What failed startup Anki did for robotics

The well-funded passion project of Carnegie Mellon robotics experts, San Francisco-based Anki had a long-range master plan to progress from cute toys to sci-fi level robotic maids, security guards, and other servants. But after nearly a decade and over $200 million in venture funding, the furthest the company got–in terms of publicly revealed products–was a $250, palm-size toy that drives around making cartoonish faces and gibberish chitchat.

And now it won’t get any further. As first reported by Recode’s Theodore Schliefer, Anki is shutting down, bringing an end to the story of one of consumer robotics’ highest-profile startups.

It would be too simple to call Anki (a multi-year Fast Company Most Innovative Company designee) a flop, however. Its final product, Vector—which I wrote about last August—pioneered what Anki called emotional intelligence. It constantly surveys its environment, getting excited by new stimuli like lights, sounds, and human interaction–or depressed when they are lacking. Anki’s goal, CEO and cofounder Boris Sofman said, was to create robots that feel alive–including naturalistic imperfections, say, in how they move around–in order to appear more friendly and sympathetic to human companions.

Anki also helped teach new generations of roboticists. Vector’s predecessor, the $180 Cozmo, came with programing tools ranging from a simple drag-and-drop interface based on MIT’s Scratch Blocks, to a full software development kit in the Python programing language. It was even used in college-level robotics classes at the founders’ alma mater, Carnegie Mellon.

Cozmo appeared to be a commercial success, too, as the best-selling toy by revenue on Amazon in the U.S., U.K., and France in 2017. It was also featured prominently in once-revered Toys R Us. On the strength of Cozmo and its smart toy racing cars, car line, Drive and Overdrive, Anki claimed close to $100 million in annual revenue.

But that wasn’t enough to sustain Anki and its plan to progress from toy making to ever-more advanced (and larger) robots over the coming years. “We’ve always known from the beginning that this is not a toy company,” Sofman told me last year, mentioning Rosie the Robot from The Jetsons as a model. Rosie won’t come from Anki. But by inspiring and educating robot fans and researchers, the company may have made it more likely that others may build her.