Major institutions predict that BTC will break through its previous high this year. AI Agent, DeSci, RWA, DePin, etc. are all seen as tracks worth paying attention to in 2025. 2025 is worth looking forward to! At the beginning of the new year, BlockBeats invited crypto KOLs to talk about the trends and opportunities worth paying attention to in 2025.
BlockBeats: Please introduce yourselves first.
Anthony Z: Hello everyone, I am Anthony, the head of The Graph Foundation in Asia Pacific. I have been working at The Graph for about 5 years. I have been in this industry since 2018, and I have about 7 years of experience, so I can be considered an "old leek". I mainly do non-technical work in Web3 data infrastructure, and also face the technical community. In the past, we rarely talked about the market with friends in the circle, but I want to try it this year. Recently, I also have some content about AI that I want to share with you.
Sanzhi: I am Sanzhi, from AmberAC, which is the accelerator of Amber Group. We mainly focus on some projects in the AI and Crypto tracks. If Anthony calls himself an "old leek", then I am a "new leek". Before, I mainly worked in the developer community and was close to developers. I started to get involved in market transactions in March this year, and recently I have also been studying AI Agent and Meme coin.
CM: I’m Chen Mo, and I mainly do research and project investment related to DeFi. Recently, I’ve also been paying attention to AI agent related areas, and I’m also in the learning stage. I’m very happy to accept the invitation and communicate with you all.
EO: Hello everyone, I am the investment manager of Future Money Group. We mainly invest in DeFi and AI. We started investing in DeFi in 2020, and we have invested in several AI projects this year. We are currently optimistic about the AI agent track, but due to AI computing power and energy issues, the infrastructure is not yet perfect, resulting in slow development of AI agents. Despite this, we are still optimistic about this track and believe that there will be many excellent projects in the future.
BlockBeats: Please share with us the best deal you made in 2024, or the most profound lesson you learned.
Anthony Z: I rarely trade. This year, when the price of Bitcoin rose to more than $60,000, I added some positions. When the price of Ethereum was less than $3,000, I also bought some. I mainly hold. Some of my friends used to be resistant to cryptocurrencies, and were not even very interested in them. They mainly speculated on US stocks. But this year, some new friends said they started to speculate on ADA. I didn't dare to touch ADA before, but seeing that it is indeed rising, I think this may be because these newcomers have become new price setters. The cognition of us old leeks may no longer be applicable because we are a little outdated.
Sanzhi: I remember that I didn’t know much about tokens in January or February. I heard that Pepe was good, and a friend of mine also said that Pepe was good. Although I didn’t know much, I bought Pepe because both of them recommended it. I didn’t buy much, about a few hundred U, but later the profit tripled. At that time, I only dared to buy 100 U of all tokens.
Recently, because I need to study the hot spots and trends of the market, I started to operate on my own. I bought SOLANA before, and after the purchase, I went to Pump.fun to directly order a Meme coin. Although I am not familiar with these quick-in and quick-out operations, if I am optimistic about a project, I will do some research. Now this Meme coin has probably doubled or tripled. At its lowest point, it was only $0.018 a few days ago, and now it is about $0.1, which is almost 10 times higher. At $0.02, I frantically recommended to my friends that I entered the market at a price of 0.05, but later I was stuck. I think it's quite interesting.
Throughout the trading process, I did not do very well. I prefer to hold the token for a long time because I do not have time to watch the K-line all the time. If the token falls after I buy it, I will add to my position, and if it rises, I will hold it first. But I have a psychological expectation, for example, I think it will rise 10 times, so if the token does not meet my expectations, I will not sell it. If it returns to zero, I will accept it. I entered the market at $0.05, but when it fell to $0.02, I only added 20U, which was not fully implemented according to the strategy. However, although the price has been falling, I did not sell it. Now it has risen back, one or two times higher than before, but I still did not sell it and continued to hold it for a long time. I tend to be conservative. I must verify that my strategy is correct. I will buy it after doing a lot of research, and I will also formulate a trading strategy and execute it, keeping the mentality that it doesn't matter if it returns to zero.
CM: In recent years, I have paid more attention to the DeFi track, especially in the last few months of 2024, DeFi has regained some market attention. Some top assets, such as AAVE and LINK, have also performed well, and these are my long-term holdings. In the middle or late stages of the bull market, they performed well, and this performance was related to some policy factors, such as Trump's team or project indirectly holding some DeFi-related tokens.
