From DePIN to Robots: How I Got Here and Where I Think This Is Going #

A builder’s perspective on Web3, AI, and why the next decade of infrastructure will look nothing like the last one.

I’ve been building stuff my whole life. I wrote my first program as a kid in Brazil, studied electrical engineering because I wanted to understand both the hardware and the software, built Arduino projects, hacked intercoms, assembled speaker cabinets from scratch, ran Helium miners on DigitalOcean before most people knew what LoRaWAN was. If there’s a wire, a board, or a protocol involved, I probably tried to make something with it.

This post is my attempt to explain how I ended up where I am today, what I learned along the way, and why I think the next big wave isn’t just AI or just crypto — it’s physical infrastructure powered by both.

But first, some context.

Finding Helium (and falling down the rabbit hole) #

Back in 2020, I stumbled upon Helium. A decentralized wireless network where regular people deployed hotspots and earned crypto for providing real coverage. I was immediately hooked. This wasn’t another DeFi fork or NFT project. It was real hardware, real signal, real utility. People were actually building a network.

I wrote tutorials about running miners, built DIY hotspots with Raspberry Pis and LoRa hats, and pretty quickly realized that the tooling around Helium was… well, it didn’t exist. So I co-founded Hotspotty to help people manage their deployments. What started as a side project turned into a real company, and I had a front-row seat to one of the most ambitious experiments in crypto.

We started calling this whole movement DePIN — Decentralized Physical Infrastructure Networks — because there wasn’t a name for what was happening. I built Depinhub to track the ecosystem as it grew from a handful of projects to hundreds.

So after years of being deep in this world, what did I actually learn?

The 3D chess problem #

Here’s what nobody tells you about building in DePIN. You’re not playing one game. You’re playing three at the same time.

Game 1: Build a real product. You need actual technology that works. Hardware that ships. Software that doesn’t break. Users who get value from what you built. This is hard enough on its own — most startups fail right here.

Game 2: Build a token economy. You need tokenomics that incentivize the right behavior without creating a speculative death spiral. You need liquidity. You need to navigate regulations that change every six months. You need to convince people that your token has value beyond “number go up.”

Game 3: Play both worlds simultaneously. Your Web2 users don’t care about your token. Your Web3 community doesn’t care about your product metrics. Your investors want both. Your team is stretched across two entirely different cultures, two different sets of expectations, two different definitions of success.

Most DePIN projects fail not because the technology doesn’t work, but because playing all three games at once is nearly impossible. You end up being mediocre at all of them instead of great at one.

I learned this the hard way. After years of building Depinhub, I realized we had fallen into the trap I was warning others about. We were becoming a marketing agency for DePIN projects instead of building something that stood on its own. So I made the hard call — shut it down and stepped back to think about what’s next.

What I still believe #

Despite everything, I’m not a DePIN skeptic. The core thesis is solid: token incentives can bootstrap networks that would be too expensive or too slow to build through traditional means. Helium proved it. The network exists. People use it. Real devices connect to it.

The problem isn’t the idea. The problem is execution, and the fact that most teams underestimate how brutally hard it is to serve two masters.

The projects that will actually win? They’re the ones building a product so good that the Web2 side works on its own — and using Web3 as an accelerant, not as the foundation. If your product dies without the token, you don’t have a product. You have a token.

Now, about robots #

While I was deep in DePIN world, something was happening in parallel that I couldn’t ignore.

The cost of robotic hardware is dropping fast. Foundation models (the same tech behind ChatGPT and Claude) are giving robots the ability to reason, plan, and adapt — without being explicitly programmed for every scenario. Companies are raising absurd amounts of money. Autonomous vehicles are racking up hundreds of millions of real-world miles. Pre-trained robot skills are becoming licensable platforms.

And here’s where it gets interesting for someone coming from my background: robots are physical infrastructure.

Think about it. A fleet of delivery robots is a logistics network. A fleet of autonomous vehicles is a transportation network. A swarm of agricultural robots is a farming network. These are all infrastructure problems — and they have the same bootstrapping challenge that DePIN tried to solve.

How do you deploy thousands of robots across a city? How do you coordinate them? How do you incentivize people to host, maintain, or operate them? If you squint, this looks a lot like what Helium tried to do with wireless hotspots, except the “hotspot” can now move, manipulate objects, and make decisions.

I’m not saying every robot network needs a token. That would be falling into the same trap again. But the intersection of decentralized coordination, AI-powered autonomy, and physical hardware? That’s where I think the next decade gets really interesting.

What I’m doing now #

I’m writing this post because I want to start thinking out loud. I spent years building tools and communities in DePIN, and I learned a ton — about what works, what doesn’t, and where the real opportunities are hiding.

I don’t have all the answers (who does?), but I have a perspective that comes from actually building in these spaces, not just tweeting about them. I’ve shipped hardware, written smart contracts, built SaaS products, shut down projects that weren’t working, and started new ones from scratch. Multiple times.

This Substack will be where I share what I’m thinking about: the convergence of AI, robotics, and decentralized infrastructure. The opportunities. The traps. The things I wish someone had told me earlier.

If you’re building in any of these spaces — or just curious about where this is all going — stick around. It’s going to be a wild ride.

Let’s build.