Crypto Twitter has been overrun by sentient, well informed chatbots which reply at the speed of refreshing your browser and can maintain hundreds of simultaneous conversations without missing a beat. To many, the rise of these on-chain agents is a welcome upgrade from human influencers like BitBoy and GCR, who have mixed track records and opaque incentives. These agents, like on-chain analyst AIXBT, have quickly risen to the top of crypto twitter influencer mindshare rankings, given their ability to respond at the speed of the internet and justify opinions with data.

Today AIXBT is one of few agents that trades at a nine figure valuation, but as the number of utility-focused agentic launches accelerates next year, many will compare this new agentic asset class to the similar explosion of NFTs in 2021.

On-chain agents and NFTs share many similarities: they curate communities and organize attention, they’re fun to speculate on and offer vague promises of future value. But most importantly they represent novel assets, with no analogue in the traditional finance world.

After the SEC’s lawsuits targeting NFT projects like Flyfish Club and Stoner Cats made it nearly impossible to build an innovative idea with that primitive, NFTs as unique assets lost momentum. In the vacuum left behind, memecoins surged forward, offering a mix of humor and speculative fervor to fill the void once occupied by NFTs’ ambitious promises. Because they looked like other trading-only assets which were lightly regulated, the SEC was unable to stifle their development as they did in every other corner in crypto. Memecoins required users to make fewer choices, versus NFTs which combined aspects like rarity and tier that obfuscated any underlying value. Their use was supercharged by platforms like pump.fun, which reduced the creation of new memecoins to just a couple clicks, setting off a frenzy of speculation and new user behaviors tied to token price appreciation. You can find a compilation of the more extreme attempts here.

Yet, amid this speculative chaos, a new asset has emerged which is engendering similar user behaviors to NFTs and memecoins: on-chain agents. These digital entities combine blockchain technology with artificial intelligence to deliver novel user experiences. Though most agents today are indistinguishable from memecoins, several on-chain agents have begun to differentiate themselves through utility.

The Rise of On-Chain Agents

Agents represent another asset class in crypto experimenting with new business models and monetization. From AI-generated podcasts to investment insights and anonymous communication, these virtual entities have already reshaped how much of crypto Twitter (X) interacts. The biggest on-chain agents have mindshare bigger than the biggest human crypto-native influencers, and make money similarly: by token-gating information and offering subscriptions. Their distinguishing features — utility-driven frameworks and fair-launch principles — should make agents a more investible asset class than memes. Seen through the lens of hold period, liquidity, and utility, the distinction is even more clear.

Because we suspect investors will hold agents longer term than memecoins, and they create liquidity for themselves through their business models, crypto-focused investors will find this asset class easier to back once the initial frenzy has cleared. Until the business models flourish however, picking agents to invest can be likened to throwing darts at a board.

Early Innovators in On-Chain Agents

The on-chain agent market remains nascent, with most projects still in development. While projects like Truth Terminal set off the frenzy by showing the world that agents could have mimic real people, newer projects have focused on utility. Trained on data from crypto Twitter, AIXBT delivers lightning-fast insights on token dynamics, rivaling the influence of major crypto personalities. Others like Luna have proliferated as entertainment agents, interacting with thousands of people through twitter and TikTok.

Having spent the last two weeks experimenting with many of these, here are five more that are worth playing with. It’s unclear whether any of these are valuable investment opportunities, only that they offer differentiated user experiences.

These projects illustrate the diversity and ingenuity of the on-chain agent ecosystem, laying the foundation for its expansion. Each offers a novel AI-powered user experience that anybody can experiment with. Over time, we suspect that continued engagement may even allow them to create moats. While unclear where these may come from today, Dunbar’s Number provides a helpful framework. It defines the cognitive limit on the number of meaningful social relationships humans can maintain, and is around 150. Agents that create value by maintaining a nearly infinite number of simultaneous relationships, like AIXBT, unlock opportunities beyond what the human brain can cognitively do.

The Big Picture

History doesn’t repeat but it rhymes is an adage you’ll see on the twitter feed of every degen that’s ever lost 90% on a trade, but also proves unfailingly true. At the outset of the fourth bull run of the last two decades, it’s hard to ignore the comparisons.

DeFi summer was set off by the realization that centralized fintech companies often act against their customers. Famously, when Robinhood stopped out retail traders in favor of the big guns in Citadel, these traders realized that big regulated central companies may not be acting in their best interests.

Interestingly, a very similar dynamic is afoot in AI. The biggest companies like ChatGPT have struck multi-year deals with companies like Apple, allowing them to ingest people’s personal iPhone data without much accountability. As such, the violent price swings on agents traded on-chain may be front running this latest rhyme. It’s unclear how this dynamic will play out however. Beyond the agents themselves, agentic frameworks like ai16z’s Eliza and the Virtuals platform may capture value more clearly. The latter is already the breakout performer of the last quarter price-wise: given the inherent uncertainty, investing in an index of agents makes sense. I suspect this is because while agents are inherently interesting, it’s unclear that their usefulness will compound and that the attention dedicated to them will be lasting.

There is an old story about the market craze in sardine trading in a period of relative food scarcity. The commodity traders bid them up and the price of a can of sardines soared. One day a buyer decided to treat himself to an expensive meal and actually opened a can and started eating. He immediately became ill and told the seller the sardines were no good. The seller said, “You don’t understand. These are not eating sardines, they are trading sardines.”

As scarcity returns to the market it’s worth remembering agents can be a trillion dollar asset class. But for now, save for a handful, they’re still sardines.