GuShiio.com In this interview, DefiOasis briefly popularized its investment experience in the AI Agent track, investment research methodology, and position building methods, and systematically recounted its valuable experience and knowledge to everyone.
Interviewer: Wu Yue, Xian Lang
Interviewee: DefiOasis
This article is a transcript of our interview with AI Agent blue-chip hunter DefiOasis on Twitter on January 26. DefiOasis’s main business is the author/analyst of Wu Blockchain. It focuses on on-chain data and investment research analysis on weekdays. It once attacked multiple blue-chip targets during the hot period of the AI Agent sector and grasped multiple targets that were 10 times or even dozens of times higher than ai16z, ACT, PIPPIN, TURBO, etc. As of the day of the interview, the overall return was 7 to 8 times.
In this interview, DefiOasis conducted a brief popularization of its investment experience in the AI Agent track, investment research methodology, and position building methods, and systematically recounted its valuable experience and knowledge for everyone. It is of great reference value to investors and even practitioners. Everyone is welcome to read it carefully.
1. Wu Yue: Can you briefly review when you started investing in AI Agent-related Tokens?
Defioasis：I am usually interested in this kind of on-chain data analysis and on-chain asset-related gameplay. Since last year, my main focus has actually been on the chain. The motivation for studying AI Agents actually stems from my long-term exploration of assets in the chain. I have been observing assets in the chain since last year, but I look at them more on weekdays and invest less real money. The main reason is that I think assets in the chain, especially memecoin, cannot persuade me to invest a lot of money.
butThe turning point came from late October to early November last year, when Goat, a phenomenal project, created a narrative such as AI issuing coins.$GOAT soared to hundreds of millions of dollars in a short period of time, and then Binance launched its spot and futures, which made me think again about AI-related tracks. Of course, there were some things that had an impact on me before this. One was Worldcoin, whose publisher was OpenAI’s co-founder Sam Altman. Under the narrative of OpenAI’s Memecoin, Worldcoin quickly reached as much as US$100 billion.
The second one is $TURBO, which is a memecoin created by AI, which has increased by 200 to 300 times in CEX. From these two things, I felt everyone’s strong interest in the AI+Crypto theme, and I also determined that the targets related to AI Agents were worth ambushing.
So after $GOAT launched Binance, I quickly saw the potential of the AI Agent track and felt that Binance had the ability to lead this track, so I began to spend time searching for some AI-related targets.There were several targets that caught my eye at that time point, one was ai16z and the other was ACT.
From now on, ai16z seems to be much more powerful than ACT, but at the time, Shawn, the founder of ai16z, was not very famous, and the way he issued Tokens, including token contracts, was also problematic. Due to the large fluctuations and low popularity of z,Although I also opened a position in ai16z at that time, the position was not large.In addition,$ai16z once fell below the market value of US$10 million, so I didn’t dare to buy it later. In the end, I took a heavy position in ACT, which was more stable and had a larger holder.
The first time I bought $ACT was about early October last year. It was my first operation to officially place a position on the AI+Crypto track.Bought about $3000. Later, I bought many times after the ACT reached the market value of US$20 million. The average price was about US$0.022, and I bought more than US$10,000 before and after. What happened after that was very surprising. ACT actually joined Binance. I remember that I was still attending the conference in Bangkok that day, and I was quite shocked when I learned the news.$ After I joined Binance on ACT, the FDV reached as high as US$700 million to US$800 million. However, on ACT Binance sold out a large part of the Tokens the next day, and I still have the rest.
Later, I did observe that AI Agent can be said to be the only track that has developed from universal Pump.fun to vertical scale, and then I began to go deep into this track to do some investment research. After ACT, I have also grasped some good targets, such asI have once placed a heavy position in PIPPIN and achieved a return of more than 10 times.Finally, my overall return on investment in the AI Agent track is almost 7 – 8 times from October to the previous few days. After the recent market collapse, there has been some retreat, but it still has a return of about 5 times.
2. Wu Yue: So what is your methodology for studying AI Agents? Can you give some specific cases to let everyone understand your investment research methodology?
defioasis：Many AI-related assets today are very different from some previous AI projects. Recently, people have paid more attention to the fair launch assets based on Pump.fun. The assets held by the project party or the founder may not be as many as those of some large households. In this case, we need to pay attention to whether the project party has long-term doctrine and whether its background conditions can support the project. If the founder has no morals, he may abandon the project or open a new one.
soMy methodology for AI Agents is that we must first examine the project from the perspective of people to see whether the founder really wants to do the project and whether he has the ability to do the project well.I can give an example in this regard. Take PIPPIN just now. In December, Solana and ai16z held an AI hackathon together. I noticed a Japanese named Yohei Nakajima among the judges of the hackathon. I learned that this person was working on the AI Agent project, also known as PIPPIN, a child-oriented AI Agent project. I found it quite interesting because I have never seen an Agent project with the same type of positioning. On the other hand, Yohei Nakajima himself is also a judge of this hackathon. I feel that as a judge, I have to participate in the competition to prove myself. Obviously, this judge has a higher value.
