I’m going to say something that might get me in trouble with my software friends: building a SaaS company is easy. Okay, not easy, but the playbook is written. Find a niche, build a product, get distribution, and scale. The physics are known. Hardware is a different beast entirely. It’s a world where the laws of physics trump the laws of Moore, and I learned that the hard way.
After two successful exits in the software world with RemoteTeam and MovieLaLa, I had a bit of capital and a lot of curiosity. I saw the explosion of AI and thought, "where does this meet the real world?" The answer, obviously, was robotics. So I jumped in headfirst, investing over $1 million across a portfolio of robotics startups. Some are building humanoid robots, others are working on autonomous drones, and a few are creating specialized machines for industries I barely knew existed. It felt like the Wild West again, a frontier where the rules haven’t been written yet.
It’s been a wild ride. And an expensive education. For anyone thinking about building or investing in this space, here are the ten biggest lessons I’ve learned.
1. Hardware is Hard. No, Seriously.
Everyone in tech parrots the phrase "hardware is hard," but you don't truly understand it until you see a six-figure prototype brick itself because of a faulty $2 capacitor. With software, a bad deployment can be rolled back in minutes. With hardware, a bad batch of circuit boards can bankrupt you. I remember one of our portfolio companies, a promising drone startup, getting a perfect demo ready for a major investor. The day before the meeting, the drone’s custom-designed landing gear, a piece of molded carbon fiber that cost $5,000 a pop, snapped during a routine test flight. Just like that. No quick fix, no patch. The team was devastated. The meeting was a disaster.
Your iteration cycles are measured in months, not hours. A simple change to a circuit board means a new design, sending it to a fab in China, waiting for it to be manufactured, waiting for it to clear customs, and then finally testing it. That’s a three-month cycle for something a software engineer could change in an afternoon. Supply chains are a constant nightmare. A single component shortage can derail your entire roadmap. It’s a brutal, unforgiving business that requires a level of patience and resilience that most software founders simply don’t have.
2. The "AI" is Often Silent
Every robotics pitch deck is dripping with "AI." Machine learning, neural networks, computer vision—you name it. But when you get under the hood, you find that 90% of the real-world problems are still mechanical and electrical engineering. Can the robot’s gripper actually pick up a variety of objects without crushing them? Can the motor run for 10,000 hours without failing? Does the battery last long enough to be useful?
I’ve seen teams with brilliant AI talent get stuck for a year because they couldn’t solve a basic thermal management problem. Their processor would overheat and throttle performance after just 20 minutes of operation. The fancy AI algorithms were useless because the hardware couldn't keep up. AI is the brain, but the body is what interacts with the world. And the body is where most startups fail. Don’t get me wrong, the AI is important, but it’s only one piece of a very complex puzzle.
3. The Demo is Not the Product
A slick demo is the lifeblood of a robotics startup. It’s how you raise money, attract talent, and get customers excited. But a demo that works once, in a controlled lab environment, is a world away from a product that works thousands of times in the messy, unpredictable real world. I’ve seen robots that can flawlessly fold laundry in the lab but fall to pieces in a real home with different lighting, fabrics, and floor surfaces.
One of the first questions I now ask is, "How many times has this demo run consecutively without human intervention?" If the answer is less than a hundred, they don’t have a product. They have a science project. Another question I love is, "What’s the mean time between failure (MTBF)?" If they don’t have an answer, it’s a huge red flag. It means they aren’t thinking about reliability, which is everything in hardware.
4. Unit Economics are a Physics Problem
In software, the cost of serving a new customer is close to zero. In robotics, your unit economics are governed by your bill of materials (BOM). Every screw, every sensor, every piece of plastic has a cost. And that cost doesn’t magically go down at scale unless you can commit to massive volumes, which is a huge risk for a startup.
I invested in a company making autonomous delivery robots. The tech was amazing, but the robot cost $20,000 to build. They were charging $5 per delivery. You can do the math. It was never going to work. They were so focused on the tech that they completely ignored the physics of their own business model. Now, I look at the BOM before I even look at the pitch deck. I want to see a clear path to a healthy gross margin, even at low volumes.
5. The Best Teams are Weirdly Diverse
My software companies were built by software engineers. My robotics companies are built by a motley crew of mechanical engineers, electrical engineers, software developers, industrial designers, and even materials scientists. You need people who can think in atoms and people who can think in bits, and they need to be able to speak the same language.
