Author: Marco Manoppo
Compiled by: Deep Tide TechFlow
In a bull market, our social media is always flooded with success stories of achieving 100x returns (like the hype around Hyperliquid, please stop flooding!). However, we rarely discuss the missed opportunities.
In today's article, I want to review the story of Virtuals Protocol, share how I got in touch with the founding team early on, and how, as a venture capitalist, I missed my first 100x opportunity.
Note: Strictly speaking, the first 1000x opportunity I missed was participating in Solana's seed round investment in 2019 through an angel friend, but I was not yet a true investor at that time.
Disclaimer: I am an investor in @primitivecrypto (PV). Although PV is not a venture capital firm in the traditional sense, we also engage in investment operations similar to venture capital. The views expressed in this article are solely my own.
Virtuals is one of the biggest investment opportunities I missed in this cycle. The founding team first contacted me in July (during ETHCC), when their fully diluted valuation (FDV) was only $50 million. Before that, I had actually heard about this project through mutual friends in the first quarter when their valuation was even lower. Fast forward six months, and this AI agent tokenization platform has become one of the focal points of the current cryptocurrency cycle.
The co-founders of Virtuals, Jansen and Wee Kee, have indeed shown extraordinary perseverance.
I clearly remember how tirelessly they introduced Virtuals to investors and industry insiders. Since they spent most of their time working in Southeast Asia (SEA), I heard from some friends in the crypto space about their rebranding from the PathDAO era and their theories on AI agent tokenization. The perseverance to continue pushing the project forward through a bear market without major centralized exchange (CEX) listings is admirable. Many other founders might have chosen to return funds or abandon the project, but the Virtuals team has persisted and returned to the market with greater strength.
Why did I make a poor decision?
Earlier this year, we saw many projects combining crypto and AI attempting to achieve decentralized computing or inference. To be frank, many of these projects are just talk. Most projects lack effective ways to engage ordinary users. Of course, you might earn some airdrops by joining networks and running some computations, but these pale in comparison to GameFi or DeFi's pool2 models in attracting large-scale ordinary investor participation.
Initially, I thought these projects would gather unique data through some gamified approach and combine it with consumer-centric applications to make the experience more unique—perhaps even adding some elements of 'Ponzi economics'. After all, data is still at the core of any AI model; using 'free internet currency' to incentivize people to share unique data seems like the best way to do it.
Do you remember the third season of (Westworld)?
However, it turns out that the crypto market is heading towards polarization. We directly skipped the part I just mentioned and entered the asset issuance phase—which is still the most important product-market fit (PMF) in the cryptocurrency space.
The Virtuals team is well-positioned to seize this opportunity based on all their previous efforts.
The emergence of GOAT
People often say that luck happens when preparation meets opportunity.
Here, I don't need to elaborate on what GOAT is; if you are not familiar with this phenomenon, you can refer to this explanatory article. In short, GOAT sparked a frenzy around AI agent tokens because it made the market start imagining the possibilities when AI agents could interact with some currency. Although GOAT has some limitations, such as requiring a certain degree of human intervention, the key point is that it makes people believe that when AI agents meet cryptocurrency, a whole new experimental field will open up.
Upon realizing this opportunity, the Virtuals team acted quickly to showcase their technical prowess.
Their tokenized AI agent LUNA launched on October 16, just one week after GOAT's release. If you are familiar with the crypto industry, you know that merely being the 'Beta of the protagonist' is far from enough. At a breakfast meeting in Bangkok, Jansen told me that the Virtuals team sprinted to make LUNA the first AI agent capable of autonomously conducting on-chain transactions.
Reflection: How to seize winning opportunities in the crypto market?
Reflection often carries subjectivity, but here are some of my takeaways:
Resilience: The Virtuals team has demonstrated extraordinary perseverance through continuous product iteration. Although different founders may perform differently due to varied backgrounds and motivations, investing in teams that do not easily give up, maintain high ambition, and keep up with market trends is key.
Geographic advantages and rapid experimentation: Generally, projects attempting rapid experimentation (such as platform or launchpad models) find it difficult to succeed in Western markets due to higher costs. However, the Virtuals team, located in Southeast Asia, can iterate quickly at a lower cost while benefiting from a dollar-denominated user base and capital markets.
Resilience and pragmatism: The advantage of Southeast Asian founders lies in their resilience and pragmatic spirit. The business culture in the region has always prioritized 'pragmatism first', whether in traditional business, Web2, or the crypto industry, localizing successful experiences from Western or Chinese markets. This pragmatic and business-oriented mindset is fully embodied in the Virtuals team.
What are the future trends?
The hype around AI agents has only lasted about two months, but I feel like I have gone through two years of baptism. Although the market has shown some signs of fatigue, I believe that by 2025, we will see more AI surges combined with cryptocurrency. Innovations in the crypto industry often start in the most fervent areas and gradually evolve into more mature real-world application scenarios.
An undeniable fact is: without cryptocurrency, AI experiments would be greatly limited.
This is particularly evident in the experiments with AI agents. Imagine that you cannot get a random AI agent in touch with real capital in the traditional financial system (TradFi) unless you prepare a mountain of paperwork and legal support. Not to mention directly handing cash to an AI agent. Cryptocurrency, as a purely digital form of currency, provides the most suitable vehicle for these experiments.
Therefore, experiments with AI agents will gradually evolve from simple functions (like a GPT wrapper that can tweet, valued at $100 million) to more interesting application scenarios. Here are some directions I personally look forward to:
More frameworks and platforms for AI agent tokenization: While the Virtuals team has been rapidly launching new products, there is still significant competitive space in the market. For example, platforms like @ai16zdao, @MoemateAI, @Spectral_Labs, and @griffaindotcom have begun to emerge and gradually capture market share.
Niche AI agent experiments: Some projects (such as @freysa_ai, @aiwdaddyissues, and @Big_Pharmai) showcase more niche experiments and application scenarios. The key for these projects lies in how to evolve from an interesting experiment into a real protocol with long-term commercial value.
Consumer-facing crypto x AI applications: How can we transform AI agents into practical applications that attract consumers while maintaining their uniqueness and innovation? This could even be combined with other AI products (such as data collection, model training, or inference services). The key is to make the user experience of AI agents both novel and practical.
AI agents combined with 'side hustles': I don’t dare to predict too much here, but it’s foreseeable that more AI agents will create significant cash flow by participating in certain 'side hustle' fields (such as gambling, adult industry, etc.), rather than relying solely on token issuance or crypto market trading.
The integration of AI agents with payment systems: As the interaction capabilities between agents improve, we can explore how to utilize AI agents to achieve a more seamless on-chain and off-chain payment experience, thereby optimizing the payment process.
When faced with community-driven innovation, the mindset of traditional venture capital can sometimes seem limited. The core of learning is to maintain an open attitude toward new experiments, unbound by traditional notions, and to be able to adapt quickly rather than blindly pursuing idealism. Primitive is always looking for brave founders. If you are researching any of the aforementioned directions, feel free to reach out to us!
