Original text by Marco Manoppo

Compiled by | Odaily Planet Daily Golem (@web3_golem)

In a bull market, we are always bombarded with information about people successfully achieving 100x returns (like Hyperliquid), but we don't talk enough about missed opportunities. In this article, I reflect on my story with Virtuals Protocol, how I met the founders early on, and how I missed it as a VC (which now seems to be my first 100x return opportunity).

Note: Theoretically, one of the first 1000x opportunity I missed in my life was in 2019 through an angel friend for a seed investment in Solana, but at that time, I was not yet a VC.

Virtuals is the biggest liquidity risk investment opportunity I missed in this cycle. The founders first contacted me in July at ETHCC when their FDV was around 50 million dollars. Before that, I even heard about this project from mutual friends in the first quarter when their valuation was lower. However, I didn't expect that six months later, the AI agent tokenization platform would become the protagonist of the current crypto cycle.

The co-founders of Virtuals, Jansen and Wee Kee, truly possess extraordinary courage.

I remember they tirelessly shared information about Virtuals with investors and operators in the field. Since I spent most of the time in Southeast Asia (SEA), I heard about their rebranding from the PathDAO era and their story about AI Agent tokenization from friends in the crypto industry.

Even in the bear market and without any mainstream CEX listings, their steadfast commitment to the project is impressive. If faced with the same situation, many other founders would refund LP funds or abandon their projects, but the Virtuals team has persevered, and they are now stronger than ever.

What made me think Virtuals would not succeed

Earlier this year, we saw a large number of Crypto x AI projects trying to perform decentralized computing or reasoning. I probably don't even need to mention how many of these projects are hollow gimmicks, as most projects do not truly allow retail participation. Of course, users can earn some airdrops by joining networks and running some computations, but nothing can excite the retail market more in terms of airdrops than GameFi or DeFi.

Initially, I thought Virtuals might collect unique data through some gamified way, combined with consumer-centric applications, differentiating user experience - perhaps mixed with some 'Ponzi economics'. After all, data is still the cornerstone of any AI model; what better way to incentivize people to share unique data than magical internet currency? Just like in Season 3 of (Westworld).

But it turns out that the cryptocurrency market is becoming increasingly complex and extreme (barbell). We skipped the aforementioned process and went directly into the asset issuance domain - this is where the main product-market fit (PMF) of the crypto market lies.

And the Virtuals team, with their previous efforts, is fully capable of seizing this opportunity.

GOAT opens the era of AI Agent tokenization

People often say that luck comes when opportunities are well prepared.

AI Meme coin GOAT has sparked a frenzy for AI Agent tokens in the market as it opened the imagination of what could happen when AI agents can interact with some form of currency. Although there are still many questions about the limitations of GOAT, whether there is some level of human intervention, etc., the most important thing is that GOAT has made people believe that when AI agents combine with cryptocurrency, they will open up a whole new experimental world.

Realizing this opportunity, the Virtuals team knew they had to leverage their technology and act quickly.

The tokenization of AI agent LUNA by the Virtuals team was launched on October 16, about a week after the GOAT launch. If you've been in the crypto industry long enough, you know that merely appearing as a 'beta version' before mainstream projects is far from enough. So at a breakfast meeting in Bangkok, Jansen told me that the Virtuals team is doing everything they can to make Luna the first AI agent capable of autonomously conducting on-chain transactions.

Summarizing lessons leads to success

Essentially, reflection is often subjective, but here are some insights I gained after this missed opportunity:

  • Resilience: The Virtuals team has demonstrated extraordinary resilience through continuous iteration of various products. Different founders have different motivational structures based on their personal reasons and socio-economic backgrounds. Regardless of these factors, the key is to bet on those teams that do not easily give up, maintain strong ambition, and always keep up with market trends.

  • Geographic arbitrage and rapid experimentation: Typically, projects attempting to conduct rapid experimentation (more of a platform/startup model) struggle to succeed in Western markets due to high cost structures. However, Virtuals is located in Southeast Asia (SEA), allowing them to iterate quickly at a lower cost structure while benefiting from a dollar-denominated user base and capital markets.

  • Resilience and business-first priority: The advantage of Southeast Asian founders lies in their resilience. This region has always prioritized business, drawing on the successful experiences of Western/Chinese markets, whether in traditional business, Web2 business, or crypto business, and has been able to localize. The founders of Virtuals fully exemplify this mindset of prioritizing business over ideology or technocracy. They know what they don't understand, what they lack, and what their moat is.

What might happen next

The AI agent craze has lasted about two months, but I feel like I've aged two years. Although the market has shown some signs of fatigue, I believe we will see an even crazier crypto x AI frenzy in 2025. The verticals of cryptocurrency always start from the most 'degenerated' corners and gradually evolve into more mature applications.

One thing has always been true → Without cryptocurrency, AI experiments would be severely limited.

This is most evident in the AI agent experiment. Without mountains of paperwork and legal support, an ordinary AI agent cannot access real capital in the TradFi system. We cannot casually give a real cash to an AI agent, while cryptocurrency is the purest form of currency for digital intelligent life forms.

Therefore, the AI agent experiment will evolve from a GPT wrapper that can only simply publish tweets but is valued at over 100 million dollars into more interesting use cases - personally, I hope to see the following next:

  • More AI agent tokenization frameworks and tokenization platforms. While the Virtuals team has been launching new products, the market is still fiercely competitive. Other platforms, such as ai16z, MoemateAI, Spectral_Labs, griffain, etc., have emerged and taken a certain market share.

  • Niche AI agent experiments. Projects like freysa, Fi, and Big Pharmai show that we can explore more niche experiments and use cases, with the key being how to evolve from interesting experiments into legitimate protocols.

  • Consumer cryptocurrency x AI agents. How can we maintain the advantages and novelty of AI agents while packaging them in an enticing way as consumer applications? It can even integrate with other AI products (like data, training, or reasoning) - the key is to maintain the novelty of the AI agent.

  • AI agents x vice (unethical industries). I dare not elaborate too much here, but we will see more AI agents creating significant cash flow by entering unethical industries (gambling, adult industry, etc.), rather than just through their tokens or trading in the crypto market.

  • AI agents x payments. Advancing the interaction between agents and how we use agents to unlock a more seamless off-chain and on-chain payment experience.

When it comes to community-driven projects, the VC mindset can sometimes be flawed. The key to breakthroughs is always having an open mind towards new experiments, unbound by traditional norms, and being flexible rather than idealistic.

Related Reading

What stories are AI projects within the VIRTUAL market telling as its market cap reaches 3 billion dollars?

Want to mine gold in the Virtuals ecosystem? I contributed an idea