I have studied the DeFi field, but I am not very good at other tracks. I always think that DeFi is one of the few tracks in the cryptocurrency circle that can change the market structure. Before the emergence of DEX or lending platforms, users did not have many choices and could only trade and borrow on centralized exchanges. Now, many excellent projects have appeared on the chain, and there are more choices. For example, Hyperliquid, which has been very popular recently, has also received a lot of attention in the derivatives track. This is still a trend, so I will continue to pay attention to this track for a long time. Correspondingly, for hot tracks like Meme, I didn't catch many opportunities. Part of the reason is path dependence and limited energy. For the AI track, I bought the top assets, and I didn't catch other small projects. I have some of the top tokens such as AI16Z, VIRTUAL and AIXBT. I mainly rely on some previous experience.
EO: After experiencing two waves of market fluctuations in 2024, I have some painful experiences to share with you. During the two fluctuations in April and December, one of my important feelings was that you must choose strong currencies for investment. Although weak currencies are cheap, they are actually not suitable for layout. This kind of currency is strong for one or two days, and will soon weaken again. The overall performance will not be very good, and the K-line chart will fluctuate greatly. Especially when the market is bad, this kind of currency will fall more. During the market fluctuations in November, I opened two orders for AAVE and DYDX. At that time, I could clearly feel that the bottom of AAVE was very strong, and it would not look back every time it rose; while DYDX was relatively weak. Although I have feelings for DYDX, judging from the periodic table, the performance of weak currencies is indeed not very good.
In addition, some once strong currencies also need attention, such as PNUTS. I did not seize the opportunity when it was rising, and when it fell to about 1.2 US dollars, I buy the dips the buy the dips again, but was slapped twice, and finally gave up this coin. Recently, the price of PNUTS has rebounded to 0.6 US dollars. You need to be vigilant about the relatively weaker coins in the near future. In my experience, if this coin absorbs more at the bottom, the price will rise faster. So I still recommend that you buy strong currencies in the bull market. If this coin is not strong in the bull market, there is basically no hope.
Another experience is that the market usually gives us some buffer time when it goes up or down. For example, since the flash crash on December 9, the subsequent performance of the market has not been strong, and many currencies have not performed particularly well. Many people have begun to liquidate their positions and sell their chips, while large institutions or large investors usually do not take a large number of orders at this point in this case, because this is not the stage of the main upward wave. At this stage, it is recommended that everyone must be vigilant about risks and set a stop loss line to avoid some risks, otherwise it will be difficult to stop losses in time during a sharp drop.
Anthony Z: I would like to ask Professor Chen Mo some questions. After the bull market, the DeFi track can be said to be the most mature. Hyperliquild is quite unique. It will gradually develop Layer1 from DEX. What do you think of this project?
CM: My personal opinion is that Hyperliquild actually seized the right time, right place, and right people. Many people think that it is not much different from DYDX in terms of project mechanism. DYDX had good data during the early airdrop, but the final effect was not good. The derivatives track has always been considered to be gradually falsified. Many people believe that people who open contracts on centralized exchanges will not come to the chain. First, the learning cost on the chain is high; second, people who speculate on contracts and players on the chain are two completely different types of people. It is similar to people who played NFT in the previous cycle not knowing what tokens are.
I don’t think Hyperliquild can attract users from centralized exchanges, but some users may come. But the main force is still the existing users on the chain. These people are already accustomed to interacting on the chain, and Hyperliquild just provides them with a new place.
Hyperliquild did a better job than many projects in early market operations and community building. Although there is FUD in the market, good products are controversial. Many people are skeptical about whether Hyperliquild can continue to develop, because the project's tokens have not been fully distributed, so people may increase trading volume due to the expected tokens. There is a very clear indicator in the future derivatives track. If the trading volume does not decrease and the number of users does not decrease after getting rid of token incentives, then this project will most likely continue.
