Author: EigenLayer Research
Compiled by: TechFlow
Ethereum launched Maker in December 2017, ushering in the era of decentralized finance (DeFi). Subsequently, Uniswap and Compound were launched, building a new economic ecosystem around ETH and ERC20 tokens. Since then, we have witnessed the booming development of on-chain finance, the centralized liquidity has improved the efficiency of capital use, the perpetual contracts (perps) have continued to evolve, and even the flash loan, an innovation that cannot be achieved in traditional finance, has emerged.
However, we seem to have hit a wall. Since the “merger,” Automated Market Maker (AMM) liquidity providers (LPs) have lost over $700 million in Miner Extracted Value (MEV). Derivatives exchanges have centralized risk management and order books to improve efficiency. In addition, we are unable to provide personalized loan services, offer more favorable rates to users with low default risk, or easily provide fixed-rate loans for a fixed period.
Many of the problems stem from the limitations of Ethereum as a finite state machine. It is limited by gas fees, has a 12 second block generation time, and cannot natively receive off-chain data. Modular architecture provides us with a way forward by offloading heavy computations and integrating external data without sacrificing Ethereum's core security.
If the EVM is the glue that lets developers write arbitrary business logic, then what form should these coprocessors take? While Vitalik refers to these coprocessors as precompilations or opcodes, we need a broader solution. We need coprocessors that can handle tasks that are computationally expensive or impractical for the Ethereum finite state machine, and most importantly, these coprocessors must be verifiable.
Figure: Adhesive and coprocessor architecture modified from Vitalik
Developers have been building efficient and specialized services for years, but verifiability changes all that. This is where EigenLayer's value lies: it provides the infrastructure to create a decentralized network of node operators that can cost-effectively run any of your node software.
We call these decentralized networks Active Verification Services (AVSs), and they significantly reduce the cost of building verifiable and trustless services.
The combination of decentralized finance (DeFi) and active verification services (AVSs) opens up a range of powerful new use cases:
Trustless off-chain computation (coprocessor): Perform heavy computation off-chain and return the results to the chain with minimal gas fees, with security guaranteed by zero-knowledge proofs or crypto-economics. Think of it as free limit orders or even AI model calls, all of which are verifiable and decentralized.
Trustless off-chain data (verifiable oracles, zkTLS): Securely bring real-world data — like prices, volatility, real-time liquidity, and even sports data — into DeFi.
Going a step further: auction networks, policy layers, decentralized order books — AVSs expand DeFi to areas that were previously unreachable.
We call this new paradigm Smart DeFi because it brings real-time adaptability and personalization to decentralized finance. By leveraging trustless off-chain computation and data, Smart DeFi enables smarter decision making. In this article, we’ll dive into 10 use cases that demonstrate its potential.
Exchanges
Exchanges are a core component of DeFi, but less than 15% of spot trading and only 6% of derivatives trading are conducted on-chain. Smart DeFi has the potential to narrow this gap and make decentralized exchanges (DEX) more attractive in competition with off-chain exchanges.
VIP Tiers: Fee tiers based on trading volume
Centralized exchanges offer tiered fees based on trading volume, not only to foster user loyalty, but also to subsidize market makers so that they can offer tighter spreads and better prices to retail traders, thereby bringing more trading volume to the exchange.
Implementing volume-based fees on a DEX is a challenge. In order to calculate a trader’s volume, a DEX needs to:
Dynamic calculation of transaction volume
Store and update the trading volume of each trader
Tracking the past 30 days of trading volume adds complexity and requires historical data storage and calculations.
Both of these methods are very expensive on-chain. But by outsourcing the computation to a co-processor like Lagrange or Brevis, we can verifiably calculate the trader's transaction amount for each transaction.
How is it achieved specifically?
The coprocessor indexes and stores portions of the blockchain data in a queryable relational database.
The AMM (or Uniswap hook) smart contract calls the coprocessor to execute a SQL query to calculate the trader’s transaction fees over a certain period of time.
