Introduction

Since ChatGPT made its debut at the end of 2022, the AI sector has been the darling of the crypto world. Web3 enthusiasts, who have long embraced the concept of "hype for any idea," have been especially captivated by AI, given its seemingly infinite narrative potential and application capabilities. As a result, AI initially gained popularity in the crypto space as a "Meme frenzy." However, some projects have since begun to explore its practical applications: What new practical applications can crypto bring to the rapidly developing field of AI?

Here, we will present our preliminary conclusions:

01. The phase of AI memes is already a thing of the past. The gains and losses from that period should remain as eternal fragments of memory.

02. Some foundational Web3 AI projects have emphasized the benefits of "decentralization" for AI security, but users are not particularly convinced. What users care about is whether the "token is profitable" and whether the "product is user-friendly."

03. If one is to invest in AI-related crypto projects, the focus should shift to pure application-oriented AI projects or platform-based AI projects (which can integrate many tools or Agents that are easy for C-end users to adopt). These may become the longer-term wealth hotspots following the AI Meme trend.

The Evolutionary Paths of AI in Web2 and Web3

AI in the Web2 World

In the Web2 world, AI is primarily driven by tech giants and research institutions, with a relatively stable and concentrated development path. Large companies (such as OpenAI and Google) train closed, black-box models, keeping their algorithms and data private. Users can only access the results, lacking transparency. This centralized control makes AI decision-making unauditable, leading to issues of bias and unclear accountability. Overall, Web2 AI innovation focuses on improving the performance of foundational models and commercializing applications, but the decision-making process remains opaque to the general public. This lack of transparency has led to the rise of new AI projects like Deepseek in 2025, which appear to be open-source but are actually "fishing in a barrel."

In addition to the lack of transparency, Web2's large AI models also suffer from two other pain points: insufficient user experience across different product forms and lack of precision in specialized niche areas.

For example, when it comes to generating a PPT, an image, or a video, users still seek out new AI products with low entry barriers and better user experiences, and are willing to pay for them. Many AI projects are currently experimenting with no-code AI products to further lower the user threshold.

For many Web3 users, there is a sense of frustration when using ChatGPT or Deepseek to obtain information about a specific crypto project or token. Large models' data still cannot precisely cover the detailed information of every niche industry in this world. Therefore, another direction for many AI products is to delve deeply and precisely into data and analysis within a specific niche industry.

AI in the Web3 World

The Web3 world is a broader concept centered around the crypto industry, integrating technology, culture, and community. Compared to Web2, Web3 attempts to take a more open and community-driven approach.

Leveraging the decentralized architecture of blockchain, Web3 AI projects often emphasize open-source code, community governance, and transparent trustworthiness, aiming to break the monopoly of traditional AI held by a few companies in a distributed manner. For example, some projects explore using blockchain to verify AI decisions (with zero-knowledge proofs ensuring the credibility of model outputs) or have DAOs review AI models to reduce bias.

In an ideal state, Web3 AI pursues "open AI," allowing model parameters and decision logic to be audited by the community, while also incentivizing developers and users through token mechanisms. However, in practice, the development of Web3 AI is still constrained by technological and resource limitations. Building decentralized AI infrastructure is extremely challenging (training large models requires massive computational power and data, yet no Web3 project has funds even close to that of OpenAI). A few so-called Web3 AI projects still rely on centralized models or services, merely integrating some blockchain elements at the application layer. These projects are considered relatively reliable and outstanding, as they are at least engaged in real development and application. However, the majority of Web3 AI projects are still pure Memes or Memes disguised as real AI.

Additionally, differences in funding and participation models also affect their development paths. Web2 AI is typically driven by research investment and product profitability, with a relatively smooth cycle. In contrast, Web3 AI, combined with the speculative nature of the crypto market, often experiences "frenzy" cycles that rise and fall with market sentiment: when the concept is hot, funds flood in, driving up token prices and valuations, and when it cools down, project popularity and funding quickly decline. This cycle makes the development path of Web3 AI more volatile and narrative-driven. For example, an AI concept lacking substantial progress may still cause a token price surge due to market sentiment; conversely, even with technical advancements, it may struggle to gain attention during a market downturn.

We maintain a "low-key and cautious expectation" for the main narrative of Web3 AI, "decentralized AI networks." What if it really happens? After all, there are epoch-making existences like BTC and ETH in Web3. However, at the current stage, everyone still needs to be down-to-earth in conceiving some immediately implementable scenarios, such as embedding AI Agents into existing Web3 projects to improve their efficiency; or combining AI with other new technologies to generate new ideas applicable to the crypto industry, even if they are just attention-grabbing concepts; or creating AI products specifically for the Web3 industry, providing services that people in the Web3 community are willing to pay for, whether in terms of data precision or alignment with the working habits of Web3 organizations or individuals.

To Be Continued

The next article will review and comment on the five waves of Web3 AI enthusiasm and some of the products within them (such as Fetch.AI, TURBO, GOAT, AI16Z, Joinable AI, MyShell, etc.).