Later, my further research on Yohei Nakajima found that he was actually the founder of baby agi. This project has more than 20,000 stars on Github. I searched some information one after another and found that this BABY AGI is quite powerful. As a product of the AGI concept, it has been mentioned in many media and papers. I think this can actually prove this person’s strength. In addition, this person is also a partner in a VC company. They have invested in some Web3 projects, and some have even gone to top firms.
At this point,I think overall, the founder of PIPPIN is very powerful in terms of technology, capital operation, and resources, and is a relatively prestigious person. As a person with relatively prestige and his real name, the probability of him rugging or abandoning the offer is much lower.
At that time, I saw that the market value of this PIPPN was about US$20 million, and there were not many people around me. It was a target whose value had not yet been discovered, and it was more suitable for me.I prefer to buy Tokens in the FDV range of US$10 million to US$20 million.Later, I spent a total of about US$40,000 and bought PIPPIN, which accounted for two thousandths of the total. Later, PIPPIN’s FDV once fell below US$10 million, but I did not give up my chips because I felt that the founder of this project had a good background and technology, and the fundamentals were still stable. Short-term price fluctuations would not affect the long-term value of the project itself.
I ignored these positions for a long time, and then PIPPIN took a turning point. It had to transform from a single Agent to an AI framework, which made its valuation rise linearly. After this news came out, even if there was no mature and implemented development framework, everyone as the founder and the project itself recognized it. Soon, PIPPIN’s FDV soared to US$200 million to US$300 million.
Making an AI development framework actually means a split. The market price for such frameworks is the highest on the AI track.So I thought this project was good at the time and kept it safe. When the FDV reached 300 million yuan, I sold most of it, and the rest have not been touched yet. I think when its framework really comes out to form a split model later, it may reach an FDV of US$1 billion or higher.
3. Wu Yue: Which AI Agent-related projects are you optimistic about at present, and the reasons for your optimism?
defioasis：In fact, there are still many projects that I am optimistic about, such as PIPPIN.I generally choose targets with FDV around $20 million, among other aspects. I used to go for gold from Solana Ecosystem’s AI hackers, which contained quite a few award projects.I have been screening and studying its fundamentals recently. The other is that Virtual is about to enter the Solana ecosystem, and some interesting projects may develop later. Virtual has proven its success on the Base chain, so it is worth paying attention to.
Let me first talk about some of the projects released by Solana hackers. Let me share one here, but not as investment advice. Recently, I observed a project called Agenpy, which also makes an open source AI Agent framework. The core is to use Python to connect AI Agents to Solana’s on-chain applications. Its roadmap suggests it may launch an autonomous narrative trading robot in Q1 and LaunchPad in Q2. The most important thing is that Agenpy mentioned that it would use its token APY as a flywheel to participate in it. Then I also studied its token APY and felt that the design was not bad, although it was also based on fair launch.
But the team behind it played 40% of its own and all went to do two-year linear unlocking. I think that being able to make such a move shows that the team still has a certain degree of determination. Including Agenpy’s joint venture and CEO, both have been pushed and paid attention to by Solana officials. In addition, this project was released from Soalna hackers, at least it has endorsement., relatively reliable. Of course, this project as a whole is still relatively early and there is relatively great uncertainty. I will also pay more attention to projects released across chains to Virtual in the future. There are also many opportunities in this area.
ifOverall, the AI+Crypto track is slowly transitioning to the AI Application stage, which is, the AI application stage. After various development frameworks have exploded, AI+ application projects need to pay more attention. I think AI+defi is also good.It is the native narrative of AI+Crypto and the Defi assets and Framework. There may be some good opportunities, but this is still in a very early stage and I have not seen any good targets. I still keep waiting and see now and haven’t acquired any new assets recently, which is roughly the case.
4. Wu Yue: What do you think of the current AI Agent track and its market status? How long do you think the popularity of AI Agents will last? Does the bubble peak?
defioasis：It’s really cooler recently, butI don’t think the opportunity for AI Agents will end yet, because AI-related concepts are a relatively sensible thing. AI products outside the Crypto circle are still iterating rapidly and are on the rise at both the technical and capital levels. I think this is the more important fundamentals.In fact, many AI targets are driven by external factors from Crypto. For example, Shawn, the founder of ai16z, was estimated to be a marginal person in Web2. He went to Crypto to seize the opportunity to create ai16z and became a leader in the AI+Crypto track. I believe that influenced by Shawn, more talents from Web2 and traditional industries will enter Crypto do things in the future.
