Back in 2014, we wrote about Baxter, an anthropomorphic factory robot that was gentle enough to work alongside human workers. We visited Baxter at a robotics trade show in New York, and it seemed like the company was getting a lot of interest from potential customers. Unfortunately, we learned last week that Rethink Robotics, the company behind Baxter, is shutting down.
Traditional factory robots are dangerous—they’re often put behind cages to avoid accidentally injuring human workers nearby. Baxter, by contrast, is designed to work directly alongside human workers. All of Baxter’s joints are designed not to pinch fingers. Its arms can “feel” if they encounter unexpected resistance (like a human body part) and stop. Company representatives liked to put their arms or heads in the path of the robot’s arms to show off this safety feature.
Traditional factory robots are also difficult to program. By contrast, almost anyone can learn how to teach Baxter a new task by grabbing its arms and guiding it through the desired steps.
The cost of traditional factory robots can easily run into the six figures. Baxter was initially priced at $22,000 in 2012 (the price rose a bit over time). Rethink Robotics also introduced a one-armed version of the Baxter, called Sawyer, that sold for $29,000.
Baxter seemed to have a bright future when it was unveiled in 2012. Rethink Robotics was cofounded by Rodney Brooks, who had previously cofounded iRobot, the maker of the popular Roomba robot vacuum cleaner.
Baxter and Sawyer faced strong competition, weak demand
So what went wrong? Rethink Robotics suffered from lackluster sales, due in part to strong competition. A Danish rival called Universal Robots has emerged as a leader in the market for these “cobots”—robots capable of interacting safely with human workers. The company said earlier this year that it had sold more than 21,000 robots over its lifetime.
The robots offered by Universal Robots are similar to Baxter and Sawyer in a lot of ways—they’re easy to train and designed to stop gently if they come in contact with unexpected obstacles. But one significant difference is that the Universal Robots models are smaller and lighter.
Rethink Robotics chose to make its robots look a lot like human workers. Baxter is designed to be mounted to the floor and has a big LCD screen that shows Baxter’s “eyes” while it’s doing its work. This makes Baxter seem more approachable, but it also means that it takes up a lot of space and weighed 165 pounds, without its base.
Universal Robots, by contrast, sells simple robot arms. The company’s large UR-10 robot weighs 64 pounds, while the smaller UR-3 weighs just 24 pounds. These robot arms were small and light and could be mounted to the floor, wall, or ceiling. Baxter draws attention to itself in ways that the UR robots don’t. In factories where space is at a premium, this gave UR an advantage.
Rethink Robotics recognized this issue and released Sawyer in 2015. It was 42 pounds with a payload of 4kg, compared with the UR-3’s 3kg payload.
Baxter also faced criticism for imprecise handling. “The accuracy and repeatability is not there,” one reviewer wrote in 2016, describing Baxter as “shaky.” Sawyer has a more rigid frame, yielding more precise handling, but it evidently wasn’t enough.
The larger problem may have been that the market for industrial robots just isn’t that big. Universal Robots has enjoyed healthy growth over the last decade, but, in the grand scheme of things, 21,000 robots just isn’t that many. In a truly fast-growing market there would have been plenty of room for a bunch of companies making human-friendly factory robots. But so far, at least, this is still a relatively niche market.
The failure of Rethink Robotics provides a valuable reality check for people worried about robots rendering human factory workers obsolete. There are hundreds of millions of factory workers around the world. By contrast, only 387,000 industrial robots were sold, globally, in 2017. That’s far too few to have a significant effect on the human labor force. Robot sales are growing at a healthy rate, though, so things might look different in a decade or two.