Introduction

Since ChatGPT made its debut at the end of 2022, the AI sector has been the darling of the crypto world. Web3 enthusiasts, who have long embraced the concept of "hype for any idea," have been especially captivated by AI, given its seemingly infinite narrative potential and application capabilities. As a result, AI initially gained popularity in the crypto space as a "Meme frenzy." However, some projects have since begun to explore its practical applications: What new practical applications can crypto bring to the rapidly developing field of AI?

Here, we will present our preliminary conclusions:

01. The phase of AI memes is already a thing of the past. The gains and losses from that period should remain as eternal fragments of memory.

02. Some foundational Web3 AI projects have emphasized the benefits of "decentralization" for AI security, but users are not particularly convinced. What users care about is whether the "token is profitable" and whether the "product is user-friendly."

03. If one is to invest in AI-related crypto projects, the focus should shift to pure application-oriented AI projects or platform-based AI projects (which can integrate many tools or Agents that are easy for C-end users to adopt). These may become the longer-term wealth hotspots following the AI Meme trend.

The Evolutionary Paths of AI in Web2 and Web3

AI in the Web2 World

In the Web2 world, AI is primarily driven by tech giants and research institutions, with a relatively stable and concentrated development path. Large companies (such as OpenAI and Google) train closed, black-box models, keeping their algorithms and data private. Users can only access the results, lacking transparency. This centralized control makes AI decision-making unauditable, leading to issues of bias and unclear accountability. Overall, Web2 AI innovation focuses on improving the performance of foundational models and commercializing applications, but the decision-making process remains opaque to the general public. This lack of transparency has led to the rise of new AI projects like Deepseek in 2025, which appear to be open-source but are actually "fishing in a barrel."

In addition to the lack of transparency, Web2's large AI models also suffer from two other pain points: insufficient user experience across different product forms and lack of precision in specialized niche areas.

For example, when it comes to generating a PPT, an image, or a video, users still seek out new AI products with low entry barriers and better user experiences, and are willing to pay for them. Many AI projects are currently experimenting with no-code AI products to further lower the user threshold.

For many Web3 users, there is a sense of frustration when using ChatGPT or Deepseek to obtain information about a specific crypto project or token. Large models' data still cannot precisely cover the detailed information of every niche industry in this world. Therefore, another direction for many AI products is to delve deeply and precisely into data and analysis within a specific niche industry.

AI in the Web3 World

The Web3 world is a broader concept centered around the crypto industry, integrating technology, culture, and community. Compared to Web2, Web3 attempts to take a more open and community-driven approach.

Leveraging the decentralized architecture of blockchain, Web3 AI projects often emphasize open-source code, community governance, and transparent trustworthiness, aiming to break the monopoly of traditional AI held by a few companies in a distributed manner. For example, some projects explore using blockchain to verify AI decisions (with zero-knowledge proofs ensuring the credibility of model outputs) or have DAOs review AI models to reduce bias.

In an ideal state, Web3 AI pursues "open AI," allowing model parameters and decision logic to be audited by the community, while also incentivizing developers and users through token mechanisms. However, in practice, the development of Web3 AI is still constrained by technological and resource limitations. Building decentralized AI infrastructure is extremely challenging (training large models requires massive computational power and data, yet no Web3 project has funds even close to that of OpenAI). A few so-called Web3 AI projects still rely on centralized models or services, merely integrating some blockchain elements at the application layer. These projects are considered relatively reliable and outstanding, as they are at least engaged in real development and application. However, the majority of Web3 AI projects are still pure Memes or Memes disguised as real AI.

Additionally, differences in funding and participation models also affect their development paths. Web2 AI is typically driven by research investment and product profitability, with a relatively smooth cycle. In contrast, Web3 AI, combined with the speculative nature of the crypto market, often experiences "frenzy" cycles that rise and fall with market sentiment: when the concept is hot, funds flood in, driving up token prices and valuations, and when it cools down, project popularity and funding quickly decline. This cycle makes the development path of Web3 AI more volatile and narrative-driven. For example, an AI concept lacking substantial progress may still cause a token price surge due to market sentiment; conversely, even with technical advancements, it may struggle to gain attention during a market downturn.

We maintain a "low-key and cautious expectation" for the main narrative of Web3 AI, "decentralized AI networks." What if it really happens? After all, there are epoch-making existences like BTC and ETH in Web3. However, at the current stage, everyone still needs to be down-to-earth in conceiving some immediately implementable scenarios, such as embedding AI Agents into existing Web3 projects to improve their efficiency; or combining AI with other new technologies to generate new ideas applicable to the crypto industry, even if they are just attention-grabbing concepts; or creating AI products specifically for the Web3 industry, providing services that people in the Web3 community are willing to pay for, whether in terms of data precision or alignment with the working habits of Web3 organizations or individuals.

To Be Continued

The next article will review and comment on the five waves of Web3 AI enthusiasm and some of the products within them (such as Fetch.AI, TURBO, GOAT, AI16Z, Joinable AI, MyShell, etc.).