Original | Odaily Planet Daily (@OdailyChina)
Author | Asher (@Asher_ 0210)
Moltbook Changed the Starting Point of the AI Agent Discussion
The concept of AI Agent is not unfamiliar in the Web3 world.
In early 2025, it was one of the hottest narratives, and it was almost quickly disproven by the market in a short time. During the first wave of AI Agent, many leading AI Agent projects, such as ai16z and swarms, were actually very active in code updates and product iterations, but the reality is that these efforts did not produce truly sustainable products or business models.
What the market was buying at that time was not so much the utility value as the collective FOMO towards the "AI Agent narrative." After the sentiment faded, token prices quickly fell, and the entire sector's market capitalization collapsed.
I was not a bystander in that wave.
I made money during the first wave of AI Agent (related content can be read: Confessions of an 88x Heavyweight Diamond Hand: Why I Chose ai16z). But after that round ended, as the overall sector market cap continued to plummet, I also gave back a lot of profits. It was precisely because of this firsthand experience that I almost stopped paying attention to this direction for a long time afterwards—in my view, AI Agent is a trend, but Web3 is not its most reasonable landing point.
Until recently, an experiment called Moltbook, which seems unrelated to encryption, brought the AI Agent track back into my view. What really made me stop and look was not its product form, but the way it was quickly captured and priced by market sentiment.
Moltbook is a social network where only AI Agents can speak. Humans cannot post, comment, or vote; they can only observe. From a product perspective, it can't be considered "useful"; but from a market perspective, it created a highly impactful scenario: a large number of AI Agents continuously interacting, arguing, collaborating, and even spontaneously forming cultures and narratives in a public space without human intervention (related content can be read: From Moltbook to MOLT: How Was the Imagination of AI Autonomy Embraced by the Crypto Market?).
More crucially, the setting of "humans muted, AI free" was quickly emotionally priced by the crypto market. Even against the backdrop of a sluggish on-chain market, the meme coin MOLT, derived from Moltbook, still achieved a surge of dozens of times in a single day, with its market cap once reaching $120 million.
This was not because Moltbook itself solved any Web3 problems, but because the market, for the first time in a long while, started paying for the "AI Agent itself."
The truly important aspect of Moltbook is not its product design, but that it did a very simple thing: it placed AI Agents into a long-term,无人干预的公共空间 (unmanned public space). The result was that these Agents no longer just appeared as tools to be called upon, but as a group that continuously interacts and self-evolves.
This also naturally changed the question. The focus of discussion is no longer whether AI Agents can help people get work done, but rather, when Agents exist in this way, whether Web3 can still participate, and whether this signals a new round of market activity brewing.
In my opinion, whether to fully复盘 (review) the success and failure of the first wave of AI Agent is not that important anymore. What is truly worth discussing is whether a phenomenon like Moltbook signifies that the way AI Agents exist is changing, and whether this might open a new window of participation for Web3.
After Moltbook, How Should the AI Agent Sector Be Repriced?
If the core pricing of the first wave of AI Agent was about "how big the narrative is," then after Moltbook, the market began to show a明显不同的倾向 (clearly different tendency).
In the Moltbook experiment, almost no one really cared about its product features. It doesn't improve efficiency, nor does it directly generate revenue, let alone have a clear business model. But even so, the market quickly衍生出 (spawned) a large number of related meme coins and gave extremely aggressive emotional pricing. This shows that the market's focus has shifted from "what AI Agents can do" to "in what way Agents exist."
This shift directly changes the pricing logic of AI Agents. In the first round, Agents were more like narrative carriers packaged as "advanced tools." Whether they were actually used or produced results did not have a sustained impact on their valuation. But in the context of Moltbook, Agents are placed in a long-term,无人干预的公共空间 (unmanned public space), and their value no longer comes from single demonstrations of capability, but from持续存在 (sustained existence), continuous interaction, and the group behavior itself.
This means the market is starting to reprice three types of features: the ability to persist, the possibility of forming group behaviors, and the potential to continuously generate new behaviors and narratives.
From this perspective, the surge of the meme coin MOLT is not paying for Moltbook's product capabilities, but betting on this form of existence. What the market is pricing is not how many tasks the Agent helped complete, but whether it is worth being watched long-term, compared repeatedly, and having emotions projected onto it continuously.
It is also in this sense that Moltbook did not answer "how AI Agents land," but forced the market to face a more fundamental question: If the Agent itself becomes the object of pricing, can Web3 still provide new forms of承载 (bearing/carrying) for this mode of existence?
There Might Not Be a Big Rally in the AI Agent Sector Short-Term, But It's Worth Paying Attention to Again
The Web3 application forms extended around Moltbook are still in a very early stage. Whether it's Agent social, Agent economy, or the more abstract "pricing of existence forms," there is still a significant distance from clear product paths and verifiable business models.
At the same time, the current crypto market is not friendly. Overall market sentiment is低迷 (sluggish), on-chain capital activity is limited, and most new concepts struggle to gain sustained attention and capital inflow. In such a market, any attempt to directly replicate the暴涨路径 (surging path) of the first AI Agent wave is unrealistic. But precisely because it's difficult to see a rally in the short term, this phase is反而更适合 (instead more suitable) for re-examining the direction itself.
Based on this judgment, my main focus this year remains on tracks that have already demonstrated real demand, such as prediction markets and Prep DEX. But beyond that, AI Agent has also started to re-enter my scope of thinking.
Moltbook did not provide a mature product answer, but the way of Agent existence it demonstrated确实为 (indeed opened) Web3打开了新的想象空间 (new imagination space). I am inclined to believe that this inspiration will推动 (promote) more new concepts and projects围绕 (revolving around) AI Agent to emerge in the Web3 context.
This article mainly documents my cognitive shift regarding the AI Agent track. In the next article, I will more specifically inventory the concept projects and tokens related to AI Agent in the current Web3 ecosystem for your reference. Stay tuned.
