Editor's Note: Crypto AI empowers autonomous agents to manage assets, optimize capital flows, and operate autonomously in the DeFi ecosystem. Compared to Web2 AI, it can access decentralized data, utilize open models for collaboration, and accelerate evolution. With the development of DeFi, Darwin-style AI, and decentralized infrastructure, AI will no longer be just an assistant but a direct participant in the on-chain economy, achieving asset ownership, trading, and value creation. Crypto AI combines programmable currency with intelligent agents, building a decentralized economic system and accelerating the arrival of the autonomous agent economy, breaking through the limitations of Web2 AI.

The following is the original content (slightly edited for readability):

As the market tightens and capital gradually concentrates on stronger fundamentals, the next wave of innovation in the AI field is accelerating its collision with the core moat of the crypto world.

Here are several key areas where crypto and AI may further integrate, demonstrating and consolidating crypto-native AI application scenarios.

The most direct synergy point between AI and crypto: capital efficiency and yield optimization.

DeFi - On-chain Yields

Stablecoins

RWA

Spot & Perpetual Contract Trading

Lending Markets

Yield Markets (Interest / Points)

DeFi has always been the core of the crypto world, providing globally accessible on-chain yield and trading opportunities. The addition of AI can more efficiently capture and optimize these values, enabling better utilization of idle capital. For example, DeFi can be used as a tool to hedge against inflation or generate excess returns through AI strategies.

Stablecoins: As the core use case of cryptocurrencies, covering almost all on-chain trading scenarios.

RWA: Tokenizing assets such as government bonds, bonds, real estate, DePIN loans, GPU computing power, etc., and introducing them on-chain.

Spot & Perpetual Contract Trading: Optimizing trading fees and yields.

Lending Markets: Improving capital utilization through more efficient lending mechanisms to achieve better yields.

Yield Markets: Introducing new interest rate markets and enhancing yield optimization capabilities.

Crypto = DeFi = Capital Flow and Appreciation. Web3 AI may be better at this than closed Web2 systems because the openness and incentive mechanisms of blockchain and token economics allow AI to manage funds more efficiently.

Although DeFi AI is still in its early stages, there have been some exciting developments:

·@gizatechxyz's stablecoin yield optimization AI agent has exceeded $1M TVL, with trading volumes of $6M, and yields 83%+ higher than traditional lending strategies.

·@Cod3xOrg launched the Sophon Spark trading agent competition, with agents competing for a $1.5M reward and optimizing AI trading capabilities through data.

·Modius & Optimus by @autonolas, serving as personal portfolio management AI agents. Their team is the only one supporting users to run AI agents locally, recently launching a $1M Olas accelerator program.

·Projects like @HeyAnonai, @AIWayfinder, @slate_ceo are exploring more user-friendly DeFi entry points, though still in early stages.

Why are AI Agents Suitable for DeFi?

AI agents can continuously optimize yields and manage risks 24/7, intelligently adjusting positions. MCP (Multi-Protocol Compatible) drives deep integration of DeFi and AI, enabling AI agents to access on-chain data and integrate more protocols. Within the next year, AI agents may handle a large number of on-chain transactions, automating DeFi operations and enhancing yield optimization capabilities.

Directions Worth Watching:

Teams driving technological advancement and building developer ecosystems (hackathons, competitions, workshops, etc.).

Teams focusing on privacy, verifiability, and non-custodial modes, ensuring users truly control AI agents.

Growth data of AI agents, such as AUA (Agent Managed Assets) / TVL (Total Value Locked).

Beyond DeFi, AI is triggering an evolutionary race. Crypto AI is not just a yield optimization tool; it is accelerating the natural selection of AI agents and teams - only the strongest AI agents and teams will survive and thrive.

Darwin's Law of AI Evolution (Natural Selection)

@opentensor (AI Computation Network)

@AlloraNetwork (Machine Learning / Prediction)

@BitRobotNetwork (Robotics)

Darwinism: "Evolution of species through natural selection". In other words, this is a "Hunger Games" for AI teams - either drive technological progress and gain incentives, or be eliminated by the market.

Web3 AI provides the most suitable infrastructure for AI evolution, accelerating the process of survival of the fittest through token incentives, inflation/burn mechanisms. Bittensor pioneered this trend, with many teams building technologies around its subnets (such as SN6, 41, 44), especially in the GambleFAI (prediction markets) domain, leveraging AI/ML predictive capabilities to gain a competitive advantage in markets.

Allora is leveraging the power of machine learning to accelerate and enhance their models, covering a wide range of financial prediction application scenarios. Allora's model is similar to Bittensor but focuses on financial predictions, using "Topics" (specific financial prediction use cases) instead of subnets, where development teams compete, with the best-performing teams receiving the most incentives.

Best Case: Allora collaborating with @steerprotocol to use AI-driven liquidity provision strategies, creating higher returns for positions while reducing Impermanent Loss (IL).

Bit Robot, developed by the @frodobots team, who are also behind @SamIsMoving (in the @virtuals_io ecosystem). Currently, there is limited information about Bit Robot, but they plan to build an ecosystem similar to Bittensor, focusing on robotics. Its subnets will represent different sectors in the robotics field, such as data, hardware, visual models, LLMs, etc.

Focus points: $TAO price trend, dTAO ecosystem growth, how consumer applications/agents utilize subnet technology, Allora ecosystem integration, case studies, and TGE (Token Generation Event).

Key Elements of Decentralized Infrastructure:

Data

Model Creation / Training

Verifiability

Confidentiality

DePIN (GPU)

This type of infrastructure supports open collaboration, open innovation, and prevents technological innovation from being monopolized by a few centralized players. As I mentioned in previous articles, with the continued advancement of DeFAI and Darwin-style AI evolution, we will see continued adoption of these infrastructures, especially as more mature and clear application scenarios emerge.

In the short to medium term, the areas I'm most interested in:

Social & Sentiment Data:

·@KaitoAI Yap Leaderboard and recently launched Yaps Open Protocol, allowing teams to build products based on Yap scores

·@aixbt_agent tracking & mapping project Alpha / social trends on Twitter

·@cookiedotfun providing AI agent marketplace / social intelligence

On-chain Data:

·Currently, no absolute leader has been seen in on-chain data compared to social & sentiment data.

Other Data Players

·Data Scraping: @getgrass_io collecting data using idle bandwidth

·Data Ownership: @vana incentivizing data ownership through DataDAOs

·Confidential Computing: Blind Compute by @nillionnetwork, with related applications and upcoming $NIL TGE (coming soon)

More in-depth reading about the data domain:

About DePIN (GPU)

Currently, two interesting protocols are emerging that promote GPU asset financialization through on-chain lending, helping data centers and operators massively expand their GPU businesses.