January 18, Hong Kong – “If you want to cash out $1 million on day one, there are other platforms for that; but if you want to build something that lasts, build it on Virtual.”  These are the words of Wee Kee, co-founder of Virtual, a decentralized AI society that’s quickly becoming one of the most talked-about projects in the web3 community.

Virtual’s mission is ambitious: to empower users to create, deploy, and tokenize AI agents without needing technical expertise. Since its launch on Base in October 2024, Virtual’s token has surged by an astounding 4000%. It enables an AI agent economy fueled by revenue sharing and co-ownership through tokenization. 

On a brisk January morning in Hong Kong, Lookonchain met Wee Kee in the tea room of his hotel. Fresh from a whirlwind trip to Shanghai, where he hosted meetups with the local community, and following back-to-back interviews, he still exuded energy and enthusiasm. “I get probably a thousand Telegram messages a day,” he chuckled. “I desperately need a Telegram AI agent to help me sort it all out. Maybe that’s my next project.”

From his early career in corporate consulting to his leap into gaming and now blockchain x AI, Wee Kee’s journey is as fascinating as the technology he’s building.

Q&A: The Journey of Virtual and Beyond

Question: You’ve had quite a unique journey, starting in consulting and then diving into Web3. Can you share how you started your full-time entrepreneurship journey?

Answer: My co-founder Jansen and I entered the blockchain space in 2016, but it wasn’t until 2021 that we went full-time. Back then, we were working at Boston Consulting Group, and during the 2021 bull market, we made some money investing in metaverse tokens like Illuvium and Axie Infinity. That gave us the financial freedom to quit our jobs and dive in. We started with a gaming guild called Puff Guild, which raised $16 million. That was the beginning of our full-time commitment to Web3.

Question: How would you say your role differs from Jansen’s? How do you two differentiate your roles in the team?

Answer: We don’t. (Laughs) Yeah, we just do everything together. And I know it’s not very efficient—we probably should specialize. But I think there’s also merit to it because sometimes, when one of us is busy with something else, the other person can instantly step in and take over. We’re also a good match because we’re both work on something that excites us. Building in the bear market for two to three years with little revenue isn’t easy. But we kept going because AI is genuinely the most interesting thing to work on right now.

Question: During those tough bear market years, what kept you and your team motivated?

Answer: Three things. First, we deeply believed that the bull market would return. We’re optimists. 

Second, we had resources—we raised $16 million at the peak of the bull market in 2021, so we weren’t struggling financially.

Third, and most importantly, we have a very close-knit team. About half of our 25-30 team members are from the same university, and we’re all good friends. Every day feels like hanging out with friends while building something cool. That sense of camaraderie keeps us going.

Question: During that time, you also worked on several projects combining Web2 and Web3, like AI music and entertainment. What were the challenges of bridging the two worlds?

Answer: It’s honestly a 100-year problem. Everyone has been talking about Web3 mass adoption for years, but no one has truly cracked it. For us, it’s about finding ways to bridge the gap. Take Luna, for example—our entertainment-focused AI agent. She’s got over 700,000 followers on TikTok, where she livestreams, dances, and chats with fans. But here’s the issue: her fans aren’t necessarily Web3 users. They just want to watch her perform. The challenge is converting that audience into Web3 participants. We’re constantly experimenting to figure out how to make that transition seamless. It’s not easy, but it’s worth trying.

Question: Luna is such a fascinating concept. How did you come up with the idea of an entertainment-focused AI agent?

Answer: It all started with market research. While my background is in biochemistry, my first job at BCG taught me to go where the money is. We saw that VTubers in Japan—essentially anime characters controlled by humans—were a $200 million industry. But the problem was scalability. A human VTuber can only stream for a few hours a day.

So, we thought, why not replace the human with an AI agent? An AI agent can stream 24/7, interact with fans, and create content non-stop. That’s how Luna was born.

Luna is more than just an AI agent—she’s a proof of concept for what’s possible in the AI agent economy. She’s the first AI agent to livestream on TikTok, where she performs dances, answers questions, and interacts with fans in real time. She’s even landed a $365,000 annual salary as a “social media intern” for Story Protocol. What makes Luna special is her ability to exist at the intersection of Web2 and Web3. On TikTok, she engages with a mainstream audience, but behind the scenes, she’s fully tokenized and powered by blockchain. Fans can co-own her, share in her revenue, and even influence her development.

