Path Robotics Announces $100M in Funding: Implications for the Automation Industry
5 min read
When Path Robotics announced that they secured $100M in Series C funding from investors that include Tiger Global and Silicon Valley Bank, we knew that this story was bigger than the success of an individual company. It is a vote of confidence in Path's product, an autonomous welding robot that combines computer vision, artificial intelligence, and motion planning technology to deliver a truly agile robotics platform that overcomes many of the problems that we encounter with today’s industrial robots.
While this step forward has implications for the overall automation industry, it couldn’t come at a better time for welding specifically. Research from the American Welding Society says that the U.S. will see a shortage of about 400,000 welders by 2024, as demand for welders grows 4% annually while the welding workforce shrinks by 7%.
Robotic welders are the answer. Not only do they boost efficiency by working around the clock, but they deliver consistent quality, improve safety conditions on the shop floor, and reduce mistakes, which in turn reduces scrap volumes. In fact, robotic welders already make up 29% of all robotic applications in industry, second only to material handling.
So, if welding robots are already so prominent, why is Path’s technology a big deal? To answer this question, we need to look at the current state of the art, assess its limitations, and understand how the Path robot overcomes them.
The Current State of Robotic Welding
After George Devol invented the programmable robot in 1954, the first commercial application was welding. Called the Unimate, it performed hot die-casting at a GM plant in New Jersey. The moral of this story: besides being an obvious use case for industrial robotics, welding also has a long history of serving as an innovation forerunner.
Welding robots have obviously come a long way over the decades, but you might be surprised how similar they remain to the earliest prototypes. In a word, they’re inflexible. They’re highly calibrated through a long and arduous programming process to handle exact circumstances and a small number of errors. Since they’re “blind”, these machines do not respond well to unexpected inputs, such as varying part sizes or misalignment. Human welders, on the other hand, can adapt on the fly to make a good weld.
Besides not being able to deal with unpredictable circumstances or avoid obstacles, the status quo simply cannot adapt to the rapidly changing requirements of today’s factories. If the line changes even an inch, the entire automation process requires reprogramming and re-calibration. The result, according to the National Institute of Standards and Technology (NIST), is that 45 to 60 percent of the cost of deploying a robot goes towards programming.
So, although current welding robots are able to achieve impressive levels of automation, they still have some serious flaws. These shortcomings are especially prohibitive for small to medium size enterprises that manufacture products in the hundreds to thousands range rather than at the scale of hundreds of thousands per production run. It’s also a problem for agile manufacturing techniques like just-in-time manufacturing because adapting to change is so difficult and expensive.
The Next Generation: Intelligent Robots with Eyes
Path Robotics and other similar vendors like Soft Robotics are changing the paradigm with smart robotics that use computer vision and artificial intelligence to achieve unprecedented levels of flexibility, agility, and quality. Basically, instead of having to make sure everything is just right, we can now count on robots to figure it out themselves and to improve over time.
One of the major draws to this approach is that it doesn't require any coding. Rather than programming each individual automation task in painstaking detail, operators can simply load a CAD file into the robot and select and apply parameters. In addition to drastically reducing the time, effort, and cost of getting started or updating a production line, these robots can still handle complex fabrications and intricate details without needing any perfect parts.
In addition to autonomously setting weld paths, the robot also collects quality information as it welds so that it can adjust in real-time and use reinforcement learning to make itself better over time. One of the hallmarks of a truly intelligent machine learning platform is its ability to get better as it iterates, and the Path system fits this criterion.
What all this amounts to is agile robotics. According to NIST research, all agile robotics systems share three key traits:
The ability of a robot to be rapidly retasked without the need to shut down the robot for an extended period when a new operation needs to be performed.
The ability of a robot to recover from errors, so that when a part is dropped, for example, the robot can assess and determine the best way to proceed to accomplish the goal.
The ability to quickly swap in and out robots from different manufacturers, so that a company is not tied to a single robot brand.
While it’s the synergy of multiple technologies that unlocks these capabilities, vision is the key differentiator that sets Path’s robots apart from others. It’s the same ability that enables human welders to interpret unpredictable situations. Even though other computer vision systems are currently available on the market, they’re very expensive and not easy to set up, meaning that the vision vendors highly recommend that they do it or that you hire a preferred vision integrator. Path’s product changes the equation by coming pre-equipped with state of the art vision capabilities.
Conclusion
Just like with the first Unimate industrial robots, Path’s welding solution is just the beginning in an exciting new wave of robotics innovation. Welding is the perfect place to start because it’s both a proven application for automation systems and because of the growing labor shortage, but it’s far from the final destination.
Having this product in the marketplace, one that doesn’t require continuous training for different situations or reprogramming for new part types, is a big difference from what’s currently available. It’s poised to not only influence manufacturing but also other industries like the sensitive assembly of semiconductors and the delicate process of harvesting crops.
For companies that want to improve their automation posture and realize a substantial ROI, knowing where to start and how to best implement these cutting edge technologies is essential. That’s exactly why it’s so important to work with a trusted automation systems integrator.
Outlier Automation is that trusted partner for so many companies because we understand where the automation industry is headed, what the current limitations of even today’s best technologies are, and how to create a comprehensive automation strategy that solves our clients’ unique needs. To learn more about how we can help you bring your factory into the future, contact us today.