs are solving DeFi’s MEV problem, revisit part one of our series on the AMM Renaissance here and discover how HOT fits into the Arrakis ecosystem here.

This piece was inspired by an in-depth research report compiled by Ari Rodriguez and Daniel Contreras at Arrakis Finance. Thanks to both for their feedback and editing contributions to this piece. Read the full report here.

References

Subscribe to ✨Arrakis Finance
Receive the latest updates directly to your inbox.
Mint this entry as an NFT to add it to your collection.
Verification
This entry has been permanently stored onchain and signed by its creator.
Arweave Transaction
o-FrbBWZZTF6rbv…m-3KAUWzIIfg-gM
Author Address
0xD431B7c1433FbA9…211F4773B95E9d9
Content Digest
6JfFl8XicC_ut7d…CUVIOIE_PS-Jig0