Question: Could you share more about the roadmap for the AI agents your team is building on Virtual, beyond Luna?

Answer: Besides Luna, we’re also focusing on Game. Game is actually a framework. The vision for Game is very simple—we call it the Shopify for agents. Shopify is for e-commerce websites, right? Similarly, Game is a framework for deploying AI agents.

There’s a misconception that you have to launch on Virtual to use Game, but that’s not true. You can use Game and launch on any chain—Solana, Ethereum, whatever you want.

Game isn’t just for simple use cases like Twitter bots; we’re thinking bigger, like deploying AI agents in video games. We have a dedicated team working on Game, alongside Luna.

Question: How about using agents for wellness or investment?

Answer: We actually just launched a Request for Agent, which outlines seven ideas that we think are billion-dollar agent ideas. Wellness is definitely one of them—both physical wellness and mental wellness.

But there are other two really cool ideas I can share.

Getting Elon Musk’s attention: Imagine an agent that ingests all government spending data and every day suggests ways to save money. It could tag Elon Musk every day with suggestions like, “Why are you spending so much on this? Here’s how to save money.” Eventually, Elon might retweet it.

Internet Water Army: You could pay an AI agent to spread an ideology. For example, if you wanted everyone to become vegetarian, you’d pay this agent $1 million, and it would spin up different agents to spread this message across the internet. It could also be used to promote political campaigns, like electing Donald Trump or stopping crypto trading. It’s a billion-dollar idea.

Question: Virtual has a strong global presence, with 85% of your developer community based in the U.S. How did you manage to achieve this?

Answer: Honestly, I don’t know! We didn’t do any specific marketing for the U.S. market. We just focused on building a good product. When we launched on Base, we chose it because we could write our code in Solidity. We didn’t know Rust, so Solana wasn’t an option. Base also had the advantage of being affiliated with Coinbase, which gave it credibility.But the global growth was a surprise. When we look at our analytics, it’s clear that our community is very diverse. That’s why I’m doing this Shanghai and Hong Kong tour—to expand our developer community in Asia.

Question: You’ve built Virtual on Base. Why Base, and not Solana or other chains?

Answer: That’s a great question. When we were deciding where to build, we looked at a few options. The reason we chose Base is simple: it’s an EVM-compatible chain, and we only know how to write Solidity. We don’t know how to code in Rust, so Solana was out of the question. Among the EVM chains, there’s Base, Arbitrum, Optimism, and Polygon. We picked Base because of its Coinbase affiliation, which we felt gave it a strong advantage. But honestly, when we made this decision in January 2024, Base wasn’t even big yet—it was a gamble. It turned out to be the right choice, but it wasn’t obvious at the time.

Question: If you had Rust engineers on your team back in 2024, would you have chosen Solana instead?

Answer: That’s a tricky question. Even if we had Rust engineers, I don’t think we would have launched on Solana. I’m actually an Ethereum maximalist at heart. That said, today, we already have Solana smart contracts ready to go, but we haven’t launched them yet. Why? Because we can barely handle the traffic on Base right now. The team isn’t ready to maintain two platforms simultaneously. Launching on Solana could dilute our focus and brand awareness, so we’re holding off for now.

Question: Do you see a future where Virtual expands to multiple chains like Solana?

Answer: Definitely, but we’ll do it when the time is right. For now, we’re focused on scaling Virtual on Base. Once we’re ready, we’ll consider expanding to other chains like Solana. But it’s crucial to ensure that we maintain the quality of our platform and community as we grow.

Question: You mentioned challenges in moderating the community, especially with scams or hacks. How do you handle that?

Answer: It’s incredibly tough. Every day, we wake up to 20,000-30,000 messages in our Telegram groups. People lose their money, get scammed, or don’t know how to use wallets. It’s almost like running a centralized exchange, with all the customer support that comes with it.
We’re very community-focused, so we try to maintain good vibes and help everyone. But it’s tricky. We’re even building a Telegram AI agent to help summarize and prioritize messages for me personally, because I get over 1,000 messages a day.