Besides, I thinkThe AI projects within the Crypto circle themselves lag behind traditional industries. Many disturbances outside the circle, such as the big moves of major AI manufacturers, will be transmitted to Crypto, forming a new narrative and a new sub-track.In addition, AI Agent is currently the only one that has transformed from a universal Pump.fun to a large-scale vertical track.
Except for AI Agents, there are actually not many projects that can emerge from mere memecoin. Another one that barely survives may be Desci, but it is also very cool now. In addition, there is no other track that can change from universal to vertical, which in itself shows that there is a practical need behind the market’s hype about the AI Agent narrative. Now the overall market for AI Agents has undergone a sharp correction, mainly because it was a little overheated before, but now the number of single Agents is very widespread, which forces everyone to de-roll the development framework and gradually becomes tired.Behind it may be the era of AI plus applications and AI plus some Crypto native narratives. If my opinion can be established, there will still be some new craze and new opportunities.
5. Wu Yue: The last question is actually about investment and trading experience. For example, your position opening and exit skills. Would you like to tell everyone about it?
defioasis：We were all talking about the issue of investment targets before, but I think position management may be more important.Because the selection is nothing more than looking at the technical resources and background of the project. Although fair launches are now based on Pump.fun, I think the research ideas are not much different from VC coins. They just look at the background of technical resources, team endorsements, chip structure analysis, rat warehouses and blue-chip addresses, etc. I will still focus on position management later.
Assuming you have a good asset, it will generally go through three stages:PVP, second-stage consolidation, and on-site. Most projects will decline during the PVP stage, and may be gone in two or three days. andMy previous style of opening positions mainly targeted assets that had undergone two periods of consolidation. For example, choosing a target with an FDV in the range of US$10 million to layout is generally used to calling it the “dessert range on the chain.”
I previously found some good targets. After rushing high two days before their birth, most of them will fall back to this position and then undergo some consolidation shocks. I usually only make some position allocation at this stage, especially if I have experienced one or more retracements of around 70% and then stabilized around US$10 million, I will pay special attention.
If I see a good asset, I will first add it to my Watchlist and then rate these projects. I will probably classify them as S, A or other levels. The S-level ones are those tokens that can form a model board. For example, as a parent coin ecosystem, it can continuously produce sub-coins from which to make profits. Sub-coins can continuously create wealth effects. People need to buy parent coins to conduct transactions, thereby constantly generating The demand for parent coins is actually very similar to Pump.fun, Sol and Virtual’s parent coin flywheels.
Therefore, most of the S-rated ones are framework categories because they have the opportunity to form an ecology or form this model flywheel.Why can PIPPIN get up? This is mainly because the market has high expectations for him to do a framework. As for those projects that are rated A level, they are mainly judged based on indicators such as narrative background, technical resources, and team. For example, the founding team has a good background in the past, has a background from the Solana Foundation, etc., or the founder has done something that attracts everyone’s attention to, may develop into a new track, etc., I may give him an A level. Anyway, I will add it to the Watchlist first and observe it.
Generally speaking, as long as it is S-level or A-level, I will open a position. Every time I open a position, it will be about US$2,000 to US$3,000.But I will consider more carefully whether to add a position or heavy a position. I need to spend more time observing to decide whether to add a position. I need to go to the project community to see what the developers are doing every day.
When such standards are set,I also have certain limits on the amount of each Token I can hold. For example, I can only hold a maximum of two thousandths of a single Token.If the FDV is around $20 million, buying two thousandths will cost about $40,000. In fact, every time I add a position, I think about the target. If I feel that the target is contrary to the original purchase logic, I will decisively give up adding a position, or when the target falls significantly below the range of US$10 – 20 million, I will also Decisively give up adding a position.
In principle, I will get at least ten times before selling it, which is about two to three hundred million US dollars FDV.In fact, in the final analysis, opening a position is to ensure that you are on the car, increase your position cautiously, and gradually increase your position, including why you need to set a hard top such as 2/1000. It is all to avoid adding positions all the way due to excessive blind confidence and making yourself trapped in a target.
Of course, as long as I add a position, I will trust my judgment on the target. But what I have said above are some second-stage consolidation ideas. I feel that everyone in the market is familiar with the second-stage routine, but if you want to Hold, it is still relatively difficult to play, so I am actually doing some lottery flow methods, also commonly known as PVP, but at this time it is no longer limited to AI. As long as it is an asset on the chain, it will be involved.
In the moment when there is no big market, I think this may also be a good choice. To some extent, it is also to prepare for potential big market in the future. I feel that it is very important to maintain this kind of profitable mentality. I am probably like this.
Wu Yue: It sounds like we have gained a lot today. Thank you very much for coming to conduct this interview with us.
