Compiled by: Blockchain in Plain Language
Artificial intelligence is the main theme of the current cycle, and many firmly believe this field will exist long-term. On the other hand, there are many valid criticisms that most current intelligent agents are just chaotic entities, and we need another 3 to 5 years to see these technologies become truly meaningful.
As a preface, the combination of cryptocurrency and artificial intelligence encompasses many layers. The true technological potential of cryptocurrency and artificial intelligence lies mainly in guiding better reasoning processes through cryptoeconomics or providing decentralized computational access. The Delphi article is a great starting point to understand the entire tech stack.
Delphi Labs: https://x.com/delphi_labs/status/1834247706103160939
However, the focus of this article is on the current state of intelligent agents. While exciting innovations are happening in other layers of the tech stack, intelligent agents seem to have captured the attention of the mainstream cryptocurrency community. As this layer continues to evolve, here are six trends in cryptocurrency and AI intelligent agents to watch:
1. Framework + Launch Platform: Value accumulation is beginning to matter; frameworks may exist long-term.
Value accumulation is beginning to matter. Why is that? Stepping back, why would someone hold an asset? There are two reasons people trade assets with dollars:
They believe they can sell this asset to others for a higher price, benefiting from capturing a market narrative that can attract new buyers.
This asset will generate more cash flow over time.
Jez @izebel_eth mentioned in his article (Old Coin Bad, New Coin Gud) that the only thing that matters is liquidity. These two reasons reflect two bullish liquidity perspectives on assets:
New Holders
Token Consumption
However, in most cases, we have not seen anything resembling true token consumption or token value accumulation. Using tokenized intelligent agent terminals, such as AIXBT, feels more like staking rather than traditional value accumulation.
This is also why frameworks like Virtuals, AI16z, Zerebro, and Arc have recently become so popular. Virtuals has accumulated $60 million in protocol revenue. ai16z initially started as an investment DAO, but since announcing the upcoming launch platform and details on related token value accumulation, it has become one of the top protocols in the field.
Today, the market for frameworks and launch platforms has become exceptionally saturated, leveraging the influence of early success stories. There are many reasonable doubts about the roles of these launch platforms, especially when many intelligent agents themselves have no practical use. However, many frameworks (like Eliza V2 + launch platform, Zentients, Arc, and their handshake protocols) haven't even launched their main products yet. If they successfully attract developers and users, they may continue to lead the entire industry.
Reasons why frameworks may exist long-term: Regardless of whether intelligent agents remain purely "chaotic," the frameworks for launching agents will perform well because they still have product-market fit (PMF) in speculation. Frameworks and launch platforms allow users to own both factories and casinos. In many ways, Virtuals has simply replaced pump.fun in the Base ecosystem.
More optimistically, with ongoing technological advancements, more advanced intelligent agents may emerge from leading frameworks, accelerating development alongside open-source codebases like Eliza. Many of these launch platforms are also positioning themselves as a coordination layer for communication between collectives and agents, facilitating some form of value transfer through their tokens. For a deeper article on Eliza's value capture, see here:
Project Case: Currently, @virtuals_io, @ai16zdao, @0xzerebro, @arcdotfun are the major names, but the market for launch platforms is becoming increasingly saturated. The fastest innovators, the most scalable frameworks, and those offering the most unique features ("Does this framework build capabilities that others cannot?") are worth paying attention to.
2. DeFAI
DeFAI: The next wave of intelligent agents will focus on practicality and value accumulation, and DeFAI (DeFi x AI) may be the first area to achieve product-market fit (PMF). Most intelligent agents only possess meme tokens without actual functionality. To drive the development of the intelligent agent space, the next wave of agents must perform specific tasks for us. New opportunities will stem from genuine value accumulation and agents capable of taking action. I believe this will be a process of leapfrogging from level one agents to level three agents, expected to occur within the next year.
We first see the realization of this trend in decentralized finance (DeFi). We will see more terminals that allow users to express desired outcomes through natural language or voice, along with intelligent agents executing tasks behind the scenes. Existing wallets and protocols will also integrate intelligent agents to enhance user operational processes.