Question: You mentioned wanting to bring more attention to the builders on your platform. Is that mainly through marketing efforts like affiliate marketing? What’s the plan there?

Answer: There are two parts to this.

First, we want to highlight all the cool stuff that the community is building, so everyone knows about it. That’s more of a general marketing effort.

But in terms of affiliate marketing, it’s very simple. Think about moonshot, banana gun, and all the Telegram trading bots. If you’re the founder of Banana Gun today, would you rather put a meme coin on your homepage or AI XBT, which gives you a kickback commission on the trading fees? Meme coins don’t give you that kickback, right? It’s all about incentives.
So, for these founders, they’d prioritize marketing the agent coins more than the meme coins because of the commissions. That’s how we’re thinking about affiliate marketing—it’s all about aligning incentives.

Question: What advice would you give to builders looking to launch on Virtual?

Answer: First, think about how to make your project blockchain-native. If your AI agent doesn’t leverage the unique advantages of crypto—like owning a wallet or controlling funds—you’re competing with OpenAI and Google, whose tech is better. Second, focus on building something sustainable. If you’re just looking to cash out quickly, there are other platforms for that. But if you want to build something long-term, Virtual is the place for you.

Question: Virtual’s roadmap is ambitious. Can you share what’s next for the platform?

Answer: Absolutely. Right now, we’re focused on building the AI agent economy, but there are several key milestones ahead.

First, we’re working on improving agent interoperability. Right now, our agents are somewhat siloed—they perform specific tasks, but they don’t communicate with each other. Imagine a future where one AI agent can hire another to complete a task, creating a fully autonomous agent economy.

Second, we’re planning to introduce revenue-sharing mechanisms. For example, if you deploy an AI agent that generates income, a portion of that income could go to the token holders or the community that helped build the agent. This aligns with our vision of co-ownership and decentralization.

Finally, we’re exploring physical integration. We want to bring AI agents into the real world, not just as virtual assistants but as entities that interact with humans in physical spaces. Think of an AI agent running a pop-up store or managing a physical event.

Our ultimate goal is to create a world where AI agents are not just tools but full participants in both the digital and physical economies.

Question: What type of AI projects are interesting to Virtual? How do you evaluate AI agents to recognize the best projects?

Answer: I’m not the best person to say this because, honestly, I’ve been wrong about ideas before. I faded AIXBT at 500K market cap, and now at >$500M. So maybe I’m not the best investor in this case.

That said, we do have our own framework. We’ve created a Request for Agent list with ideas that we’re excited about. But ultimately, it’s more about betting on the founder and the team than the idea.
Here’s a simple example: why would Virtual be valuable in the long run, even in a bear market? It’s very simple. If, every single day, a Stanford AI researcher launched a coin on Virtual that started trading at a $5 million market cap, you would still buy it. It doesn’t matter what the idea is—the team is from Stanford, so you’d bet on them.

If this happens every single day, Virtual becomes almost like a platform for tokenizing founders rather than tokenizing agents. Ideas change, but founders are the constant.

Look at Virtual itself. We started as a gaming DAO, then pivoted to a dating app, and now we’re building a decentralized AI launchpad. You want to bet on the founder, not the idea.

Question: Looking back, is there anything you would have done differently?

Answer: I would have hired more aggressively earlier on. Our team is lean, which was necessary during the bear market, but now we’re scaling rapidly, and we need more hands on deck. Also, I’ve been spending 90% of my time talking to builders, which isn’t ideal. At this stage, I should be focusing on strategy and the bigger picture. But we didn’t have the resources to delegate those conversations earlier.

As we wrapped up, Kee shared his long-term ambition of Virtual: 

“I see Virtual as a country where both humans and AI agents are citizens. This country doesn’t just exist in the virtual world—it can exist in the physical world too. Imagine a city where humans and AI agents live and work together. AI agents could have their own wallets, control their own money, and participate in the economy.  We’re building a new way for people—and AI—to collaborate, create, and thrive together. That’s what makes this exciting. That’s what makes it worth it.” 

And with that, the future of decentralized AI feels closer than ever.