The Crypto AI Agent Market: Beyond Hype and Memes
In this deep dive, we’ll explore the current state of crypto and AI, the Agentic Economy, market insights, and key challenges and opportunities.
AI agents are shaking up the crypto world. While much of the current landscape is flooded with hype and low-quality projects, the potential for innovation is enormous. We are in the early innings of what could be a revolutionary shift. The question isn’t just about separating noise from substance—it’s about identifying and leveraging the real opportunities driving this revolution.
As it stands, 99% of AI agents and adjacent tech in crypto are low-effort projects, scams, or gimmicky chatbot memes tied to a coin. But the remaining 1% is where innovation is taking root, and within that, 0.1% could prove to be truly game-changing.
In this deep dive, we’ll explore:
The origins of crypto AI agents
Market analysis of crypto and AI
The emergence of the Agentic Economy
Composability: Crypto AI agents working together
The impact on NFTs, art and culture
Challenges, opportunities, and the road ahead
Crypto AI Agents: How We Got Here
Thanks to the nature of crypto, where decentralized access can create industries seemingly out of nowhere due to the ease of capital formation, AI has found a natural home on the blockchain, where its potential aligns with crypto’s decentralized ethos and rapid innovation. I believe there are two main reasons for this, the first being access to capital.
On the blockchain, nearly everything becomes financialized. This has merged AI agents with cryptocurrencies, creating new opportunities—but also driving speculation around project success. This is nothing new to crypto. Beyond fungible tokens, we saw this phenomenon with NFTs starting in 2021. This time, however, it’s happening with AI agents and AI adjacent technology.
The origins of this trend date back to late 2024, when an autonomous agent called Terminal of Truths gained popularity on X. This agent, trained on a less restricted large language model (mainly fueled by Reddit), began promoting and pumping a coin called GOAT. Over time, GOAT surpassed $1 billion in market capitalization, sparking a new wave of funding mechanisms for projects—but also a surge in speculation, scams, and other problematic behaviors common in crypto.
Like it or not, this mechanism has created a massive market where founders and developers of AI agents or adjacent technologies are rewarded either by the rising value of their tokens or by the transactions generated by their projects. This allows them to fund their startups without relying on venture capital or private investment. In my opinion, this is a positive development for talented teams.
The second reason, which I believe we haven’t fully grasped yet, relates to the ease of transacting on the blockchain compared to traditional systems. For humans, creating a crypto wallet, buying crypto, interacting with smart contracts, and sending or receiving assets can be quite challenging. For agents, however, these processes are tailor-made. It’s much easier for agents to transact and operate in Web3 (and even parts of Web2) than in the traditional banking system. While buying and selling on platforms like Amazon remains difficult for agents due to credit card checks and being designed for humans, purchasing or selling an NFTs, or transacting with crypto is comparatively simple.
Now, the most important point:
Just because something is labeled as an "AI agent" doesn’t mean it has inherent value—just like NFTs, value comes from utility and meaningful use cases. Unfortunately, AI jargon and complexity make it easier for bad actors to disguise low-quality projects as innovation.
The majority of crypto AI agents today fall into one of two categories:
Chatbots: Generic, repetitive, and often tied to a token with little real utility.
Entertainment: Projects that focus on hype rather than solving real-world problems.
On the bright side, some entertainment and chatbot projects are developing strong brands and communities, which means they can continue to improve and spin off new agents, use cases, and capabilities. The key question is how much of their current value is genuine versus being driven by hype.
In addition to these two dominant use cases, there are new categories emerging, which I will refer to: AI agent-adjacent technology.
AI Launchers
AI Frameworks
AI Infrastructure
Decentralized Finance AI (DeFAI)
Market Analysis: The State of Crypto AI Agents
If you’ve been paying attention, it’s no secret that most AI activity in crypto is happening on two blockchains: Solana and Base (an Ethereum L2).
Let’s compare both blockchains and explore the different categories in the AI market.
For this analysis, we’ll use Cookie.fun, a platform that tracks the crypto AI market. Cookie divides this market into three main categories:
AI Agents: Projects with tokens tied to their ecosystems.
AI Infrastructure: Tools and frameworks that enable the development of AI agents.
DeFAI: Decentralized Financial AI, which encapsulates AI tools for DeFi.
Total Crypto x AI Market Size: $18.9 billion
At the time of this writing, the Crypto AI Market Cap is $18.9B. It reached an all-time high of $20B before recently dropping to $11.7B (almost 50% down from all time highs). Remember, this is the total market, including the three categories listed above.
Base Share: $6.81B
Solana Share: $9.71B
Agents Market Size: $9.39B
Base Share: $2.81B
Solana Share: $6.1B
The crypto AI agents dominates the market with over 50% of share. This isn’t surprising, as the most popular terms have been AIXBT, GOAT, Zerebro, Luna, which are mostly AI agents.
Infrastructure Market Size: $7.93B
Base Share: $4.01B
Solana Share: $2.81B
Base’s market cap in the infrastructure category is nearly double that of Solana. While this doesn’t necessarily mean Base has better infrastructure, it does indicate that the market places a higher value on Base’s ecosystem. This could be due to the number of projects, higher quality, or simply better positioning. Regardless, strong infrastructure often leads to more projects, better tools, and an overall healthier ecosystem.
