Written by: 0xJeff, Crypto Kol
Compiled by: zhouzhou, BlockBeats
Editor's note: In this article, Jeff discusses how DeFAI (Decentralized Finance + Artificial Intelligence) simplifies, optimizes, and enhances the DeFi experience through abstraction layers, automated trading agents, and AI-driven dApps, introducing several emerging DeFAI projects such as Almanak, Cod3x, and Mode, highlighting the role of AI in enhancing trading strategies and portfolio management.
The following is the original content (for readability, the original content has been reorganized):
DeFi has always been a pillar of Web3. It makes blockchain practical, providing tools for instant transfer of funds globally, investing in assets on-chain, peer-to-peer lending, and stacking strategies across DeFi protocols. This is the financial freedom within reach.
More importantly, DeFi solves real-world problems. It allows unbanked individuals to access financial services, removes intermediaries, and operates around the clock, creating a truly global and inclusive financial system.
But when we face reality, we realize that DeFi is complicated; setting up wallets, managing transaction fees, avoiding scams and rug pulls—it's not suitable for the average person. The ever-expanding L1, L2, and cross-chain ecosystems only make things more complicated. For most people, the entry barrier is just too high.
This complexity has limited the development of DeFi, but with the emergence of DeFAI, this situation is beginning to change.
What is DeFAI?
DeFAI (DeFi + AI) makes DeFi more accessible. By leveraging AI technology, it simplifies complex interfaces and removes barriers for ordinary people to participate. Imagine a world where managing your DeFi portfolio is as simple as chatting with ChatGPT.
The first wave of DeFAI projects has started to emerge, focusing primarily on three areas:
1. Abstraction Layer
2. Autonomous Trading Agents
3. AI-driven dApp
1. Abstraction Layer
The goal of the abstraction layer is to hide the complexity of DeFi through an intuitive interface. They allow users to interact with DeFi protocols using natural language commands instead of complex dashboards.
Before AI, abstract layers like intent-based architecture simplified trade execution. Platforms like CoWSwap and symm io allow users to get the best pricing in decentralized liquidity pools, addressing liquidity fragmentation, but they do not solve the core issue: DeFi still feels challenging.
Now, AI-driven solutions are filling this gap:
Griffain is the first project to launch a token and is currently in early access, requiring an invitation to use.
Griffain is more versatile, allowing users to execute a wide range of actions, from simple to complex tasks such as task automation (DCA), launching Memecoins, and conducting airdrops.
Orbit / Grift is the second project to launch a token, aimed at enhancing the on-chain DeFi experience. Orbit emphasizes cross-chain functionality, having integrated over 117 chains and 200 protocols, making it the most integrated of the three protocols.
Neur is the third project to launch a token but has quickly outpaced Orbit in valuation due to its open-source nature. Neur positions itself as a co-pilot for Solana, designed specifically for the Solana ecosystem. Neur is powered by Sendaifun's Solana Agent Kit.
I personally use slate ceo, which is still in its early stages and hasn't launched a token yet, but I love their automation features. I mainly use it to set conditional trades, such as selling 25% of my position if xxxx reaches a market cap of $5 million, or buying $5,000 worth of tokens if xxx reaches xxxx price.
AIWayfinder is another interesting project worth noting. This is a behemoth created by the PRIME / ParallelTCG team, and it is worth anticipating.
2. Autonomous Trading Agents
Why spend hours digging for Alpha, manually executing trades, and trying to optimize your portfolio when you can let an agent do it for you? Autonomous trading agents have pushed trading bots to a new level, transforming them into dynamic partners that can adapt, learn, and make smarter decisions over time.
It needs to be clarified that trading bots are not new. They have existed for years, executing predefined actions based on static programming. But agents are fundamentally different:
They extract information from unstructured and constantly changing environments.
They reason about data in the context of their goals.
They identify patterns and learn to leverage these patterns over time.
They can perform operations that their owners have not explicitly programmed.
This subfield is rapidly evolving; initial agents may have been for entertainment purposes—like shilling some junk coins for fun—but have now shifted to more practical, profit-driven tools that help users trade more effectively. However, an important challenge remains: how do you verify that an 'agent' is not just a bot, or even a person operating behind the scenes?
This is where the DeAI infrastructure plays a critical role.
The role of DeAI in verifying agents
Key infrastructure like Trusted Execution Environments (TEE) ensures agents can operate securely and tamper-proof.
For example:
TEE: Promoted by PhalaNetwork, TEE provides a secure isolation zone where data can be processed confidentially. Phala's experiments—like Unruggable ICO and Sporedotfun—demonstrate how agents can execute tasks while maintaining data integrity.
Transparent execution / verification framework: Innovations like zkML (Zero-Knowledge Machine Learning) or opML provide reasoning and computational verifiability. Hyperbolic labs' Proof-of-Sampling (PoSP) is a prominent example. This mechanism combines game theory and sampling techniques to ensure computations are accurate and efficient in a decentralized environment.
Why is this important?
