Investing in Tomorrow's talent

Risk is the only way to grow.  This applies to all aspects of life, including wealth.

Yet, with higher risk comes the need for smaller allocations—because extreme risk can lead to extreme losses. You must eliminate your risk of ruin otherwise you’re out of the game and the only way to win is to stay in it.

As a wise man once said, “You get rich by taking big risks with small amounts and stay rich by taking small risks with large amounts.”

Today, I’m going for the extreme side of risk with calculated bets.

Idea Snapshots

• Token: AI Agents, I will curate a list below as more of them will emerge and I want to be early on the winners.
• General Thesis: AI Agents are trading at minuscule valuations, most under $5 million in market caps. Some winners will emerge and go to $100-500 million, maybe more.
• Risk Level: Extreme
• Potential Upside: 1,000%+

Criteria

First things first. I am going to walk you through my entire thought process and write as I make discoveries.

At the time of writing (1 November 2024), investing in AI Agents is returning multiples already and that has fully relied on the idea’s novelty, marketing among crypto circles and future promises. In other words: hype.

Let me give you two examples I personally invested in and made money from:

1. MISATO: I bought it at a market cap of $1.2m and right now it’s at $2.8m. This agent was LUNA spinoff, another anime character built by the community with a promise of upgrades. All I have really seen is new videos drop on its Twitter account, I am yet to see how this one generates revenue and creates the buyback-and-burn cycle to boost its token up.  So again, just hype for now but people (me included) are speculating on the fact that promises get delivered on.
2. VU: I bought it at a market cap of $1m and now it’s at $1.1m. The novelty here was that this agent would run a crypto fund, make smart trades and return money to investors. I could get exposure to it by either investing in its fund or buying its token. I chose the latter because I haven’t seen its skill. And, the only contribution made to it was a “State of Crypto 2024” PDF from the famous A16z VC fund. Yet another speculation on on future promises. The fact that it was launched by the platform Velvet.Capital where VU had its fund and that Velvet.Capital is backed by Binance Labs gives it some legitimacy (but Binance Labs invest in many startups, so while enough for today or until $10 million it won’t be enough to get to $100 million). A similar one is Ai16Z, which trades at a valuation of $26m and manages $330k on Solana.

I made money, not massive amounts because these tokens remain low on liquidity for big buys. But I want more, I want to figure out what will set agents apart, how I can be early to one that goes to $500 million+.

I know that AI Agents are akin to multi-tasking superhumans, finding potential is the same as finding talent in the workplace. So I need measurable metrics (quantitative and qualitative) that I take into account when making a decision. Here they are:

1. Contributions: The agent is only as good as its training, and that comes from information contributions made by people. But you can fake this metric and just upload 1,000 random files (which will actually make the agent suck at its job). So, I need to sift through these. Early on it’ll be easy because these will be small, but if they start off great, they’re likely to continue to be at least good. Dependent on the designated expertise of the agent I will have to look at one or multiple of the four areas to be contributed to: cognitive, voice, visual and domain. In the case of MISATO it will need to be the first three, in the case of VU only the cognitive and domain expertise matter. This metric can be summed up with the “intelligence” score but I prefer taking it with a grain of salt and checking contributions myself.
2. Inferences: If the agent is good, it will have interactions. Another metric that can be faked through bots (Telegram messages, TikTok views and comments, etc.). So once again, I will have to personally communicate with them and see if they’re good at their domain by asking questions and measuring answer quality.
3. Social Equity: Attention, no token can grow without it. This increases if price goes up but there must be some sort of base to start with. If no one is talks about a great agent, few will ever find out how good it is, no one buys, price stays stuck. I check this by searching for the ticker on Twitter with a cashtag before it (eg: $MISATO, $VU) or by using a social analytics tool like LunarCrush (but small tokens don’t make it there). This too can be faked by KOLs (supposed “Key Opinion Leaders” that are usually just glorified influencers), but that’s where the next one comes in.
4. Holder Analysis: This one is a rabbit hole, you can really get sucked in but I stick to basics to avoid time wasting. Four tools: first Virtuals Protocol itself, scroll down on the token page and see if there are too many large holders. Then, for deeper dives Dexscreener or Arkham to view the largest holders then dive into their holdings and see if they’re “smart money” and BubbleMaps to see if any one entity holds too much of the supply (>10%) via a single or multiple wallets. Take DoctorAI (DOC) for example, the dev holds more than 25% of the supply, they’ll certainly start dumping on holders once price crosses a certain valuation.