DeFAI Market Size: $2.55B
Base Share: $1.05B
Solana Share: $1.38B
Although DeFAI is a term that has been recently coined, it already exhibits close to $2.5B market size. In this category we can find agents and tools dedicated to push decentralize finance forward with the use of AI. Among the use cases that we can already find are ways to invest, place orders, take loans and navigate the DeFi space smartly.
Speculative Outlook:
In the coming months, I foresee a significant shift in market dynamics. Currently, entertainment agents make up just over 50% of the market. However, I predict their share will decline, with infrastructure emerging as the dominant category, capturing approximately 50% of the market share. Meanwhile, DeFAI is likely to grow to 30%-40%.
This projection doesn’t imply that the agent market won’t grow—it likely will. However, the "entertainment" segment faces greater uncertainty, which creates opportunities for utility-driven categories to take the lead.
As we’ve recently witnessed the rise of DeFAI, I anticipate the emergence of new, theme-specific categories in the future. For instance, a category centered on art and culture could develop, leveraging AI's potential to transform creative industries. I can see this category involving all the aspects of media, art, and culture, including creation, curation, preservation, collection and trading.
Of course, this is speculative, but if you keep reading, I have some interesting examples of how crypto agents are starting to shape this area.
The Agentic Economy: A New Paradigm
The rise of AI agents marks the beginning of what some have labeled the Agentic Economy. This new paradigm centers around autonomous agents that can:
Execute transactions onchain with ease
Launch and interact with other agents, digital objects (NFTs), and applications
Make autonomous or semi-autonomous decisions based on a given goal
These agents don’t just simplify blockchain interactions for humans; they also pave the way for autonomous systems to transact and interact at unprecedented speeds and reach. This could lead to an explosion of onchain activity, fundamentally altering how we use blockchain technology.
Let’s review some of these use cases and agent-to-agent cooperation that is already happening.
Composability: Crypto AI Agents Working Together
These examples show how agents can cooperate and make the entire ecosystem stronger by enhancing each other’s capabilities.
Take Clanker, for instance, an AI agent that makes it simple to launch tokens. You don’t need to know complex coding—just send a message through the Farcaster feed, specifying what coin you want to create and information such as an image, a ticker, and tag Clanker. Bracket Game, a sports betting and gaming platform used Clanker to launch a token for its own AI agent, Bracky.
Bracky is designed to make Bracket Game’s platform more engaging. It serves as a token for players but also provides helpful information, like game stats and betting tips. It’s not just about one agent, though—Bracky works even better when paired with Bankr, another AI agent that helps users buy and sell cryptocoins directly through Farcaster or X. With Bankr, you can specify what and how much you want to buy or sell, and it does it for you (your wallet should be connected in advance). In the future, features like limit orders will be available.
What’s truly exciting is how these agents can work together. Imagine this: Bracky identifies a great betting opportunity and then uses Bankr to execute the transaction. Or, Bankr handles a complex trade while another agent gathers data to guide the decision. This kind of teamwork between agents creates a powerful, scalable system that benefits everyone involved.
By working together, these agents aren’t just solving individual problems—they’re building an ecosystem where each component strengthens the others. This is, a controlled and early scenario of the true Agentic Economy.
Adjacent Crypto AI Technologies and Infrastructure
As we have already mentioned, the long-term value of crypto AI agents likely lies in non-entertainment use cases. Infrastructure, tools, and additional projects play a critical role here—think of those as the shovels.
“Don’t dig for gold, sell shovels.”
There are two more elements that have risen in popularity when it comes to crypto AI-adjacent technology: AI Frameworks and Launchers.
Crypto AI Agent Frameworks
AI frameworks enable developers to easily create AI agents by simplifying the process and reducing the time required to launch. They also standardize different components, making them reusable and interoperable. Three frameworks stand out at the moment, though more are likely to emerge.
These frameworks primarily focus on messaging, with the main use case being to create an agent personality and launch a chatbot on X. While this is a logical first step, further improvements are needed to enhance their ability to transact on the blockchain.
Examples include Eliza by AI16z, Zerepy by Zerebro, G.A.M.E. by Virtuals, Rig Framework by Arc. It’s worth noting that Virtuals also functions as a launcher, and Zerebro itself is an agent.
Crypto AI Token Launchers
Virtuals stands as the first mover in this space, with a market cap of +$3 billion. It initially launched as a platform for entertainment-focused AI agents, with its first streaming agent, Luna. Over time, Virtuals has begun to expand into new categories, though it remains to be seen whether robust agents and projects will be developed.
Current Challenges:
High transaction fees and limited integration with decentralized exchanges (DEXs) and apps.
Poor user experience.
A proliferation of low-effort projects and rug pulls.
Despite these challenges, Virtuals shows promise. However, the space is so new that maintaining leadership is difficult, especially without a clear roadmap for product improvement…. Because, it isn’t clear what’s next or how will everything play out. In addition to Virtuals, numerous other launchers have entered the market, including Loomlay, Tokito, Creator.Bid, Hat, and many more.