This interdisciplinary collaboration is incredibly difficult to manage, but it’s also where the magic happens. The best solutions come from the intersection of these different fields. I remember a team struggling with a robot arm that was too slow. The software team had optimized the control algorithms as much as they could. The mechanical team was convinced they needed a bigger, more expensive motor. Then the materials scientist on the team suggested a new, lighter-weight alloy for the arm itself. It cut the weight by 30%, allowing the existing motor to move the arm twice as fast. That’s the kind of breakthrough you only get with a truly diverse team.
6. Don't Bet on General-Purpose Humanoids. Not Yet.
I know it’s exciting—Tesla’s Optimus, Figure AI, the whole idea of a robot that can do anything a human can do is the ultimate sci-fi dream. And I’ve invested in a couple of them, more as a long-term bet on the future. My investments in companies like Anthropic, OpenAI, and Scale AI give me a front-row seat to the pace of AI progress. But the reality is, we are still decades away from a truly general-purpose humanoid robot being commercially viable.
The money, for now, is in specialized robots that do one thing exceptionally well. Robots that clean commercial floors. Robots that move boxes in a warehouse. Robots that inspect pipes. These are not as sexy, but they solve a real, immediate business problem. They are the workhorses of the robotics revolution, and they are the ones that are actually generating revenue today.
7. The Factory is the Real World
If you want to see the future of robotics, don’t go to a tech conference. Go to a modern factory. The level of automation is staggering. But it’s also a reminder of how structured and controlled the environment needs to be for most robots to function. The real world, meaning our homes, our cities, and our farms, is infinitely more complex.
Startups that embrace this complexity from day one are the ones that will win. The ones that test their robots in real-world conditions, with all the messiness and unpredictability that entails, are the ones that will build robust, reliable products. The lab is a necessary starting point, but you have to get out of the building as quickly as possible. One of my most successful investments is a company making autonomous tractors for farming. They spent the first year just driving their prototype through muddy fields in the rain, letting it fail, and learning from each failure.
8. Data is the New Oil, but Robots are the Drills
One of the most underrated aspects of robotics is the data. A fleet of robots is a mobile sensor platform, constantly collecting information about the physical world. A robot cleaning a supermarket floor isn’t just cleaning; it’s also creating a real-time map of the store, tracking inventory on the shelves, and identifying spills or hazards.
This data can be incredibly valuable, often more valuable than the physical task the robot is performing. The smartest robotics companies I’ve seen are not just selling a robot; they are selling a data service. They are using the robot as a means to an end, and that end is a unique and valuable dataset. They can sell this data back to the retailer to optimize store layouts, or to CPG brands to track how their products are displayed.
9. Full Autonomy is a Myth. Collaboration is the Future.
The goal of robotics is not to replace humans, but to augment them. The idea of a fully autonomous, "lights-out" factory or warehouse is a long way off. The most effective systems I’ve seen are collaborative, where humans and robots work together, each playing to their strengths. Humans handle the complex, edge-case tasks that require creativity and problem-solving, while robots handle the repetitive, physically demanding, or dangerous work.
This "human-in-the-loop" approach is not a compromise; it’s a more practical and powerful model. It leads to more flexible, resilient, and cost-effective automation. The future is not humans or robots; it’s humans and robots. The best companies are building tools that make human workers more productive, not trying to engineer them out of the equation.
10. It’s All About the People
This is true for any investment, but it’s especially true in robotics. Given the long timelines, the technical complexity, and the sheer number of ways a hardware startup can fail, you are betting on the team above all else. Are they resilient? Are they resourceful? Are they obsessed with solving the customer’s problem?
I’ve learned to look for founders with a healthy dose of paranoia and a deep understanding of the physics of their business. The ones who are not just in love with the technology, but with the messy, difficult, and ultimately rewarding process of bringing a physical product to life. I’ll take a scrappy founder who has spent months in a Shenzhen factory over a PhD from a top university with a perfect simulation any day of the week.
Investing in robotics has been a humbling experience. It’s taught me to respect the complexity of the physical world and the ingenuity of the people who are building the machines that will shape our future. It’s a long game, but I’m more convinced than ever that it’s a game worth playing.
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.