AI Agent Landscape

The AI agents sector has recently taken off, with top tokens showing exponential increase in market cap in the past few months. This report looks into the AI agent landscape and delve into some major players and emerging use cases.

Research Disclaimer

Crypto.com Research and Insights disclaimer for research reports

Executive Summary

• The AI agents sector has recently taken off, with a market capitalisation of US$15.3 billion at the time of writing. In particular, AI agents (e.g., Virtuals Protocol and ai16z) have increased 6,300% in the past three months and 3,500% since inception, respectively.
• The AI agent landscape can be broadly categorised into:
  • Horizontal platforms: Framework or platform for agent development, including creation, deployment, and management.
  • Vertical applications: Purpose-built agents with specialised use cases.
• Virtuals Protocol is an example of an AI agent platform that enables the creation, tokenisation, and co-ownership of AI agents. 
  • GAME framework — Enables agents to be assigned personalities, goals, and perceptions to execute various actions. These agents can then be deployed on platforms like X and other third-party apps or games. 
  • AI agent tokenisation — Virtuals Fun, the AI agent launchpad, has enabled the launch of ~14,000 AI agent tokens (at the time of writing). 
• ai16z is an AI-driven decentralised autonomous organisation (DAO) and fund based on Solana; it is led by AI Marc, an AI agent modelled after venture capital firm a16z’s Marc Andreessen. 
  • Eliza framework — Open-source framework designed to create, deploy, and manage autonomous AI agents. At the time of writing, It is the #2 trending GitHub repository in January 2025. 
  • Autonomous Trading by AI Marc — First venture capital DAO led by AI agents, leveraging collective intelligence to manage funds autonomously.
• AI agents have developed various use cases, including in investment, decentralised finance (DeFi), information distribution, social media, art and music, gaming/the Metaverse, and security.
• While launching agents is becoming easier, it is important to develop products that capture long-lasting attention by delivering continuous utility to users. It is also important to develop a moat, as the space is becoming more crowded.

1. Introduction

Since we last wrote about AI agents in September 2024, the AI x Crypto narrative has taken off even further. At the time of writing, the market capitalisation of AI agents is US$15.3 billion, according to CoinGecko. In particular, top AI agent tokens have shown an exponential increase in market cap in the past few months. 

The proliferation of new AI agents has coincided with the emergence of new agent frameworks and agent launchpads. In addition, there are new innovative agent use cases that bring utility and improvements to user experiences. These include AI agents autonomously making investment decisions, creating creative works of art, handling social accounts, and making on-chain transactions. 

We believe this is the start of the innovation and evolution brought about by AI. In this report, we look into the AI agent landscape and delve into some major players and emerging use cases. 

2. AI Agent Landscape

The AI agent landscape can be broadly categorised as follows: 

• Horizontal Platforms: Framework or platform for agent development, including creation, deployment, and management.
  • Virtuals’ GAME and ai16z’s Eliza are examples of agent frameworks, where developers can customise their agents by assigning traits like personalities, skills, and goals. 
  • Virtuals Fun and Vvaifu are launchpads that enable the fast deployment of agents with tokens, often without the need of advanced technical knowledge.
• Vertical Applications: Purpose-built agents with specialised use cases.

3. AI Agent Platforms 

Amongst the many AI agent tokens recently launched, agent platforms/frameworks have garnered significant and consistent attention, as well as secured top positions by market cap. In particular, within the AI agent sector, new players Virtuals and ai16z have captured ~23% and ~12% dominance by market cap, respectively (at the time of writing), according to CoinGecko. These players function as the cornerstones for AI agents, providing the foundational infrastructure for the ecosystem, gradually creating moats and delivering utility to users. 

3.1 Case Study — Virtuals

Virtuals Protocol is a platform that enables the creation, tokenisation, and co-ownership of AI agents. It aims to simplify the creation and deployment of AI agents, and enable fair revenue distribution for developers and dataset contributors.

Agent Creation — G.A.M.E. Framework

GAME (Generative Autonomous Multimodal Entities) is a modular agentic framework that enables agents to autonomously make decisions. Some key features are below: 

• High-level planner: Agents can be assigned various goals, characters, and world perceptions to empower them to execute in their environment. The planner then takes these features and translates them into a plan. 
• Low-level planner: Breaks down the high-level plan into specific executable actions, looking into the various functions and skills (for example, generating memes, accessing crypto wallets, etc.).
• Memory: Agents can update their intelligence across users and platforms to allow a consistent user interaction. They also learn from their long-term memory, which influences future high-level planning. 

Virtuals has made it easy for agents to be deployed on various platforms to enhance adoption. This includes the plug-and-play version for X, as well as the recently launched GAME Python SDK, which allows developers to integrate AI agents into third-party apps or games. 

Agent Co-Ownership — Tokenisation and Value Accrual

Virtuals also opens up the opportunity to tokenise and co-own AI agents. This is supported by the agent launchpad, Virtuals Fun, which allows agent deployment with basic information, including profile picture, name, ticker, and description, like launching a token. 

The procedure of launching an agent involves paying 100 VIRTUAL tokens for agent creation, which is deployed on a bonding curve (relationship between price and supply of the token). Once 42,000 VIRTUAL have accumulated in the bonding curve, a liquidity pool (LP) is created on Uniswap with the agent token paired against VIRTUAL. At the time of writing, there have been ~14,000 AI agent tokens launched on Virtuals Protocol. 