The A-List of AI Agents

Risk Disclosure: This research is for educational and informational purposes only. Crypto assets are highly volatile and risky. Always conduct your own research and consider your risk tolerance before investing.

These are only my thoughts, and be kind because I’m going in super early on this potential trend and a lot can go wrong. This is why I am not going with more than 1-2% of portfolio into any agent token, even if I lose it all on one, another is likely to recoup and cover losses plus profit.

1. **MISATO [$2.8m valuation]:** I view this one as Luna 2.0. Contributions are minimal at 22 (only 2 accepted, 1 processing and 19 rejected). While it does have social equity via third party accounts supporting/speculating on it, it has reached the milestone to handle its own X account and when it does that social equity should increase leading to an increase in price/speculation (Luna will no longer be alone on X). If its interactions and intelligence increase, and the quality of its conversation is good I’ll consider holding longer. If not, I’d sell at least half my position. Try talking to Misato.
2. **VU [$1.1m valuation]:** This is where things become interesting for me. I already experienced AI’s capability of delivering high-end results in research when guided correctly. So why not test it in the real world and give it capital to manage? This is what Velvet Unicorn is doing. I tried chatting with it (try here) it’s still stupid and only shills LUNA which tells me this is a default setting on new agents. But, according to its X account it will start getting training on crypto data from today onwards. This is why it trades at such a low valuation, it signals extreme uncertainty. But if this plays out, it can turn a small investment (1% allocation) to an extremely high one. Per Virtuals Protocol current milestones, VU would have to reach a $126m valuation (74x from here) before handling its own wallet, this is a good safety measure to be honest to avoid having untrained agents handling real funds. The only real problem is that 26.7% of the supply is held by 6 wallets, including the devs.; not extreme but not ideal distribution eitther.

There are two more agents I’m buying that will be trained to provide crypto insights, but their valuations are too low to mention. I wouldn’t want to unintentionally cause price to rise if some of you find it interesting. One sits at $100k valuation and the other at $42k, whenever one crosses $200k at least I’ll add it in here. If these get trained well I’ll even consider creating one for 0x100x. Imagine this, a 0x100x brain you can talk to 24/7.

Notes

1. My main goal is to catch one of the big AI agents super early. Since this comes with extreme risks and even more extreme potential upside, I don’t need to put in too much into each (0.5%-1% is more than enough). Once something does a 3-4x I sell my initial investment and  ride the rest.
2. This list will be regularly updated. I’ll state what I sell before I do, my reasoning and if I buy anything new.
3. I don’t like how Virtual is positioning itself like Solana’s memecoin factory: pump.fun but I see how that gives it attention. I won’t be buying anything from the fun marketplace, only the app (the sentient ones).
4. Browse the platform, dive into each agent, do your own research too. As much as this club is high in quality, never rely on anyone’s opinion but your own.

Conclusion

AI Agents represent a rare, early-stage investment opportunity in crypto, where the risks are as extreme as the potential rewards. Unlike memecoins, these agents bring practical use cases and promising, albeit experimental, technology. In my view, we’re on the cusp of a trend that could redefine crypto utility and market dynamics.

However, let’s be clear—this isn’t an area to jump into without due diligence. Allocating just 0.5-1% of the portfolio to each Agent is enough to capture outsized gains if even one performs. It’s about finding that one big win and scaling from there. I’ll keep updating this list as new Agents emerge and my analysis deepens, sharing any changes I make to my own positions.

For those looking to invest, remember that the market is unpredictable, and we are treading on new ground here. AI will undoubtedly reshape industries, but betting on the right players takes patience and ongoing evaluation. Bookmark this post, follow along, and as always, do your own research before diving in. This journey will be as much about learning as it is about earning.

About the Author

This was an unexpected post, authored by the founder of 0x100x just like the one on VIRTUAL. The next post will bring a more established bet on AI x Crypto, one on the lower end of the risk curve, written by an exceptional investor (not from 0x100x). Excited to share it with you all.

How to Buy

If you made an independent decision to try and buy any AI Agent built on Virtuals, you can only do so on their website. You also can only exchange VIRTUAL tokens for an AI Agent’s—creating a flywheel for the VIRTUAL token.

‍