Up Next: Path Robotics

Industrial applications of autonomous welding robots and artificial intelligence

Beginning

When you hear about a robotics and AI company these days, 90% of the time you’d expect it to be a self-driving or autonomous systems company, but Path Robotics is building a platform like no other. They’re automating welding, a seemingly mundane task, but a process used by millions across the world.

Although welding may seem boring, it is very hard. You have to set up rigs to hold parts in place, define a robust weld path to securely fasten parts, and spend a lot of time grinding down the welds. It’s a long and tedious task, but necessary when building robust and strong mechanical assemblies.

Manual welding is hard, but the introduction of large-scale industrial automation robots changed the game and made these processes a little easier. Because of repeatability in manufacturing, companies are able to automate some of the tasks by manually programming a weld path and building fancy robotic rigs to gather and hold parts together from any part of the line. While this significantly optimized the welding process it brought up a ton more issues. If the manufacturing line even changed an inch, the entire automation process would have to be re-programmed and calibrated to account for the new changes. Big manufacturers like Tesla tried to work around this flaw in current automation technology by hiring more human workers and although it saved time, it introduced a lot of errors and cost a lot of money to get things built.

Path Robotics has different plans for industrial automation. Their platform is automatically able to classify parts and figure out how to assemble them from a 3D scan. Once they have a simulated scan, their deep learning-based vision system is able to autonomously identify a weld line and continuously learn how to be precise using reinforcement learning.

Middle

Path Robotics is located in Columbus, Ohio… not your typical tech haven. At first, this made no sense to us — tech companies are all about attracting the best talent and being in the ‘hub’ of technology, but since Path Robotics is focused on welding it started to come together. Many manufacturing companies such as vehicle part OEMs are located in the Midwest, so it’s easy for Path to demo solutions and stay close to their customers.

Path’s focus on welding is also unique compared to other robotics companies as they aren’t replacing ‘existing’ worker’s jobs, but rather helping solve a shortage of workers and jobs. It’s becoming increasingly difficult to find good welders, with an estimated shortage of 400,000 welders by 2024.

Another interesting point to be made is that, unlike the other big players in robotics and AI, the founders of Path Robotics aren’t from Stanford, CMU, or MIT, but studied at OSU. Don’t get me wrong, OSU is still a top 20 school for engineering, but goes to show that to be a successful player in deep tech, you don’t have to chase the top 3 schools. The point that I’m trying to make is that it’s important for aspiring engineers and investors alike to recognize that talent is everywhere and that determining value from a team’s educational institution is a flawed process

End

In a technology world obsessed with autonomy, I’m surprised that Path Robotics wasn’t thrust into the spotlight earlier. They have a huge market opportunity because they have the ability to be integrated with any functional manufacturing line in the world with huge cost cuts from existing processes. Today’s industrial automation is antiquated and not dynamic at all. Think about Tesla, which makes changes to their cars every few months as they develop the technology. If they have to move any one of their manufacturing robots even an inch out of line, they have to rewrite code and change their manufacturing for the entire process. Path solves this, allowing companies like Tesla to move faster and build better cars. We love Path Robotics, but assessing an investment in hardware is a little more difficult. Turnaround time for building a product can take 4–5 times longer than software and even if a product gets built, hardware companies are typically worth less.

We are both broke, but we’d invest $4.50 of the 10 dollars in our bank account.