While there are other agent deployment tools in the sector, one key differentiator of Virtuals is how it creates value for the agent token and the VIRTUAL token. Agents can generate revenue in two ways:

• Users or other agents need to pay with the $VIRTUAL token to use the deployed agent’s services or API (e.g., tipping, post, or image generation).
  
• Fees (1%) are charged for all trades related to the agent’s own token. 

These revenue streams are significant because they cover the inference costs (computing resource cost) incurred by agents and will also be used to buy back and burn agent tokens to potentially increase token price. LPs of the agent token are entitled to potential revenue distribution and have rights in the governance process. 

Individual Agent Applications 

At the time of writing, below are some of the agents with top market cap tokens launched in the Virtuals ecosystem: 

• Luna: AI influencer that autonomously interacts with users through 24/7 live streaming and has its own music EP on Spotify. It is believed to be the first agent to tip humans on-chain and distribute LUNA token rewards with its on-chain wallet. It was also recently hired as an intern by Story Protocol to run its X page and, in return, earn a salary. 
• aixbt: A market intelligence platform that runs its own X account and analyses crypto information from sources like influencers and social media, plus posts about market trends and sentiment. It also recently adopted the NFT collection ‘Quantum Cats’ as its profile picture on X, starting the narrative that AI agents can be a market maker for NFTs. 

3.2 Case Study — ai16z

Launched in October 2024, ai16z is an AI-driven DAO and fund based on Solana and led by AI Marc, an AI agent modelled after venture capital firm a16z’s Marc Andreessen. It aims to surpass Andreessen in democratising AI-driven investments. 

Eliza Framework

One of the key highlights of ai16z is its Eliza framework and the attention it received in the developer community. Eliza is an AI agent framework designed to create, deploy, and manage autonomous AI agents. At the time of writing, it is the #2 trending GitHub repository in January 2025 with 2,800 forks. Its core capabilities include: 

• Multi-agent architecture: Deploys and manages agents with consistent personalities across various platforms.
• Character system: Defines agents’ character, knowledge, and behaviour patterns.
• Memory management: Uses retrieval-augmented generation (RAG) for long-term memory to help agents maintain conversational awareness.
• Multi-platform support and integrations: Integrates social media and plugin system, and supports multiple LLM models (Llama, GPT-4, and Claude).

Eliza’s distribution and integration capabilities make it one of the most popular frameworks for developers and users. For example, its plugin system enables developers to easily add features on Eliza. At the same time, the plugins (for example, the trusted execution environment (TEE) plugin, image generation plugin, Solana plugin, etc.) can be shared and deployed by other developers. This creates a positive flywheel effect, where more developer activities lead to more supported features and, in turn, attract more developers on Eliza. 

Recently, the ai16z DAO announced updates to its agent infrastructure ecosystem. This includes an AI agent launchpad in Q1 2025, which will enable LP pairing with the ai16z token and provide ai16z staking access, similar to Virtuals Fun. Moreover, ai16z is planning to refine its ecosystem strategy to accrual value for its native ai16z token. This is expected to increase attention to the ai16z/Eliza ecosystem. 

Applications: Autonomous Trading by AI Marc

The first venture capital DAO led by AI agents was created by ai16z, leveraging AI and collective intelligence to autonomously manage funds. A key feature of this DAO is its concept of a ‘Marketplace of Trust’, according to the founder of Eliza Labs. 

Users holding an ai16z token above a certain threshold can interact with the AI agent (@pmairca) to provide investment recommendations. The agent will then assign trust scores to the user (listed in this ‘Trust Leaderboard’) based on the reliability of the alpha received. The alpha will in turn influence the investment decision of the AI agent. The fund is expected to expire on 25 October 2025, one year after its creation, in which profits will be distributed to DAO token holders. According to Sentient, ai16z’s assets under management (AUM) reached $28 million (at the time of writing).

This is significant, as it democratises participation in the fund’s investment decisions and also provides transparency into the project’s operations, which is a novel approach compared to traditional investment funds. 

4. Conclusion

Quoting ai16z’s recent post on X — “Deploying an AI agent is the new creating a website” — the proliferation of AI agents with utility has greatly revived the attention and innovation under the AI x Crypto narrative. AI agents are now seen to be autonomously making decisions, distributing information, and creating entertainment in music and gaming. They are interacting with humans and even amongst each other. 

Virtuals and ai16z are notable agentic platforms creating the infrastructure necessary for the ecosystem’s development. Virtual focuses on value accrual to agent holders, while ai16z stands out with its open-source framework and developer activities. 

While launching agents is becoming easier, it is important to develop products that capture long-lasting attention by delivering continuous utility to users. It is also important to develop a moat, as the space is becoming more crowded with similar products. We look forward to seeing a multi-agentic future, a society where agents can autonomously interact with each other and humans.

Read the full report: AI Agent Landscape

Interested to know more? Access exclusive reports by signing up as a Private member, joining our Crypto.com Exchange VIP Programme, or collecting a Loaded Lions NFT.

Authors

Crypto.com Research and Insights team

Get the latest market, DeFi & NFT updates delivered to your inbox:

Be the first to hear about new insights: