Recently, the AI Agent track has skyrocketed, and friends have been asking: Is it still possible to get in?
Here are my thoughts: Don’t hesitate, just keep going.

Next, let’s analyze the reasons why AI agents still have 10x potential:

1. Unlike the 'pure expectation + community FOMO' of memes, agents will continuously iterate, and increasingly powerful innovations will emerge.

2. The summer of AI agents is vastly different in both the perceived limits and the pace of realization compared to previous expectations.

Whether from the trillion-dollar industry prospects of web2 AI or the integration of web3 AI on-chain, there is tremendous potential. Currently, the total market cap of AI in the crypto space is only about $10 billion, while the leading protocols during DeFi Summer had market caps in the hundreds of billions, and we haven't even seen the emergence of phenomena like 'GameFi cured my depression' or 'AXS beating Honor of Kings.'

The key point is that more and more incredibly powerful agents will emerge.

For example, aixbt, a newly emerged agent analyzing cryptocurrency opportunities 24/7, has already dominated the mindset of users, surpassing the top influencers in the crypto space, and I believe this trend will only accelerate, not slow down.

Currently, various AI agents are emerging, creating a 'bewildering array of flowers.' As an ordinary person entering this new track full of opportunities and traps, how should one choose? Here’s my personal on-chain investment logic:

1) I see many people almost liquidating all their major CEX coins and going all-in on the AI Agent on-chain track. While this shows confidence in the new trend, it is not advisable.

Because during the pure AI MEME phase, the on-chain AI narrative track still had information discrepancies due to biases among a portion of mainstream groups. Now, at the AI infrastructure stage, 'information discrepancies' have been severely diluted, with a large number of geek developers building, many knowledgeable researchers, and even many VC firms locking their attention on the new AI Agent narrative.

This means that the period for blind investment and riding the trend to profit without effort is no longer available. The risks and opportunities of on-chain investment coexist at least; the risk index of going all-in on the AI narrative will be very high, possibly exceeding the risks at the end of the chaotic era when entering the early stable era. The correct approach is to gradually shift positions to on-chain, avoiding reckless show of hands.

2) The suggested on-chain position layout is 5 + 4 + 1:

Large position choices like ai16z, ZEREBRO, $AIXBT, $GOAT, VIrtual and other consensus projects that have already emerged as pioneering leaders, just like during DeFi Summer when heavy bets were placed on MKR, COMP, UNI, LINK, etc.


Although this approach will result in a short-term opportunity cost by missing out on the chaotic rise of other small coins, looking at it over a longer period, it is definitely a risk-averse strategy that can keep you at the table.

Medium position choices like $arc, $REI, $Focai, $ELIZA, $Swarms, $SPORE, $SNAI, $ALCH, $NEUR, etc., are still in the chaotic battle phase of technical frameworks and innovative models.

If we want to further subdivide, the hierarchy is: framework standards > innovative agent applications > individual AI > vertical platform services. The core logic is that the ceiling of framework standards is definitely the highest, individual AI will be limited by imagination, and innovative applications will have many uncertainties, etc. (seeking growth).

Small positions chasing various small projects that seem appealing at first glance but have significant market volatility before you can even research them, such as $MetaV, $stoic, $SYMX, $TAOCAT, $ZAILGO, $POLY, etc. Many similar projects will emerge in the early stages, but the mix of impostors among them will also be the highest. (We need to be cautious in distinguishing).


From the perspective of the crypto space, attention equates to wealth, and the attention of AI agents will continue to snowball, growing larger and faster; we just need to keep going in the current direction and become even more formidable.