Robo Strategy (BOT) is one of the first publicly traded venture funds. It backs companies building in the physical AI space: humanoid robots, general-purpose robots, and the supply-chain firms that make the parts robots need to be assembled.
Why general-purpose robots beat single-task machines
The advantage of general-purpose and humanoid robots follows the same logic as the biggest companies in the world. Apple (AAPL) makes the iPhone, a general-purpose device: it calls people, sends messages, and takes pictures. A camera does one job well, but a device that does many jobs at scale is far more useful and reaches economies of scale that a single-task machine cannot.
Nvidia (NVDA) shows the same pattern. The GPU is a general processing unit. ASICs, chips built for one specific algorithm, also serve a massive market. GPUs stay valuable because they adapt as algorithms change, and that adaptability is a main reason Nvidia (NVDA) is one of the biggest companies today.
The same idea extends to humanoid robots. They are made for a world built for humans, they adapt to changing environments, and one machine can serve hundreds or thousands of different applications.
Telling a real breakthrough from a good demo
The robotics space has "shiny penny syndrome" and many ambitious claims. Impressive demos do not always match a robot's real capability. Videos show robots like Unitree doing backflips or moving a box from one spot to another, but the real question is whether a company can make a robot do a given task reliably. A robot that succeeds 50% of the time is not useful; it needs to be right close to 100% of the time.
Testing this means visiting the companies and trying to interfere with the robot. Move the target destination and see if the robot still performs. Change the object it works with and see if it adapts. Change the lighting and see if it adjusts. Adaptability is what makes humans so useful for physical work, so robots must be tested for it, both in a lab and in the real settings where they will be deployed: factory lines, restaurants, and logistics operations.
Some companies need a long pilot, and if that drags on, they may not be ready for deployment. Others get past the pilot stage and show they can match human-level throughput at high quality, with a very low failure rate, no human intervention, and fully autonomous operation. Those are the companies worth investing in.
The Maven Robotics investment
Robo Strategy (BOT) recently invested in Maven Robotics. The team spun out of Apple's (AAPL) autonomous vehicles program, which Apple spent close to $10 billion on over about a decade before disbanding it, likely for internal political reasons. That program recruited top talent across autonomous vehicles, robot learning, and controls. Much of that team, under head of engineering Hamza, left to found Maven.
The key lesson from Apple (AAPL): systems must run perfectly, because real people sit in these vehicles and any error puts a human life at risk. That same perfection is needed for robots in production. Faulty robots endanger humans, and a robot that is right 90% of the time is not very useful. It needs to be right 99.99% of the time, as close to 100% as possible. Maven has reached that level in customer settings already, which is why the fund backed them.
