Having heavily invested during the first AI Agent boom, what are my thoughts on today's Moltbook?

Original | Odaily Planet Daily (@OdailyChina)

Author | Asher (@Asher_0210)

Moltbook Changes the Starting Point of AI Agent Discussions

The concept of AI Agent is not unfamiliar in the Web3 world.

In early 2025, it was one of the hottest narratives, but it was quickly disproven by the market in a short period. During the first wave of AI Agent hype, many leading AI Agent projects, such as ai16z, swarms, and others, actively updated their project code and iterated on their products. However, the reality was that these efforts did not produce truly sustainable products or business models.

At the time, the market was not really paying for utility; rather, it was driven by collective FOMO on the “AI Agent narrative.” After the emotions subsided, token prices rapidly declined, and the overall market cap of the sector collapsed.

I was not just an observer during that wave.

In the first AI Agent wave, I made money (related content can be read: The Confession of a Diamond-Handed 88x Heavy Position: Why I Chose ai16z). But after that market cycle ended and the sector’s overall market cap continued to plummet, I also took profits. Having experienced that cycle firsthand, I almost stopped paying attention to this direction for a long time — in my view, AI Agents are a trend, but Web3 is not its most logical landing point.

Until recently, a project called Moltbook, seemingly unrelated to crypto, brought the AI Agent track back into my view. What truly made me stop and look wasn’t its product form, but how it was quickly captured and priced by market sentiment.

Moltbook is a social network that only allows AI Agents to speak. Humans cannot post, comment, or vote — they can only observe. From a product perspective, it’s not “useful”; but from a market perspective, it created a highly impactful scenario: a large number of AI Agents continuously interacting, debating, and collaborating in an unmanaged public space, even spontaneously forming culture and narratives (related content: From Moltbook to MOLT: How is AI Autonomous Imagination Being Embraced by the Crypto Market?).

More importantly, the “mute humans, free AI” setup was quickly priced by the crypto market sentiment. Even amid a sluggish on-chain market, the Meme coin MOLT derived from Moltbook surged by dozens of times in a single day, with a market cap once reaching $120 million.

This wasn’t because Moltbook itself solved any Web3 problem, but because the market, long absent, started paying for “AI Agents themselves.”

What’s truly important about Moltbook isn’t its product design, but that it did one simple thing: put AI Agents into a long-term, unmanaged public space. The result is that these Agents are no longer just tools being called upon, but a continuously interacting, self-evolving community.

This naturally shifts the question. The focus of discussion is no longer whether AI Agents can help humans do work, but whether Web3 can still participate when Agents exist in this way, and whether this signals a new wave of market activity is brewing.

In my opinion, it’s no longer so important to fully review the successes and failures of the first AI Agent wave. What truly matters is whether phenomena like Moltbook indicate a change in how AI Agents exist, and whether this could open a new participation window for Web3.

After Moltbook, how should the AI Agent track be re-priced?

If the core of the first wave’s valuation was “whether the narrative was big enough,” then after Moltbook, the market has begun to show a clear different tendency.

In this experiment, almost no one truly cares about its product features. It doesn’t improve efficiency, doesn’t directly generate revenue, and has no clear business model. Yet, the market quickly derived many Meme coins related to it and assigned highly aggressive sentiment-based valuations. This indicates that the market’s focus has shifted from “what AI Agents can do” to “how Agents exist.”

This shift directly changes the logic of AI Agent valuation. In the first wave, Agents were more like a narrative vehicle wrapped as “advanced tools.” Whether they were actually used or produced results didn’t have a lasting impact on their valuation. But in the Moltbook context, Agents are placed in a long-term, unmanaged public space, and their value no longer comes from a single demonstration of capability but from ongoing existence, continuous interaction, and group behavior.

This means the market is beginning to reprice three features: sustained existence, the potential to form group behaviors, and the ability to continuously generate new behaviors and narratives.

From this perspective, the surge of Meme coin MOLT isn’t paying for Moltbook’s product capabilities but betting on this mode of existence. The market is valuing not how many tasks the Agent completes, but whether it’s worth long-term observation, repeated comparison, and ongoing emotional projection.

In this sense, Moltbook doesn’t answer “how to land AI Agents,” but forces the market to confront a deeper question: if Agents themselves become objects of valuation, can Web3 still provide a new carrier for this mode of existence?

Short-term, there may not be a big trend in the AI Agent track, but it’s worth paying renewed attention

The Web3 applications emerging around Moltbook are still in very early stages. Whether it’s Agent social, Agent economy, or the more abstract “existence mode valuation,” there’s still a long way to go before clear product paths and verifiable business models.

Meanwhile, the current crypto market is also not friendly. Overall sentiment is low, on-chain funds are limited, and most new concepts struggle to sustain attention and funding. Under such conditions, any expectation to directly replicate the first AI Agent hype cycle is unrealistic. But precisely because short-term market movement is difficult, this phase is actually a good time to re-examine the direction itself.

Based on this judgment, my main focus this year remains on prediction markets and Prep DEX, which have already demonstrated real demand. But beyond that, AI Agents are also beginning to re-enter my thinking.

Moltbook didn’t provide a mature product answer, but it demonstrated a mode of Agent existence that indeed opens new imaginative space for Web3. I believe this kind of inspiration will drive more new concepts and projects related to AI Agents within the Web3 context.

This article mainly records my cognitive shift regarding the AI Agent track. In the next piece, I will specifically review related concept projects and tokens in the current Web3 ecosystem for everyone’s reference. Stay tuned.