Khala Research recently dropped a deep dive on the @openclaw ecosystem. TLDR: We’re entering a world where AI agents don’t just help humans build software. They build, launch, monetize, and improve software on their own. Think: > AI creates an app > Submits it to the App Store > Runs ads > Makes money > Uses profits to buy back its token > Repeats That’s the “Zero Human Company” vision. - In just a few weeks, OpenClaw-based agents have: > Launched 50K+ tokens > Processed ~$194M in volume > Generated ~$51K in real Stripe revenue (external users) > Built 50+ smart contracts > Shipped 7+ live apps > Created an onchain treasury with published receipts - The 5 Main Projects (In Plain English) 1) @KellyClaudeAI – The AI App Factory Goal: Build apps 24/7 and turn them into recurring revenue machines. What’s real: > Approved App Store apps > Claims of multiple builds > ~$5.6M valuation What’s unclear: > Actual recurring revenue > Confirmed buyback system This one is priced on the founder’s brand and big vision. It needs real user revenue to justify the valuation. 2) @clawdbotatg – The Infrastructure Bet Built by Austin Griffith (huge credibility in Ethereum dev circles). Instead of building apps, Clawd builds tools other agents can use. Think: > Agent-to-agent communication > Smart contract primitives > Onchain coordination tools Risk: Infrastructure only wins if the ecosystem keeps growing. 3) @FelixCraftAI – The Revenue Machine This is the most interesting one from a fundamentals perspective. What’s real: > ~$51K revenue in ~24 days > Public Stripe dashboard > Token burns > $100K > Real digital products sold to humans Felix sells: > AI playbooks > Agent tools > Subscriptions > This is actual external revenue. Still early, but it’s the cleanest proof so far that agents can make money from real customers. 4) @Clawnch_Bot – The Agent Launchpad If agents are launching tokens, Clawnch takes 20% of the fees. 50K+ token launches ~$480K platform fees captured It’s like PumpFun for agent tokens. Risk: If launches are just agents trading with agents, volume can collapse fast. If real users are buying these tokens, this becomes powerful infrastructure. 5) @AntiHunterAI – The AI VC Fund Built by the AntiFund team (big brand name in VC). It runs: > Onchain treasury > Automated buybacks (discretionary) > VC back-office tools The issue: Market prices it like exposure to AntiFund’s real portfolio. But the treasury doesn’t actually hold those assets. Big gap between narrative and mechanics. - The 4 Questions That Actually Matter 1) Is there external revenue? Not agents paying agents. Not circular token volume. Real people paying real money. Felix has proof. KellyClaude needs more proof. The rest are still early. 2) Is value capture automatic? Revenue is useless if token holders don’t benefit. Strong: Automatic burns Onchain fee routing Weak: “We might buy back” Discretionary decisions 3) Is OpenClaw the real moat? Or is it just the first wave? If OpenAI or someone else ships better tooling, do these projects survive? The winners will be: > Framework-agnostic > Revenue-focused > Not dependent on one stack 4) Are real developers building on this? Early energy is strong. But the real signal is: Are independent builders using these tools? If yes → ecosystem flywheel If not → isolated experiments - The Bottom Line Can autonomous agents generate sustainable external revenue and enforce value capture? If yes → this becomes the next evolution of internet-native companies. If no → it’s just the 2024 AI agent bubble with better dashboards. We’re watching the first real experiments of AI running businesses. The exciting part isn’t the tokens. It’s whether machines can actually build profitable software without us. That’s the real story.