DYDX's sustainability has deteriorated. Once the token rewards are reduced, users will find it unprofitable, and its sustainability will weaken. You can observe whether the project is sustainable as the market develops after the token expectations of the project end. If it is sustainable, some users will stay on it. Many people still like on-chain transactions. Even centralized exchanges have some mainstream currencies, and the Gas is lower. Hyperliquild is promising, but it cannot be said that it will definitely succeed. Once it succeeds, it will also be a phenomenal project. Many people say that Hyperliquild is not decentralized enough, but decentralization is gradually being achieved, so this is not a big problem. In general, Hyperliquild is a project worth paying attention to. It has done a good job in grasping market opportunities and timing, but it is too early to say that it is the end.
BlockBeats: AI Agent is the direction everyone has been aiming at. I would like to ask you how you view the current development of AI and the core focus areas in future development?
Anthony Z: I think the earliest AI narrative can be traced back to ChatGPT released by OpenAI. After that, many AI-related Crypto projects emerged one after another. At that time, those projects mainly talked about some grand narratives, such as decentralized computing power and computing power networks, which are relatively far from reality. Recently, due to the development of the overall AI industry, the integration of AI and Web3 is inevitable. Take AIXBT as an example. It combines some macro-currency speculation thinking with data on CoinMarketCap and discussions on Twitter to formulate strategies. But they have not yet combined specific on-chain data and transaction information. I hope this is the next step in development. We at The Graph mainly focus on infrastructure construction in this area. In general, most of the current AI projects are showing some practical use cases, and they should be developed in a more detailed manner in the future.
EO: We actually value the AI track very much. We hope that AI+Crypto will develop in the direction of automation in the future. However, AI currently still needs a lot of time and financial support in terms of infrastructure. The infrastructure cannot be greatly improved in a short period of time. There are some AI agent projects such as automatic tweeting, which are far from everyone's expectations and imagination of AI agents. These functions can also be achieved through scripts, which are not AI agents. I think it is difficult to see the emergence of real AI agents in the short term, and the cooperation of infrastructure is still needed.
There are two tracks that I am optimistic about recently. For example, AI and DeSci tracks, these two tracks usually raise funds through issuing coins, and then conduct research and product iteration after the funds are in place. The track may still be in a bubble stage. Everyone is enjoying the dividends brought by new technologies, but it is uncertain how long the bubble stage can last. In the future, I will pay more attention to the underlying technology and infrastructure of DeFi, such as privacy technology. In terms of the application layer, it depends on whether the Web2 track can develop excellent AI agent projects.
Sanzhi: I just joined AmberAC, so I share some insights from different perspectives. Earlier this year, I started an AI Crypto community. We divide the AI track into several levels: from bottom to top, there is infrastructure, including computing power layer and reasoning training. Then there is data collection, including processing, labeling, cleaning, etc. Then there is the model layer, and then the application layer. At present, many AI Agents actually belong to the application layer. In the AI track, there are several projects worth paying attention to. For example, Bittensor has innovated in mechanism, and its token economic design is quite clever. In addition, Sahara AI raised about 43 million US dollars in the first half of the year. MyShell AI belongs to the application layer, like an Agent market platform. Vana belongs to the data layer, focusing on data privacy protection, involving data and computing. In March and April last year, the wind of AI blew to the Crypto field. In March and April this year, AI began to develop from various levels such as data and application. But the overall shine should start from AI Meme.
Whether it is infrastructure, models or data, if AI wants to attract a large number of users at once, it is still limited by technology. It is better to directly extract the tokenized part. Therefore, AI Agent has quickly become popular with the help of Meme. I think this is a very good direction at present. After the bull market comes, the technical part is certainly important, but in the bull market, everyone pays more attention to the marketing aspect. People will be more FOMO. At this time, it is difficult to attract everyone's attention by talking about some technical content. If you do marketing cleverly, such as using Meme to attract attention, do more marketing during the bull market, and do more technology when the market is cold, it may be more effective. I guess AI Agent will be popular again in the future. At present, it is only in the text stage, and it may be expanded to multimodal in the future.
OnChain Data will also have good potential. For example, if an AI Agent interacts with us on Twitter, we tag an address and it can immediately tell us the dynamics of the address on the chain. Another one is DataChain. You can feed some native data to the AI Agent. When we interact with these Agents in large quantities, a large amount of data will be generated. This data is relatively real because it is the result of our interaction. This part of data is a valuable resource in the subsequent training of AI. Therefore, I think Data Chain still has a breakthrough point.