When self-agents begin to handle large TVL (assuming $100 million or more), users will demand assurances. They need to understand how the agents manage risk, verify the frameworks they operate in, and ensure their funds are not randomly thrown at some junk coins.
This field is still in its early stages, but we have already seen some promising projects exploring these verifiability tools. As DeFAI evolves, this is a direction worth watching.
For more trends in DeAI infrastructure, please refer to this article:
Top autonomous trading agents I am keeping an eye on
Almanak
Almanak provides users with institutional-grade quantitative AI agents to address the complexity, fragmentation, and execution challenges in DeFi. The platform executes Monte Carlo simulations in real environments by forking EVM chains, taking into account unique complexities like MEV, transaction fees, and transaction order.
It uses TEE (Trusted Execution Environment) to ensure the privacy of strategy execution, protects Alpha information, and implements non-custodial fund management through Almanak Wallets, allowing precise permission delegation to agents.
The infrastructure of Almanak supports the ideation, creation, evaluation, optimization, deployment, and monitoring of financial strategies. The ultimate goal is to enable these agents to learn and adapt over time.
Almanak raised $1 million on legiondotcc and was oversubscribed. Next steps include the release of the beta version and initial strategy/agent deployment, testing with beta testers. Observing the performance of these quantitative agents will be very interesting.
Cod3xOrg / BigTonyXBT
Cod3x, created by the Byte Mason team (known for their work on Fantom and SonicLabs), is a DeFAI ecosystem aimed at simplifying the creation of trading agents. The platform provides a no-code building tool that allows users to build agents by specifying trading strategies, personalities, and even tweeting styles.
Users can access any dataset, develop financial strategies in minutes, leveraging a rich API and strategy library. Cod3x integrates with AlloraNetwork, using its advanced ML price prediction model to enhance trading strategies.
Big Tony is the flagship agent based on the Allora model that predicts entries and exits of mainstream coins. Cod3x is working to create a thriving ecosystem of autonomous trading agents.
A notable feature of Cod3x is its liquidity strategy. Unlike the common alt:alt LP structure promoted by virtuals io, Cod3x uses a stablecoin:alt LP supported by cdxUSD, which is Cod3x's proprietary CDP (Collateralized Debt Position). This structure provides more stability and confidence for liquidity providers compared to the volatility of alt:alt trading pairs.
Cod3x also has its own DeFi primitives, such as liquidity AMOs (Automated Market Operations) and mini-pools, enhancing liquidity and adding more functionality to agents in its ecosystem / DeFi Lego components.
Other notable projects
getaxal / Gekko Agent—Axal's autonomous product, where agents handle complex multi-step crypto strategies. Gekko integrates autonomous driving features. I am waiting to see how Gekko executes data-driven trades in autonomous mode.
ASYM41b07—known by many as the 'cheat code for Memecoin trading', ASYM agents can analyze vast amounts of data from blockchains and social media to predict trends in Memecoins. ASYM has consistently outperformed the market and demonstrated 3-4x returns through backtesting. It will be interesting to see how it performs in live trading.
ProjectPlutus—I just love this name PPCOI
3. AI-driven decentralized applications (dApps)
AI-driven dApps are a promising yet still nascent area within DeFAI. These are fully decentralized applications that integrate AI or AI agents to enhance functionality, automation, and user experience. While this field is still in its early stages, some ecosystems and projects are beginning to stand out.
Within this space, modenetwork is a very active ecosystem aimed at attracting high-tech AI x DeFi developers as a Layer 2 network. Mode is a base for multiple teams working on cutting-edge AI-driven applications:
ARMA: A self-stablecoin mining tailored to user preferences, developed by gizatechxyz.
Modius: A self-agent supported by autonolas for Balancer LP mining.
Amplifi Lending Agents: Developed by Amplifi Fi, these agents integrate with IroncladFinance to automatically swap assets, lend on Ironclad, and maximize returns through automated rebalancing.
At the core of this ecosystem is MODE, the native token. Holders can stake MODE to receive veMODE, which grants them airdrops from AI agents, whitelist access to projects, and more ecosystem benefits. Mode is positioning itself as the center of AI x DeFi innovation, and its influence is expected to grow significantly by 2025.
Additionally, danielesesta has garnered widespread attention through the DeFAI theory of HeyAnonai. He announced that HeyAnon is developing the following:
As an abstraction layer for DeFi interfaces
DeFi agents for autonomous trade execution
Research and communication agents for acquiring, filtering, and interpreting relevant data
The market reacted enthusiastically, with the market cap of the ANON token skyrocketing from $10 million to $130 million. Daniele seems to be bringing back the excitement of TIME Wonderland, but this time with a stronger foundation and clearer vision.
In addition to these two ecosystems, many teams are building their own AI-driven decentralized applications. Once major ecosystems form around them, I will share more information in the future.
Final thoughts
DeFAI is changing DeFi by making it smarter, simpler, and more accessible. As abstraction layers simplify user interaction, autonomous trading agents manage portfolios, and AI-driven dApps optimize use cases, we are witnessing the dawn of a new era.
Rather than the DeFi summer of 2020, 2025 will be the DeFAI summer.