The coprocessor returns the verification result to the AMM via a callback, along with a zero-knowledge proof confirming that the calculation was performed on the historical blockchain data.
Figure: How on-chain contracts interact with zkCoprocessor Lagrange
2. AMM Dynamics and Asymmetric Fees
Loss and Rebalance (LVR) is an important issue that affects the profitability of liquidity providers (LPs) in AMMs. LVR arises from price inconsistencies between off-chain exchanges that are constantly trading and on-chain AMMs that trade every block or every 12 seconds on the Ethereum mainnet. Many changes can occur within a block, and at the start of the next block, arbitrageurs will take advantage of price differences between exchanges to arbitrage.
To improve LP profitability, AMMs can adopt dynamic fees and asymmetric fees:
1. Dynamic Fees: Adjust fees based on market volatility. Liquidity providers (LPs) typically perform worse during periods of high volatility. Increasing fees during times of high volatility can protect LPs from adverse trades, while lowering fees during stable periods can incentivize trading. This can reduce liquidity dispersion between different fee tiers and make the user experience for LPs smoother. You can check out this very basic proof of concept.
2. Asymmetric Fees: Inspired by Alex Nezlobin, asymmetric fees will adjust the spread based on external price data. For example, if the price of ETH on a decentralized exchange (DEX) is $1000 and the price on a centralized exchange (CEX) is $1050, the AMM can choose to buy at $980 and sell at $1060 instead of maintaining a symmetric spread around the DEX price, which can more realistically reflect market conditions.
Image: From Alex Nezlobin’s Twitter discussion
In both cases, the AMM requires reliable external data - such as CEX prices or volatility - to adjust fees. However, traditional oracles are risky: centralized operators may fail or provide outdated data. In contrast, zkTLS (network proof) provides a better solution. By cryptographically verifying data directly from the network server, zkTLS eliminates the need for trust in third parties. This provides you with real-time, tamper-proof data, ensuring that AMMs, whether on-chain or through coprocessors, can securely calculate dynamic and asymmetric fees.
3. Auction to redistribute MEV to AMM’s liquidity providers (LPs)
Another way to improve LP profitability involves not only off-chain computation, but also a decentralized auction network. Currently, searchers compete in auctions to get their trades at the front of the block. In effect, arbitrage gains are distributed to searchers, builders, and proposers, rather than LPs and traders. Instead, AMMs could auction the right to be the first to trade through a liquidity pool. If the auction is highly competitive, much of the Loss and Rebalance (LVR) will be recaptured. These gains can be distributed proportionally to the underlying LPs participating in the trade, reducing overall arbitrage and enabling LPs to offer tighter spreads. Sorella is developing this functionality in the form of a Uniswap v4 hook.
The challenge is to run a low-latency, censorship-resistant auction. Running an auction on-chain is too complex and expensive: every bid costs gas. A block is processed before a winner is chosen, so the auction cannot be completed. While a centralized entity could run the auction off-chain, this goes against the very idea of decentralized finance (DeFi) and gives them a last-ditch opportunity to potentially extract value.
The solution is a leaderless auction run by a group of decentralized operators, eliminating reliance on a single entity and ensuring the integrity of the process. The operators are responsible for selecting the winning bid and returning the proceeds to liquidity providers (LPs).
Figure: Leaderless auction from Paradigm
Derivatives
While most derivatives are traded on exchanges, Intelligent DeFi unlocks unique use cases for this asset class. Let’s take a deeper look!
4. Advanced Margin System
Currently, traders cannot express views across assets, such as SOLETH or crosses, without significantly limiting leverage. Most perp DEXs calculate margin linearly based on the sum of a trader’s open interest across different positions.
For example, if I deposit $10,000 and go long $50,000 on ETH and short $50,000 on BTC, that counts as 10x leverage. But that is a different risk profile than someone who is simply long $100,000 on ETH, and the two accounts should not be considered the same. Ideally, traders should be able to leverage more than 5x ETHBTC without being so tightly restricted.
