fb-pixel Skip to main content
Scott Kirsner | Innovation Economy

Businesses want robots that are problem solvers

Rethink Robotics’ Baxter lost out to a faster, simpler, and smaller competitor from Denmark.Image via Clearpath Robotics

On a recent Wednesday morning, I donned a virtual reality headset to look at the future of the robotics industry in Massachusetts. Inside the Financial District offices of the architecture firm Silverman Trykowski, I stood in place and used a handheld controller to navigate around a three-dimensional model of a shared robotics lab called MassRobotics, which is planning a major expansion in 2019.

The same day, news broke that one of the best-funded robotics companies Boston has ever seen, Rethink Robotics, was closing its doors. Rethink had raised nearly $150 million in capital over a decade of existence, and its first investor was Amazon CEO Jeff Bezos.


While robots may be a big part of the future, building and selling them in the present can be exceptionally difficult.

Rethink started in 2008 with a compelling and ambitious vision. Just as Apple and IBM in the 1980s introduced an affordable and easy-to-use new tool to offices with their personal computers, Rethink wanted to do the same for manufacturing facilities that couldn’t previously afford robots or considered them too complicated. Rethink had a CEO, Scott Eckert, who came from Dell Computer, a company that knows a few things about mass-producing and mass-marketing affordable tech products. Its chief technology officer and founder, Rodney Brooks, was cofounder of Bedford-based iRobot — maker of the Roomba vacuum cleaner — and a world-renowned MIT robotics professor.

Rethink unveiled its first product, a red, two-armed humanoid robot named Baxter, in September 2012. (OK, it was really a robotic torso: Baxter didn’t have anything going on below the waist.) The pitch was that by using an iPad-like touchscreen and moving Baxter’s arms, you could easily train it to perform a wide range of repetitive tasks. It could also work safely around humans, stopping the machine’s movements instantly when it bumped into something or someone. The bot’s starting price was $20,000, at a time when many industrial robots cost $100,000 and up.


It felt like a revolutionary moment in robotics. Baxter and Rethink attracted lots of media attention around the world. In February 2017, Baxter showed up at MassRobotics in the Seaport District for a ceremonial ribbon-cutting when the shared workspace was opening. Boston Mayor Marty Walsh held the ribbon, and Baxter (safely) wielded the giant scissors.

So what happened?

“Customers don’t really want robots per se, what they want is solutions to problems,” says Clara Vu, an early consultant to Rethink who is now cofounder of Veo Robotics in Cambridge. Baxter may have been sophisticated, but it was not a relevant solution to enough problems that manufacturing firms had.

When Rethink began to sell Baxter, the company found itself competing with a simpler, faster, and smaller product made by Universal Robots of Denmark, the inelegantly named UR5. The UR5 wasn’t as easy to program, but it could perform tasks and movements with more precision than Baxter — which turned out to be important to customers.

“What I’d say about Universal is that they innovated where it counted, and left well enough alone where it didn’t,” Vu says.

Rethink was also slower than Universal to build a network of distributors and resellers who would help sell Baxter to their customers, says Frank Tobe, the founder of the trade publication The Robot Report. That helped Universal build its sales momentum. “Based on all their sales,” Tobe says, Universal “began to advertise how short the return-on-investment was on their robots,” even though the company’s bot was slightly more expensive at the outset.


Rethink knew that the precision and accuracy of Baxter’s movements “was not great,” says Julian Weinstock, a former product manager at the company, mainly because of the use of lower-end joints and gearing to keep Baxter’s price low. The hope was that “we could compensate with smart software for what was a hardware problem.” But “that didn’t pan out to the degree that was necessary,” he says.

Rethink introduced a second product in 2015, a one-arm bot dubbed Sawyer that solved some of the issues related to accuracy and the ability to repeat the same motions over and over, “but it was too little, too late,” Weinstock says. “By that time, Universal had established a strong presence in the industry.”

Rethink had hoped to discover a “killer app” that would dramatically expand the use of manufacturing robots, Weinstock says. It ran out of money before it got there — though it was popular with academic researchers who wanted an inexpensive robot to help run experiments.

In 2015, Universal Robots was purchased by Teradyne, a maker of automation equipment headquartered in North Reading. The price: $285 million. The Universal Robots business generated $172 million in revenue in 2017 and “is solidly profitable,” Teradyne vice president Andrew Blanchard says. In February of this year, Universal opened an office in the Seaport District and began hiring engineers in Boston. That site, Blanchard says, should employ about 50 people by year’s end. (It’s only a few blocks away from Rethink’s office.)


Meanwhile, back in virtual reality, things were looking good as I ambled around to check out the future machine shop, conference rooms, and event space at the digitally modeled MassRobotics facility. (My tour lasted about seven minutes but ended abruptly when virtual-me accidentally fell into an elevator shaft and couldn’t get out.)

The MassRobotics space that Baxter helped to inaugurate in 2017 is now home to about 32 companies, executive director Tom Ryden said. It’s full. The expansion will add another 25,000 square feet to the facility, giving startups more room to grow, as well as access to a wider range of tools, equipment, and six robots to test on and develop with. (Among them are one Baxter and one Sawyer.)

The biggest challenge right now, Ryden says, is finding more grants and sponsorship for the build-out. “The price of construction has gone up since we started the planning process,” he says. But the hope is for the additional space to be open sometime between February and April next year, making room for even more robotics ventures.

“Despite some failures here and there,” Weinstock says, the robotics sector in Massachusetts “is still very exciting and dynamic.”

Neither of Rethink’s top two executives wanted to talk for this story, but CEO Scott Eckert told me by e-mail that he has been working to help approximately 90 former employees find new jobs. “Any company would be lucky to have this talented group of people,” Eckert wrote. “We were certainly lucky to have them for the time that we did.”


Scott Kirsner can be reached at kirsner@pobox.com. Follow him on Twitter @ScottKirsner.