I once had a founder pitch me a robot that could make a perfect cup of coffee. The presentation was slick, the animations were beautiful, and the team was brilliant. But I passed. Why? Because they’d spent all their time on the software and almost none on the cheap, unreliable plastic arm that was supposed to handle the beans. They forgot that in robotics, the ‘physical’ part of Physical AI is where most startups die.
I’ve been a serial entrepreneur in Silicon Valley for most of my life. I’ve had a couple of successful exits—RemoteTeam, which was acquired by Gusto, and MovieLaLa, acquired by Gfycat. Now, I spend a lot of my time as an angel investor, and I’ve put over $1 million of my own money into robotics and physical AI startups. It’s been a wild ride. I’ve seen brilliant successes and spectacular failures. And I’ve learned a few things along the way.
Here are ten of them.
1. It’s Not Sci-Fi Anymore
For years, “robotics” was a word you’d hear in a university lab or a sci-fi movie. Not anymore. We’re at a tipping point where robots are leaving the lab and entering the real world, doing real work. The most obvious place you see this is in warehouses. The demand for getting packages to your door in two hours isn’t going away, and humans alone can’t keep up. Companies are building fleets of warehouse robots that sort, pick, and move goods 24/7 without getting tired. This isn’t a far-future idea; it’s happening right now and it’s a massive market.
2. The ‘Physical’ in Physical AI is the Hard Part
I see so many teams with incredible AI talent get this wrong. They build these genius brains in a simulation but have no idea how to build the body. Hardware is hard. It breaks. It has supply chain issues. It doesn’t iterate as quickly as code. That coffee robot I mentioned is a perfect example. The software was 99% of the way there, but the hardware was a toy. I invest in teams that have a healthy respect, and preferably deep experience, for the grimy, frustrating, and expensive world of hardware.
3. Vertical Integration is a Powerful Moat
In a field this new, you can’t just buy off-the-shelf parts and expect to build a revolutionary product. The most promising companies are the ones building the whole stack themselves. Look at Tesla with Optimus or Figure AI. They are designing the hardware, the actuators, the sensors, and the AI that runs on it. This is incredibly difficult and capital-intensive, but it creates a huge competitive advantage. When you control the whole system, you can make it all work together in a way that no one who is just assembling parts can match.
4. Data is the New Oil, Especially Here
This is a cliché in software, but it’s even more true in robotics. The only way to make a robot smart is to feed it massive amounts of real-world data. A drone AI doesn’t learn to inspect a wind turbine by flying in a simulator. It learns by flying thousands of real missions and learning from its mistakes. The companies that win will be the ones that figure out how to deploy their robots and start collecting data as quickly as possible. That data becomes a flywheel; more data leads to a better AI, which makes the robot more useful, which leads to more deployments, which generates even more data.
5. The Team is Everything
Another investor cliché, but I can’t stress this enough. I’m not investing in a robot; I’m investing in the people who build it. Building a robotics company is a long, brutal journey. I look for founders who are resilient, obsessive, and have a mix of skills. You need the hardware guru, the AI genius, and the business person who can actually sell the thing. Without that combination, even the best technology will fail.
6. Chase the ‘Unsexy’ Problems
Everyone wants to build a robot butler. It’s a fun idea, but it’s not where the real money is, at least not yet. The biggest opportunities I see are in boring, industrial applications. Warehouse automation, agricultural robots, construction site drones—these are multi-billion dollar markets, and they are desperate for solutions. These aren’t the robots you see on the cover of a magazine, but they are the ones that will build massive companies.
7. Autonomy is a Spectrum
Full, human-level autonomy is the end goal, but you don’t have to start there. There’s immense value in building robots that are partially autonomous or have a human-in-the-loop. Think of autonomous vehicles. We don’t have Level 5 self-driving cars everywhere yet, but the Level 2 features like adaptive cruise control are already common and incredibly useful. The same is true for robotics. A robot that can do 80% of a task and have a human handle the tricky 20% remotely is still a huge win.
8. Your Go-to-Market is as Important as Your Tech
I’ve seen brilliant robots that couldn’t find a single customer. The founders were so in love with their technology that they never stopped to ask who would actually pay for it and why. You have to understand your customer’s pain point better than they do. Are you saving them money? Are you making their workplace safer? Is your robot more reliable than human labor? If you can’t answer those questions with hard numbers, you don’t have a business.
9. The ‘Humanoid’ Form Factor is Not a Gimmick
For a long time, I was skeptical of humanoid robots. They seemed inefficient and overly complex. Why build a robot with two legs when wheels are so much easier? I was wrong. The world is designed for humans. Factories, warehouses, and homes are all built around the human form factor. A robot that can walk, climb stairs, and use human tools doesn’t require us to redesign the entire world to accommodate it. That’s a powerful advantage, and it’s why I’m so excited about companies like Figure AI.
10. It’s Still Day One
For all the progress we’ve made, we are at the very beginning of the robotics revolution. The next ten years will see more change than the last fifty. There will be huge successes and spectacular flameouts. It’s going to be a bumpy ride, but I’m convinced that the founders who are building these companies today are creating the future. If you’re a founder in this space, my advice is simple: be bold, focus on a real problem, and don’t forget about the hardware. The world is waiting for what you’re building.
Frequently Asked Questions
How do I know which items apply to my situation?
Start by honestly assessing where your biggest bottleneck is right now. The items that address that specific constraint will give you the highest return on your time and energy.
Are these recommendations still relevant in 2026?
Absolutely. While specific tools and tactics change, the underlying principles remain consistent. I update my thinking regularly based on what I'm seeing in the market and across my portfolio companies.
Which item on this list has the highest impact?
It depends on your stage and context, but in my experience, the items near the top of the list tend to have the broadest applicability. That said, sometimes the less obvious items create the biggest breakthroughs for specific situations.