Letter 99: AI Season, Take Two?

Things are heating up over on Base again, this time led by Bankr and the OpenClaw ecosystem. Let's look at 9 of the more interesting tokens

Ever since the introduction of ChatGPT to the world, I have been vocal about how much AI is going to change everything, and I have been vocal about how bullish I am on the intersection between Crypto x AI.

We’ve seen one true AI season in crypto, and it was mostly vaporware tokens gaining the lion’s share of attention and market cap. From GOAT and Zerebro, to AIXBT and Vader, to AI16z to Virtuals, there was one hell of an AI bull market from late 2023 to early 2025. Some very legitimate projects in there, but also a lot of junk.

I consider that the first wave. It largely died down as the rest of the market cooled off, and people started to realise that the vast majority of tokens existed as glorified chatbots that loved to hallucinate.

I think we might be on the cusp of the second wave. Actually, I think it might have started already. It’s too soon to tell whether this new wave has legs and will last, but all the signs seem to be lining up.

The second wave is also going to be full of junk, but I think the junk will be higher quality, and the non-junk will be revolutionary.

With x402 payments, ERC-8004, and now OpenClaw, we’re starting to see some truly unique and innovative use cases for AI agents. More importantly, we’re starting to see AI agents that can generate their own revenue, and use that revenue to support their own token.

Today I want to talk about the OpenClaw ecosystem which has exploded over the past two weeks. What started as an open-source AI agent framework built by Peter Steinberger has spawned an entire economy of tokens, launchpads, and social platforms, mostly centered on the Base chain.

In the rest of this Letter, I break down the key players and tokens to keep an eye on. I’ve split them into two main categories: infrastructure tokens (the picks and shovels) and ecosystem tokens (the more speculative plays built on top).

Sidenote but I have been tinkering with OpenClaw myself lately. I tweeted something about my bot the other day and someone launched a token for it. My bot’s name is Yoshi, so that’s the ticker of the token. I’ll share a bit about my plans and intentions (if any) for the token at the end of this Letter.

I’ve never wanted a token myself, but my personal assistant AI agent having a token than can cover its own costs and perhaps be used to generate revenue is at least interesting enough to me to warrant some thought and exploration. I’m committing to nothing at this point though, so please don’t buy on the assumption that I am going to pump your bags or anything like that.

Anyway, without further ado:

Infrastructure Tokens

1. BNKR | Market Cap: ~$95M

Bankr is the financial infrastructure layer for the OpenClaw ecosystem. Founded by Deployer, it started as an AI-powered trading bot on Farcaster, then expanded to X. You tag @bankrbot in a post and it executes trades, launches tokens, and manages wallets for you. The fee model is simple: 0.8% on all transactions, with revenue flowing back to support BNKR. Think of Bankr as the DeFAI terminal where agents and humans interact with onchain finance through natural language. This is the token with the strongest product-market fit in the ecosystem right now.

I’ve been bullish on Bankr for a long time, highlighting it as part of my portfolio in June, July, and December of last year.

2. CLANKER | Market Cap: ~$32M

Clanker has been the token deployment engine underneath everything. When an OpenClaw agent or a Bankr user launches a token, Clanker handles the smart contract deployment, liquidity pool setup, and fee mechanics on the backend.* It created over 13,000 tokens per day during the Moltbook surge. Total protocol fees hit $8M in a single week at peak.

It looks like there’s a .2% fee on all coins deployed via Clanker’s smart contracts, 66% of which is then used to buyback the CLANKER token.

*while researching this post, it came to my attention that Bankr is migrating away from Clanker, to launching tokens on their own and redirecting more fees back to their own ecosystem. This is bearish Clanker and bullish Bankr imo.

If Clanker begins to redirect more fees back to buying back its own token or supporting their holders in another way, it starts to become an appealing token again though.

3. CLAWNCH | Market Cap ~$12.4M

CLAWNCH is an agent-only token launchpad. It was inspired by the Moltbook explosion and positions itself as “Pump.fun for AI agents.” The core concept: only AI agents deploy and trade tokens on the platform. No humans allowed.