CM: I am more from the perspective of Crypto users and market demand. At present, I think AI Agent is still in the stage of concept hype. Everyone knows that the current technology is not mature yet. Compared with traditional AI, the existing AI Agent does not have many substantial results. In the cryptocurrency circle, the most important thing is imagination, and the same is true for early DeFi. Although DeFi is now one of the few tracks with good landing, it was also driven by imagination in the early days. People went through various stages of mining in the early days, stepped on countless pits in the middle, but there were also many opportunities to get rich. This is a very tortuous process, but the premise is that this track requires enough imagination. Including NFT, it is also because it is full of imagination and has a strong driving force that it can generate a high level of attention.
AI Agent also has a high ceiling in the traditional AI field. Even if the current technology is not mature, you can imagine the future development potential. Even in the field of Crypto, we are aware of the limitations of current technology, but once traditional AI technology has a breakthrough, the prospects of AI Agent will be very broad. It is unlikely to expect to accelerate the evolution of AI through various means in the field of Crypto. The main battlefield of AI is still in Web2.
The Crypto field has a significant advantage, which is that it can amplify any effect, especially some good effects or effects that can make money. For some independent developers, they are fully capable of moving the mature technology of traditional AI to the blockchain, and the cost of doing so is relatively low. I think there will be many bottom-up opportunities in the field of AI Agent in the future, not necessarily all formal projects led by VC.
At this stage, there will be many opportunities for developers and entrepreneurs to enter this wave, which is very important. In the past, there was a stage where everyone was FUDing VC coins because it was difficult for many ordinary entrepreneurs to participate, and the market was already fierce to a certain extent. But for AI Agent, there are opportunities in the future. Therefore, I think it will have a greater influence in the user circle than other tracks. This is my superficial view. As for what level the technology can reach, or whether it can change something, from my point of view, we can only observe and wait.
BlockBeats: The existing AI models, whether Web2 or Web3, have already trained all the free high-quality content on the Internet. If you want to continue to develop high-quality models, you may need some paid data support. What will be the competitive landscape in the field of AI data acquisition and service agreements in the future?
Anthony Z: We have made many attempts this year. The Graph has been working on AI for a long time. The training of AI large language models requires a lot of data, and The Graph is doing data infrastructure. In March of this year, we released a white paper introducing our progress in AI. We have a team specializing in AI. I met this team two years ago, and they proposed to use The Graph data to make a tool for querying on-chain data in natural language. For example, you can ask it about the transactions of Vitalik's account in the past year and a half, and then it will convert this question into The Graph's query language, retrieve real-time data from the chain, and give answers and data lists. In March last year, the team successfully developed this tool.
In order to let you better understand The Graph, I will briefly introduce it. Many people asked me what is the difference between The Graph and Chainlink. Chainlink transfers off-chain data to the chain, such as providing the exchange's token price to the on-chain Dex. The Graph, on the other hand, takes the on-chain data down for transaction data analysis or front-end transaction history display. It is a developer tool.
In the traditional Internet, Google is responsible for indexing and filtering Internet data and providing data services to other companies or independent developers. The Graph has more than 100 Graph nodes, which provide indexing services for on-chain data. Generally, projects such as Aave, Uniswap, and SushiSwap call The Graph's API interface. When you trade on Uniswap, the displayed transaction pair prices and transaction history data are obtained through The Graph. When you trade, you are also calling The Graph's API interface. When these interfaces are called, the data is actually transmitted to The Graph's network. The project party needs to pay a query fee when calling the data. Therefore, The Graph itself is a decentralized data network infrastructure and data market with a decentralized data network architecture.
To run and train AI models, you only need to add another system. We have been developing this system for the past year and released a product called Agentc. The front end of Agentc is similar to ChatGPT. You chat with it and it can call data to give you feedback. The AIXBT I just mentioned is not actually connected to the on-chain data, but if combined with The Graph, it can obtain real-time transaction data and analyze it. For example, you can monitor the operations of some large addresses, or you can monitor the transactions of Meme coins. Through these data analysis, the transaction suggestions provided by AIXBT may be more reliable.