The problem is that on-chain computation is limited. Specifically, the system needs to take into account the collateral of each spot asset, the position of each perpetual contract asset, unrealized profit and loss, the initial and maintenance margin requirements of each perpetual contract, and correlation and delta hedging. This problem becomes particularly important as decentralized exchanges (DEX) expand to multiple asset types, such as perpetual contracts and options.
By utilizing coprocessors to perform more complex calculations to determine an account’s margin factor, decentralized exchanges (DEXs) are able to create a customized risk engine that is more aligned with the needs of the user. This enables more flexible delta-neutral strategies and ensures that liquidations are only made when truly necessary.
To provide greater flexibility, the coprocessor can dynamically adjust margin requirements, taking into account liquidity from major centralized exchanges (CEXs) and the open interest of each pool in real time.
Figure: Aevo relies on a centralized risk engine to assess the worst-case market scenarios, thereby providing more reasonable margin parameters for high-value traders. The coprocessor provides a unique margin calculation method without hindering decentralization, excerpted from Aevo's documentation
5. Pricing of Options Automated Market Makers (AMMs)
Automated market makers (AMMs) for derivatives, especially options, are both exciting and controversial. Some believe they cannot be accurately priced, while others believe that derivatives are only suitable for high-volume assets, for which order books are more efficient. Nonetheless, Panoptic, Deri, and others believe that AMMs are the best way to provide liquidity, including options.
A key factor in making options automated market makers (AMMs) truly successful is the introduction of off-chain data, such as volatility, historical prices, and real-time market signals. In addition, off-chain computation is also essential for building more advanced pricing models such as Black-Scholes. Combining this external data with on-chain trading mechanisms is critical to ensuring accurate pricing, reducing slippage, and improving capital efficiency for options traders.
Borrowing
Lending protocols face unique challenges, and artificial intelligence (AI) and off-chain computation can drive smarter and more flexible solutions.
6. Parameterized AI System
Currently, governance teams of protocols such as Aave and Compound manually update relevant parameters of the lending market. Typically, risk service providers like Gauntlet will conduct model-based simulations and recommend adjustments to parameters such as base rates, collateral factors, liquidation factors, etc. In the event of more serious incidents, they can recommend delisting or freezing certain assets.
This approach has two major flaws:
There are too many delays. When I was a delegate on the Aave DAO, it would take at least a week for a proposal to get passed.
The governance team does not have enough information about voting on lending parameters, and not all members actively participate. The recent Compound governance attack is a typical example.
Figure: According to Aave’s documentation, its governance process takes at least 5 days
Morpho and Euler v2 are important steps in the right direction. They modularize the risk management part, allowing anyone to create their own instance of a lending platform. Users can choose where to store their assets based on the history and reputation of the curators. This approach can effectively reduce the time required to update parameters.
But in an ideal system, parameters are automatically updated in real time in response to on-chain and off-chain liquidity. Artificial intelligence (AI)-based models can simulate multiple scenarios to predict and avoid the worst-case scenarios. These models rely on AI-specific coprocessors such as Ritual, Sentient, Hyperbolic, Ora, and Valence to process large amounts of data off-chain, taking into account volatility, liquidity changes, and risk correlations, and then verifiably publish the results on-chain.
7. Customize loans based on account history and liquidation risk
In traditional finance, creditworthy borrowers typically receive better loan terms, while in decentralized finance (DeFi), all borrowers enjoy the same loan terms regardless of their risk profile or credit history. While this model has its advantages, I believe DeFi can combine the best of both worlds: providing fair trustless loans to anyone, while providing better loan terms to repeat borrowers with good credit histories and low liquidation risk.
Due to the lack of differentiation, decentralized finance (DeFi) lending protocols are unable to provide personalized terms to low-risk borrowers, such as lower collateral requirements or more favorable interest rates. This lack of personalization not only limits the potential benefits of repeat users, but also leads to inefficiencies in the lending market.
Providing personalized loans first requires an anti-Sybil solution to ensure that only verified users can get better loan terms. Solutions like WorldCoin or Coinbase Verification can effectively prevent malicious actors from repeatedly exploiting the protocol through bad debt.