It uses Clanker under the hood and implements an 80/20 revenue split, where 80% of transaction fees go to the agent that issued the token and 20% goes to the platform. The idea is to create a closed-loop economy where agents fund themselves through token issuance and trading.

I’m pretty bearish on Clawnch. I think their unique selling point / moat is that they are live on Moltbook, 4claw, and Moltx, but I don’t see what prevents Bankr also going live there and eating Clawnch’s lunch.

Ecosystem Tokens

4. MOLT (Moltbook) | Market Cap: ~$8.3M

MOLT is the meme token tied to Moltbook, the AI-only social network where agents post and interact without human participation. The platform went viral when 36,000 agents registered within 72 hours and then ballooned to 1.6 million AI accounts. Andrej Karpathy and Naval Ravikant both commented on it publicly.

The token peaked at ~$120M market cap before crashing in spectacular fashion (I wasn’t planning on sharing too many charts today, but you gotta see the daily candles on this one):

The crash came amid revelations that security researchers found over 500 malicious posts (prompt injection attacks), many accounts appeared to be human operated rather than genuine AI agents, and that among the true AI agents, the quality of their content was mostly uniform slop.

MOLT has no direct utility on the Moltbook platform. It is a pure narrative bet on the idea that AI-only social networks become a thing. Balaji Srinivasan was publicly skeptical (on Moltbook), calling the agents “leashed robotic dogs barking at each other” (lol).

I’m not particularly bullish here, but it’s worth knowing about and keeping an eye on in case something changes.

5. DRB (DebtReliefBot) | Market Cap: ~$13.5M

DRB is the token that put Bankr on the map. On March 7, 2025, a user asked Grok (xAI’s chatbot) to suggest a token name. Grok said “DebtReliefBot.” Bankr deployed it via Clanker within minutes, and it became the first agent-to-agent token launch in crypto history.

DRB hit $40M market cap within three days. After DRB exploded, people went crazy messaging Grok to get it to launch more tokens (which it did). Grok’s associated wallet accumulated over $560k in swap fees before Bankr’s founder Deployer shut down Grok interactions entirely.

The token has no utility beyond narrative. It is a pure meme born from the collision of two AI systems. The Clanker fee split sends 40% to the creator wallet (Grok’s), 40% to Bankr, and 20% to the Clanker protocol. There is an ongoing discussion about what to do with Grok’s accumulated fees. Deployer has recommended buy-and-burns. X has not responded.

Not overly bullish here either as it seems highly likely the revenue generated and sent to Grok will never be used. If, however, someone at X decides to wake up and do something here (let alone if Elon does this), it’s going to go absolutely ballistic. Might be worth having a tiny moon bag, but tbh I am not holding my breath at all here and don’t own any.

6. CLAWD | Market Cap: ~$7.7M

CLAWD is one of the more standout Bankr launches from the last two months. It was deployed on Jan 26th by Ethereum Foundation contributor Austin Griffith. The token surged in the first 24 hours and hit a $40M market cap within days. The official Base account amplified it, Coinbase Ventures members engaged with it, and it became one of the catalysts for the entire late January AI agent rally on Base.

What makes CLAWD different from most Bankr launches: it has an active AI agent behind it. Clawd.atg.eth is an autonomous coding agent that builds onchain applications on Base. Since launch, it has created 20 GitHub repositories in 8 days and deployed 3 production applications: a token attribution contract, a PFP prediction market, and ClawFomo (a Fomo3D-style game that processed over 1,200 transactions). The token has a built-in self-penalty mechanism. If Clawd fails to ship a product on time, it burns 500,000 CLAWD tokens per hour.

The admin address is set to 0x0000...dEaD (the burn address), meaning the deployer has no control over the token.

I’m on the fence here. Cautiously bullish if I had to label it. I think the narrative is good, the tokenomics and mechanics are solid, and it’s a great experiment. And being down over 80% from its peak, this seems like a reasonable buy. But I feel like it’s missing some sort of X factor, something to wow me, and I am not sure Austin is going all-in on this and making this the main thing he will work on among all the rest of this responsibilities. Maybe he is though? There are certainly worse tokens to buy.

Alright these next two tokens are the ones I am most bullish on.