What we are doing now is integrating Agent building, running and execution on The Graph. We have been working on this since we released Agentc, but we found that the accuracy of the currently trained LLM model is only over 63%, which is not enough. Therefore, we are developing a new model of Knowledge Graphs & LLMs, and are also creating a new transaction data standard GRC20. ERC20 is the standard for issuing coins on Ethereum and a standard for value exchange, while GRC20 is a standard for information exchange, including the relationship between data. In general, our attempts can use on-chain data and AI for simple combination, but for complex use cases, there are more and more errors. We can not only obtain on-chain data, but also combine and mark the relationship between each data. This is what we will do next.
EO: I think AI data is actually quite critical. Theoretically, the advantage of Crypto is that it can get some data that Web2 cannot obtain through some token incentives, and many project parties are also doing this. At present, most of the useful data is controlled by large companies. I think the project party needs to think about how to obtain data from C-end users, and this data must be useful. Secondly, users need to voluntarily contribute this data. Finally, how the project party uses the contributed data is very important. If any point is missing, you may not know how to use the data, or how to transmit the data to the downstream. If the middleware can process and package the data and pass it to the downstream, it can greatly improve the efficiency of downstream data use. In addition, users are very concerned about data leakage, so encryption and privacy protection are also tracks worthy of attention. I am more concerned about the field of AI plus FHE, and I am also thinking about how to effectively avoid user data leakage.
Sanzhi: I think the projects that run well in the field of Crypto and Web3 must be innovative around asset issuance and mobilize synergy. The popularity of AI agent is because it is related to asset issuance. In the past, a token was generated by a smart contract, but now an AI agent can be bound to generate a token. In addition, data is also very important. In the past, we did not have enough awareness of our data sovereignty, and our data has never been directly monetized. At that time, the logic of data monetization was to sell the data to the application party, and then the application party sold it to the advertiser. There were many links in the middle and no direct monetization. But now the way of data monetization has changed. There are mainly two new ways: one is to monetize with data feeding models, and the other is to issue data as assets. For example, the data of a cancer patient is a data package, which can be issued as an asset and has great potential.
When trading, in addition to looking at information on Twitter, I also check the transaction data in GMGN. If there is an agent that can integrate this information and provide multiple perspectives and objective data, such an agent is very valuable. If there is an AI agent that can help me process all the on-chain data and provide some intuitive insights, I am willing to pay for such a product because it can save a lot of time. There is also a lot of room for imagination to turn such a product into asset issuance.
We may not pay much attention to privacy protection, but large companies pay much attention to it because there are many privacy protection laws. With the increasingly standardized privacy protection laws, data sovereignty, data privacy and data assetization have great room for imagination. Decentralization and ownership are interrelated. If the synergy and ownership-related parts can be combined, there may be great development potential. As for hardcore AI projects, we need to observe their development further.
CM: My point of view is similar to that of the previous guests, and I mainly focus on the issue of data quality. Currently, a lot of data is mostly in the hands of large companies, and the acquisition of user-side data can be achieved through token incentives. Many projects are also developing this model, and I think it still has great potential. In the past, everyone was quite disgusted with the play-to-earn model because its inherent hematopoietic value was relatively weak. But it is still valuable for AI data annotation, and this part of the data can be converted into AI products for profit. Token incentives may be effective in AI data acquisition, but good mechanisms are still needed in each link.
BlockBeats: Finally, let’s talk about the main tracks that you will focus on in 2025, and what opportunities do you think are currently available in the DeFi track? What specific operational plans do you have for 2025?
Anthony Z: This year I will pay more attention to the RWA track, because it is a policy-supported area that can bring in new money. Although RWA is an old concept, it has been supported by the Trump administration and related interest groups, which will open up a lot of liquidity. I will focus on the leading projects in the track. If you are in Canada or Asia and want to invest in US assets, but there is no suitable channel. If assets can be tokenized on the chain and there is relevant policy support, this will bring many opportunities. This innovation is also beneficial to the United States itself and will bring new channels and opportunities to the international financial market.
CM: Regarding DeFi, I roughly divide it into two directions. First, from the perspective of policy trends, I pay attention to RWA-related content. Many traditional institutions, such as BlackRock, are also issuing tokenized funds on the chain. There are more operational and tradable scenarios for such assets, and they may be more scalable than we imagined. So RWA is not just about moving off-chain assets to the chain, but has richer application scenarios. We can call it RWA 2.0. The general direction is still guided by compliance and policies, especially after Trump took office. Another thing is that DeFi has been developing for several years. At the beginning, there were only Uniswap and SushiSwap competing, and later there were lending and various combinations.