Project Case: @AIWayfinder @griffaindotcom @HeyAnonai @LimitusIntel @neur_sh
Long article about this field:
3. Consumer Level - Agent Proxies, Autonomous Worlds, and Gaming
Attention-based intelligent agents will develop more complex personalities and multimedia interactions. This could lead to the next wave of development in gaming and metaverse technologies.
One issue with existing intelligent agents is that they all start to blend into commoditized, beautified chatbots. While some agent projects will pivot to infrastructure (many already are doing so), others are beginning to think of useful products or applications, while others may continue down the path of attention tokens. However, the next generation of attention agents will be more powerful, with more complex personalities and higher interactivity. This could manifest through audiovisual representations or endowing agents with 3D personas and physical bodies.
@digimon_tech is a game similar to (Pokemon), where the creatures psychologically evolve based on interactions. @smolverse features Eliza, a character powered by a large language model that performs actions in a virtual world based on your instructions.
4. Agent Organization: The Return of DAOs
Decentralized agent organizations represent the next form of evolution for DAOs. Collectives or multi-agent systems are very exciting because they may be able to coordinate and execute more complex strategies, similar to operating a company. Heterogeneous collectives that include various types of specialized agents/models working in synergy may perform more efficiently than a single large model.
Although fully autonomous agents and collectives may still take some time to achieve, the next iteration of DAOs will likely include human interactions with collectives. This will reduce inefficiencies in bureaucracies and lower the costs and time of human execution. From a liquidity perspective, the next step will be for those agents that can generate income to develop into complete revenue-generating organizations.
Project Case: Agent Collective Infrastructure: Projects like @swarmnode and @joinFXN focus on providing infrastructure for multi-agent frameworks and coordination, but mainstream agent frameworks like @0xzerebro and @ai16zdao have also expressed intentions to build at this level.
DAO Launch Platforms: Most of the initial focus we see today is on investment-type DAOs. I believe @daosdotfun is the first major DAO launch platform that birthed ai16z. There are also emerging launch platforms like @daosdotworld, where AI-driven funds like @3berascapital_ are attracting increasing attention.
5. Verifiable Agents
Today's agents are developing towards greater autonomy, truly possessing their own liquidity: Currently, most agents still require a high level of human intervention. The next wave of agents will develop towards true autonomy, starting with managing their own funds. The product-market fit (PMF) of agents with cryptocurrency lies in that cryptocurrency provides financial infrastructure for real economic agents. However, most agents have an untouched treasury or are managed by human teams. To achieve true economic agency, agents must be able to manage their own funds autonomously. This will begin to change agent behaviors, as you can impose economic constraints on agents, requiring them to cover their own reasoning costs. This introduces Darwinism for agents; they must earn income to survive.
Project Case: @freysa_ai is one of the earliest agents able to control its own keys and has been performing exceptionally well (including catching Elon’s attention). They recently announced they are building a framework that allows agents to have verifiable autonomy in a trusted execution environment (TEE) with keys controlled by the agents.
@LitProtocol also has an agent framework that allows autonomous agents to transact on-chain and manage private keys through a storage and execution system. @galadriel_AI has launched an SDK called "Proof of Sentience" that enables developers to fully verify the existence of agents on-chain.
6. Summary
Rather than seeing a billion people on-chain, we are more likely to see a billion intelligent agents on-chain first.
Cryptocurrency's user experience is inherently unfriendly to human users, while intelligent agents are indifferent to this. We will start with the interaction between humans and agents, but the ultimate manifestation of crypto AI will be the interaction between agents, where clusters of autonomous agents will interact within their native economies and transact on-chain.
To realize economic agency, enabling it to incentivize behavior (pay service fees) and coordinate actual activities in the real world, intelligent agents need to have the capability to control and deploy capital. Cryptocurrency is home to these agents—blockchain infrastructure allows agents to participate in an unpermissioned financial system. Stablecoins and high-performance L1s are ideal vehicles for driving cost-effective, 24/7 global trading.
Beyond metadata and narratives, we have ample reason to be optimistic about the long-term prospects of on-chain agent economies. Perhaps many practical use cases, DAOs, and revenue-generating agents are closer than we think.
