OpenClaw had an entire arc of becoming a main character of AI on X, concluding with the founder monetizing it by getting a job working for the OpenAI machine. In between, there were security vulnerabilities exposed and everyone’s enthusiasm was sedated for now.
What this claw saga showed us is that tools like OpenClaw and NanoClaw and apps like Moltbook have massive potential. These NanoClaws, I think, are the personal websites of the internet age or the fart apps of the mobile boom. It feels like a tad silly but there is something here that can be much bigger.
If I were trying to build a billion-dollar claw co, the first thing I would do is step back and make a platform for prosumers. Not hackers or even full-time SWEs—think of something a 24-year-old TPM at Applied Intuition would buy. Somewhere between a Hacker News-obsessed full-time Gen Z SWE and a Harvard MBA for a few hundred dollars. What I’m trying to highlight with this is that I would want to lower the barrier to entry to make an OpenClaw-like platform approachable. I think V1 can run on Mac minis, but I would go for something much smaller and cheaper. More than a $10 RISC-V board, so maybe a Raspberry Pi 5 AI Kit. Add a custom case and start building ‘the brand.’
If that doesn’t feel like a billion-dollar company, it’s because it isn’t. There’s a lot more to go. Next up is to make the barrier to entry low enough that a user asks, “Should I try this thing?” Here, I would need to focus on three specific topics: security, stability, and ubiquity (which comes through usability). This is roughly the WhatsApp playbook: when there were hundreds of messaging apps, they focused on core needs really, really well. In their case making the app bullet proof in many many different environments. If Claws are the new layer on top of agents, then my company should be the WhatsApp of Claws. The next piece would be solving the data integration problem. This is to answer the user’s question of how can I feed this autonomous agent (which is secure, stable and generalized) actual data for the thing I want to point it towards.
Ideally, I would want to use this momentum to build a small group of dedicated developers and tinkerers, and feed them. What they ask for, I would build. The focus here again is: “a simple, secure, stable platform that can be used for many, many things.” It will take a lot of trial and error to really make this platform stable and able to consume data in an easy way. I would try not to be limited to my own hardware over time; that’s just a vehicle to get it into the hands of early adopters who don’t want to hand-roll their own hardware environments. But who knows, that might be the right long term wedge. iPhone & iOS vs Android.
After this, I would have to do the most difficult thing in the venture: push to the mainstream. This is getting an accountant in Troy, Michigan, to use my thing. This won’t be easy because, unlike the early adopters who will buy an environment and platform to tinker with, mainstream users will only buy applications. Like in the App Store, I would be Apple. I would want the platform to be able to host autonomous agents, but really I would want to be the maker of the best mainstream autonomous agent. What that is, I don’t know. And I’m a bit skeptical I could shortcut everything above to get to it.
Only now would I raise a big round of financing. This is necessary because markets, whether we like it or not, want to pick winners. And then consumers converge on that winner. We did this at Applied. We’ve never spent any of the money we raised in nearly 10 years, but we keep raising because it accrues momentum and confidence in us. If I were someone famous like Karpathy or a multi-time exited founder, the shortcut is to do the $100M round on a $1B valuation. I’m sure there are at least a few VC grunts who are now thinking, “Okay, I can’t get to Karpathy, but who else could do this?” Now, doing this requires actually being technical and having some commercial instincts. But VCs are such lemmings that even if I didn’t have a hint of a commercial direction, they’ll probably still give me all the money I want. Remember, they need to wash money from LPs to justify getting more money so they can collect even more management fees.
Now, if I were not famous (in a technical and academic sense, e.g., published research), there are two other shortcuts to raise a big round here. One: have some well-known, previous “wins” under my belt (aka a track record over a long period of time). Two: have a bunch of traction. Since most of us are not in the first two categories of fame and track record, the last one is what I would have to do. I would start re-reading the above on how to build momentum.
Last, but not least (as I write this out while sitting at a coffee shop in Mountain View), I would avoid selling and see how far I can take it without letting the FAANG engine consume it like another infinity stone. We had such huge hopes in the mobile age. The birth of a new FAANG, maybe. But who won the mobile age? Google, Apple, and Facebook. Sadly, the same will probably happen in this round. Look at who owns the big LLMs. It ain’t the founders or the employees who have gotten massively diluted. But that’s a rant for another time.