Once the borrower is verified, the protocol can collect on-chain information to build a liquidity profile, including:
Current and historical loan history
Repayment status of previous loans
Net assets and outstanding debt on the chain
NFTs owned (if the lending protocol partners with NFT projects to offer favorable terms)
The protocol can even look at other addresses associated with the same identity to get a more complete picture.
Finally, the coprocessor can assess liquidation risk and generate customized collateral factors and interest rates, ensuring tailored loan terms for each borrower.
8. Compliant Privacy Mixers
In August 2022, the U.S. Office of Foreign Assets Control (OFAC) sanctioned Tornado Cash for facilitating money laundering. However, privacy is a fundamental right with legitimate uses: people should be able to transfer funds to other accounts or friends without making their full transaction history public. The problem is that existing privacy mixers cannot distinguish between legitimate users and malicious actors. This lack of compliance makes them targets for sanctions and hinders wider use.
What if we could create a privacy protocol that only accepts compliant funds? This protocol would manage risk and comply with regulations, and privacy-conscious users would flock to the system. However, determining compliance requires multiple on-chain and off-chain data, which is not a simple task. Ideally, a smart contract could call an API and only approve valid transactions, thereby ensuring compliance.
Aethos is a policy layer that aims to achieve this. It enables developers to set rules at the smart contract level to ensure that transactions comply with specific compliance policies. For example, a compliant privacy mixer can set transaction limits, time locks, and block transactions where deposits or withdrawals are made to addresses that are sanctioned by OFAC or are related to DeFi hacks.
Image: From Aethos documentation
Integrating real-time, rules-based policies into smart contracts ushers in a new era of institution-friendly DeFi, where compliance and DeFi’s values are no longer in conflict.
9. Automated rebalancing yields protocol
DeFi provides a wealth of yield opportunities, covering a variety of assets and protocols, including staking, re-staking, lending, automated market making (AMM) liquidity pools, real world assets (RWAs), etc. Users have different risk preferences, which are closely related to the protocol type, chain (such as Ethereum, Solana, etc.), asset denomination, and external market risks. Faced with so many choices, some traders choose yield protocols to automate their capital allocation.
Figure: Such rich returns, non-financial advice, from DeFiLlama
These protocols can use AI models to optimize returns across multiple sources. Developers set predefined risk parameters, such as limiting exposure to 15% per protocol or avoiding protocols with a total locked value (TVL) below $100 million, and the AI model is responsible for adjusting the portfolio to meet these criteria and maximize returns.
Going one step further, AI models can create personalized revenue strategies for each user based on their on-chain activity and preferences collected through a short questionnaire. This kind of personalized service—which was previously unimaginable—is now within reach thanks to the scalability of AI.
In the background, AI-driven co-processors monitor and rebalance the portfolio. They make adjustments only when gains offset gas costs, ensuring efficient data-driven portfolio management.
10. Ultra-precise incentive plan
Incentives are at the core of crypto and DeFi. The DeFi summer really began in 2020 with the introduction of liquidity mining on Compound. By providing rewards for specific user actions, protocols are able to drive growth and activity.
But as the space matures, protocols are seeking more precise targets, often turning to off-chain processes. Automated market makers (AMMs) may focus on incentivizing active liquidity providers. NFT markets and prediction markets may reward liquidity that is close to the order book price. Lending protocols may encourage non-circular borrowers to borrow at least 20% of their interest.
Through the coprocessor, protocols can define complex reward conditions and issue rewards in real time. This gets rid of the increasingly unpopular point system and provides users with certainty of reward payments, thereby reducing the capital cost of the protocol. Gearbox has begun using Lagrange's coprocessor to handle multi-asset rewards with different payment mechanisms. By making incentives more efficient, DeFi can continue to grow while rewarding the most valuable behaviors.
Summarize
The combination of DeFi and AVS will usher in a new financial revolution. From automated market makers (AMMs) for MEV recycling to real-time policies in privacy protocols, these use cases are just a small part of the potential of decentralized finance.
