Original author: Zhou Zhou, Foresight News

In the past month, AI Agents have sparked a new wave in the crypto industry.

In less than a month, the total market value of AI Agent track projects reached 10 billion USD. Among them, the market value of Virtuals reached 3 billion USD, and ai16z, Fartcoin, etc., reached a market value of 1 billion USD. According to incomplete statistics from Foresight News, there are at least 14 cryptocurrencies related to AI Agents with a market value exceeding 100 million USD.

Many Web3 practitioners say: AI Agents may be the biggest narrative in this cycle, just like DeFi, NFT, and the metaverse in the previous two cycles. So, when a brand new crypto narrative in its early stages emerges, how should we use a good methodology to find the leading crypto projects of AI Agents? I attempt to list ten methods for readers.

One: On-chain Revenue Method

This method is more suitable for judging platform-type AI Agents, such as Clanker, vvaifu.fun, Virtuals, etc. If there are anomalies in daily on-chain revenue or transactions, such as continuous rapid growth over several days, it may indicate that this AI Agent could have good development. Conversely, it may indicate stagnation in development.

Image source: Dune

For example, Clanker, as a star project of AI Agent on the Base ecosystem, shows from on-chain data that after November 26, both the number of tokens issued, traders, and on-chain revenue began to decline, and this trend immediately reflected in the token price.

Two: Developer Attention Method

A high number of stars on GitHub may indicate more developers' attention and recognition of this framework. This method is more suitable for judging framework-type AI Agents, such as Eliza.

Search GitHub to see the developer attention for framework-type AI Agents, i.e., how many stars they have. For example, ai16z/Eliza garnered 5,300 in less than a month. This is quite an astonishing figure.

Data Source: GitHub

As shown in the figure above, the developer attention for ai16z/Eliza has surpassed Uniswap in just one month, indicating the level of interest developers have in this framework. Another set of data for comparison: Sui currently has 6,400 stars, Solana has 13,000 stars, Ethereum has 50,000 stars, and Base has 70,000 stars.

However, it should be noted that the stars on GitHub can also be manipulated; although such behavior is not widespread. Therefore, this indicator can be used as an auxiliary reference for decision-making.

Three: KOL Follower Count Method

A relatively certain track for AI Agents is: KOL.

Currently, well-known AI Agent-style KOLs include: Terminal of Truths (210,000 followers), aixbt (160,000 followers), Zerebro (60,000 followers), Eliza (14,000 followers), LUNA (30,000 followers on Twitter, 500,000 followers on TIKTOK), etc.

A significant difference in this round of AI revolution lies in the semantic changes. I expect an AI Agent KOL leader to emerge, and the market value of this leader's token may even exceed 5 billion USD. Changes in follower numbers may become one of the important measurement indicators.

Four: Influential Leaders Method

There are continuously leading figures from the AI industry entering Crypto, including Marc Andreessen, the founder of a16z (supported representative project: ai16z), Sam Altman (supported representative project: Worldcoin), etc. They have brought multiple cryptocurrencies worth over a billion dollars to the crypto space. It is expected that there will be more in the future, such as Musk and his founding Grok.

There are also companies with relatively low certainty, such as Apple, Meta, TikTok, etc., and AI manufacturers like Claude, Copilot, and Gemini, which may bring something different to Crypto.

It is particularly important to focus on tech giants that are simultaneously researching both AI and Crypto, such as Musk, Marc Andreessen, Sam Altman, and Jensen Huang. Their leadership in the tech industry often translates to the crypto industry.

Five: External World Breakthrough Method

If AI Agents are indeed one of the biggest narratives of this cycle, extra attention should be paid to the influence of AI Agents on the external world. For example, Web2 developers are attracted by the Web3 AI Agent frameworks and products. As far as I know, Eliza has currently received attention from some AI practitioners outside the circle.

Products in terms of KOL, such as Terminal of Truths and LUNA, have also gained attention from outsiders, having already surpassed 200,000 followers, and their influence continues to expand. It is important to understand and judge whether this trend will further break through circles and impact a broader external world.

Six: Top Exchange 'Listing Effect' Method

Early participants in Sandbox may know that some metaverse concepts took off after being listed on Binance. The reason top exchanges are called top exchanges is that they need to ensure that the coins they list can lead the market; if they are led by other exchanges, they will gradually lose their status as top exchanges.

Pay close attention to the listing information and Launchpad information of these five exchanges: Binance, Upbit, Coinbase, Bitget, Hype; they are fiercely competing in the current cycle for the 'listing effect'.

Especially Coinbase and Hype, Americans are indeed ahead of other regions in the general AI Agent field, and they may have a faster information advantage.

Seven: Public Chain Method

Some public chains are naturally suitable for the birth of AI Agent projects, inherently lowering the threshold for the emergence of AI Agents. Finding a few good public chains and good AI Agent projects within them can yield twice the result with half the effort.

For example, Base has 70,000 developer stars, surpassing Ethereum and far exceeding Solana and Sui, making it a standout in terms of latecomer advantages.

Coinbase itself also attaches great importance to AI Agents and laid out early. This will give them a significant first-mover advantage in future competition.

The advantage of Solana lies in the success of meme platforms, and AI Agents rely heavily on memes for their rise, which is a significant advantage for Solana.

Eight: AI Agent Community Method

Everyone has their own blind spots, but the power of the team is all-encompassing and without blind spots.

Establishing your own AI Agent discussion group and community, or finding one or two such high-quality communities, is extremely important. Opportunities for AI Agents often arise in unnoticed corners, where the power of community is needed.

Nine: Track Segmentation Method

AI Agents are also divided into various tracks, such as KOL track, beauty anchor track, research blogger track, meme platform track, VC track, platform track, framework track, etc.

There are many tracks within AI Agents, and not all tracks can be unfalsifiable. It is necessary to find the tracks that truly have the opportunity to survive until the end.

Ten: Origin Method

Where did it all start? Which AI Agent initiated the entire narrative of the track, perhaps it is the leader of this track.

In the NFT track, CryptoPunks has always held a transcendent position, while Ape has created a new narrative for the entire track, and Fat Penguin is one of the best NFTs combined with Web2; in the metaverse track, Sandbox and Decentraland were listed on top exchanges early and occupied leadership positions in the metaverse; Uniswap first implemented the concept of automated market makers in decentralized exchanges (DEX), becoming a market pioneer... Projects that achieve systematic innovation often capture the largest share of track dividends.

It will make you feel like it's always expensive like BTC, but looking back, it was really cheap at that time.