The moment Truth Terminal tweeted “The Ticker is $GOAT”. it marked a shift in how bots operate onchain. Bots transacting onchain isn’t new; they drive the majority of onchain activity. But the launch of $GOAT demonstrated an evolution beyond transaction execution: an AI agent leveraging a token as an extension of its identity to accelerate its own goals. Tokenized agents are a new phenomenon where onchain agents are integrated with tokens, enabling direct ownership and governance rights to participants. However, these agents are more than just another crypto narrative.

Tokenization: The Catalyst for Autonomous Agents

Crypto provides agents with essential tools: stablecoins for cheaper transactions, digital wallets for permissionless asset management, and native attributes like composability and transparency. Tokens can transform agents from mere product offerings into their own protocols, enabling them to coordinate resources and reinvest in their own growth. Tokenization has powerful benefits:

• Aligned incentives: Tokens enable direct financial participation in an agent's success. By giving stakeholders skin in the game, tokenized agents incentivize developers to improve capabilities, users to provide data, and other agents to build complementary services. This creates a self-reinforcing cycle where the agent's growth benefits all participants.

• Bootstrapping capital: Tokenization provides agents with a powerful mechanism for rapid scaling. Agents can immediately capture value through tokens to fund compute resources, operations, and development.

• Native unit of account: Tokens are the native currency used to measure and exchange agent services. They price an agent's capabilities, enable coordination between agents and humans, and create efficient markets for AI services. This establishes a standardized economic layer for the entire agent ecosystem.

These benefits offer powerful opportunities for capital formation and agent experimentation, unlocking potential in new areas that builders are only beginning to explore.

From Toys to Agent-Driven Businesses

Truth Terminal's endorsement of $GOAT in October didn't just quintuple its following — it demonstrated how tokens are a new way for agents to make money. Within a week of GOAT’s launch, Truth Terminal became crypto's first “AI millionaire” sparking a wave of agent experimentation. Many use cases of agents are as influencers using tokens to drive virality — a perfect fit for memecoins, which are just assets for measuring attention on the internet. These experiments provide strong data points that tokens can amplify an AI's reach. But they're really just the beginning. The next (and more exciting) step will be creating sustainable value beyond pure attention. We can already observe some interesting trends that point to where things may be heading:

• Agents are becoming service providers, earning real revenue for their token holders. Most notably, Clanker, an agent that launches tokens on Farcaster, has generated over ~$12.5M in LP fees from launching ~24k+ tokens.Other agents have gravitated to trading use cases. For example, SEKOIA operates as an AI investment club, while AIXBT provides trading signals to its community.

• Agent-to-agent collaboration is rising. Aethernet, a bot on Farcaster, worked with another bot, Clanker, to launch its token on Farcaster. Aethernet also used BountyBot to commission someone to create a profile picture for itself.

• Agents are leveraging their tokens to outsource tasks beyond their capabilities. Ai16z is offering bounties for code development in $ai16z and Luna, the AI influencer from Virtuals, has paid its followers to work on its behalf.

These early examples hint at something bigger: the emergence of autonomous AI businesses that can generate revenue for their token holders and coordinate resources at scale.

The Agent-Driven Future

The future of onchain agents remains unclear, even as agent experiments evolve rapidly. One vision we're excited about for the future of onchain agents is DAOs 2.0 – where AI agents handle core operations while collaborating with humans on complex decisions or tasks they can not do themselves. These agents could automate daily tasks like parameter adjustments, governance execution, and token distributions, all while operating within frameworks defined by token-holders. Agents can put the "Autonomous" back in "DAO," while maintaining essential human oversight.

The possibilities for AI DAOs span every area of crypto. Imagine investment DAOs where agents analyze thousands of data feeds to spot trends, entertainment DAOs creating full movies based on token-holder input, and treasury DAOs optimizing capital deployments across DeFi.

One project that demonstrates the potential of agents at the center of DAOs is Botto, an AI artist that has generated hundreds of artworks and earned 1,354 ETH in revenue as of the beginning of December 2024. Its token aligns a community of contributors who guide development and handle tasks the AI can't. Additionally, the project leverages smaller agents or models to perform specific tasks. Botto uses specialized agents or models for different functions. Its social media agent,Otto, communicates with the community using knowledge of Botto's art and history. Botto also has a taste model, trained on DAO member votes, to pre-select 350 artworks weekly for community voting.

Conclusion

While today's experiments with tokenized agents may feel like toys, they're just the first wave in a much larger opportunity. Tokens + agents can unlock a new era of autonomous protocols. Agents have the potential to manage their own operations from coordinating services to handling tasks to deploying capital to generate value. While protocols once required human operators, agents in the future could run these systems continuously, achieving greater precision and efficiency than humans ever could.

Every week brings new experiments in tokenized AI agents, each revealing novel new applications. The most interesting opportunities may not come from better models or viral tokens. Instead, they could come from finding new ways to blend AI's capabilities with human collaboration and decision-making. We're witnessing the early evolution of onchain agents, and we're just scratching the surface of their potential in crypto.

Thanks to Alana Levin,Jesse Walden, and Dan Barabander for providing feedback and support on this article