Now I think DeFi will usher in the next stage of opportunities. The first is chain abstraction. If you pay attention to OP Super Chain, it will be easier to understand. In the Ethereum ecosystem, OP's horizontal expansion almost completely beats Arbitrum, although Arbitrum's strategy is different. But from the perspective of horizontal expansion indicators, most of the current Layer 2 and Rollup are almost based on OP, including the recently popular Base chain and Worldcoin, which all use OP's underlying technology to implement their Super Chain plans. Super Chain will form a natural chain abstract state in the future. In the future, asset transfers between Super Chains, contract calls between chains, and protocol communication will become very smooth. This is also a direction of DeFi evolution.
The addition of RWA assets to the chain will definitely bring about one certain thing, that is, more assets on the chain. After the increase of assets, whether as underlying collateral to support other assets or as a circulating asset, the most direct impact is on the DEX and lending tracks. The benefits of these two areas are the most obvious, and they can also be developed in Super Chain and chain abstraction. Secondly, when it comes to the development of old DeFi, it will definitely usher in a new stage. Trump's coming to power has provided policy support for encryption. I think DeFi will have a lot of room for growth in 2025.
Sanzhi: In the long run, from the application layer, I will personally pay attention to Worldcoin. After contacting the Worldcoin fund team before, I learned that they are currently very focused on compliance and adopt a top-down strategy. Because they need to use devices to collect personal data, they have more dealings with the government. Worldcoin's infrastructure is not very good at present, but they have a large user base. If you download their APP, you will find that it is mainly point-to-point interaction, and there is no connection network. There is no chat function in it, but they plan to launch this function in the future.
In terms of the track, I mainly focus on AI and BTCFi. The reason for my focus is that after the ETF is approved, many large institutions will have demand for BTCFi. The third aspect is AI. I think the current bull market has begun, but there is no concept of breaking the circle. I think the concept of breaking the circle may arise in IP or AI Agent, because these two outsiders can also understand. If you talk about RWA and chain abstraction, outsiders may not understand. I thought IP was an important direction at first, but now I find that outsiders can also understand AI Agent.
In addition, the Ethereum ecosystem also has huge potential for explosion, such as OP Stack and OP Super Chain. Ethereum's Layer2 is not like other new public chains. Ethereum's Layer2 may develop into a specialized chain, such as a payment chain, RWA chain, and a specialized identity chain, which can drive the explosion of more applications. The Base ecosystem also has great potential. On the one hand, it may be due to the support of US policies, and on the other hand, the cooperation between Base and Warpcast is also quite interesting.
EO: The areas I am more concerned about in 2025 are mainly DeFi, AI and DePin. Our fund was mainly started by DePin, and we are very interested in this track. When many traditional funds enter this field, they often don’t know what projects to invest in, but if there are some specific physical objects, they will be more interested. For example, in the Hivemapper project, the team made a lot of money by selling cameras. We are quite satisfied with their revenue and have great expectations.
This kind of project can indeed attract new users and there is real demand. Hivemapper's maps have covered 20% to 30% of the world's area, and through a peer-to-peer token reward mechanism, users are encouraged to collect new maps. Therefore, I think these projects can attract traditional funds and there is real demand. Only DePin can connect with traditional businesses. The DePin track has not yet fully exploded. If you are interested, you can pay attention to it. The future potential is still great. I look forward to more projects in the DePin track this year.
Another one is the DeFi track, which is a track that traditional Crypto partners prefer. Because Crypto started mainly by financial innovation. In fact, I have participated in this round of DeFi since 2020, and I have a deep affection for this track and I am optimistic about it. Because it is a track that truly has the ability to generate revenue in the Crypto field, whether it is a large project or a small project, you can make some money. Another one is the AI sector. The AI sector is currently in a very hot stage. Although the income may not be as stable as DeFi or DePin, I think its wealth effect should be the largest this year.
Space link: https://x.com/i/spaces/1ZkKzRblznyKv
Welcome to BlockBeats the BlockBeats official community:
Telegram subscription group: https://t.me/theblockbeats
Telegram group: https://t.me/BlockBeats_App
Official Twitter account: https://twitter.com/BlockBeatsAsia
