Trying to make the iPhone of AI robotics

by

Roobo, maker of artificial intelligence robots, raised $53 million in a Series B funding round led by Seven Seas Partners, a Chinese VC. This is in addition to the $100 million raised last year. The managing partner of Seven Seas, Jeff Xiong Minghua, who was Tencent’s CTO for eight years, will become Roobo’s new CEO.

They make the robotic operating system ROSAI, a language processing chip, and service robots PUDDING S, PUDDING BEANQ, JELLY, and DOMGY. The PUDDING series is Roobo’s first home education and companion robot system. It is designed for kids and the two PUDDING models have sold over 100,000 units.

For the Pudding Educational social robot they have partnered with Princeton University to make a system that will help with a child’s educational development.

At the end of October, 2017, Amazon announced Amazon Alexa voice assistant sales had passed 20 million.

Amazon Echo sales to date have totaled 19.5-21.5 million units. Non-echo smart speakers with Alexa are estimated to be in the 1 million unit range through Q3 2017.

75% of the U.S. market for these Internet-connected devices, which people can use with their voices to order take out food, check their bank accounts, check the weather, and play music. Google Home is second with about 24% of the market.

Apple has sold more than one billion iPhones worldwide from 2007 to 2016.

Apple’s total iPod sales passed 10 million from 2001 to 2004. 8.2 million units sold in 2004 alone. So Amazon Alexa could be in the 2004-2005 timeframe of where iPod was.

Significant sales but not mainstream yet

As of 2017 23% of vacuum cleaners are robots.

The worldwide number of domestic household robots will rise to 31 million between 2016 and 2019. The sales value of robots cleaning floors, mowing lawns, and cleaning swimming pools will grow to about 13 billion US dollars in this period.

In the coming years, vacuum and floor cleaning robots will continue to dominate household robot markets. Sales will rise from 3.6 million units (2015) to around 30 million units within the 2016-2019 forecast period. Vacuum and floor cleaning robots account for 96 percent of domestic robot sales. Robotic mowers and pool cleaning robots rank second and third.

Tens of millions of sales are significant but not the billion level of iPhones and smartphones. Smartphones are must haves. Roombas are nice to have.

ROS.AI is an artificial intelligence solution platform to provide AI interaction capability overall solution for equipments.
ROS.AI provides developer with hardware module, software system and various artificial intelligence services, and is a more comprehensive and flexible open platform and a solution platform.

Seven Seas has invested in a variety of robotic and consumer product startups including RightHand Robotics and Lily. The reason Seven Seas put such a big bet on Roobo is that Xiong thinks it will become one of the leading companies in IoT (Internet of Things). Xiong thinks that the next-generation of AI Internet companies need to know how to integrate products with services. Roobo is trying to be the leading company in the area of AI Robotics.

Roobo has partnered with Nuance to launch multilingual versions of their robots.

10 thoughts on “Trying to make the iPhone of AI robotics”

  3. Just looking at those hideously cute abominations, I see what’s meant by “the IPhone of AI Robotics” is a product where design trumps functionality at great cost.

  4. Then just focus on treating every household task it’s like own program. ‘Empty bins’ would be a useful feature to add, so the programmers would have to spend time adding to the robots object recognition various types of bins, add various programs for how to identify places to hold said bins while carrying them, how to open them, etc.
    Then just incrementally add more and more features with each software update. The AI doesn’t have to as smart as a human, it doesn’t have to be ‘general purpose AI’ that can solve any problem. Just focus on small tasks, solve each one with a new subroutine, http://toyotadimalang.com push it out with an update, and keep going until the robot slave is capable of performing most household tasks. That kind of AI is never going to turn into skynet on us, we’ll never be able to hold a conversation with it, but it would at least be practically useful.

  6. I think we could have home robotics today if we started more simply.

    Start with a robot about the size of a short human, with arms, legs, plenty of cameras, powerful battery, and a self recharge station.

    Then, just start adding functions. Automatic mapping of an environment would be obvious, and pathfinding to find the recharge station, balance and recovery from a fall too. “Come here”, “Stay” and “Follow Me” would be obvious first new features added after that. Next, code to identify solid objects, how to pick them up safely without dropping them, etc.

    Then perhaps later on, add to the robot’s software a general purpose object recogniser that can take in images of environments and tag various sections of the image with guesses on what kind of object something is. Something like that could power a ‘Go find me this object’ feature, like, ‘Go get the old clock from the living room’, with a map of the house and object recogniser, the robot could go find the clock and bring it to the operator.

    Then just focus on treating every household task it’s like own program. ‘Empty bins’ would be a useful feature to add, so the programmers would have to spend time adding to the robots object recognition various types of bins, add various programs for how to identify places to hold said bins while carrying them, how to open them, etc.

    Then just incrementally add more and more features with each software update. The AI doesn’t have to as smart as a human, it doesn’t have to be ‘general purpose AI’ that can solve any problem. Just focus on small tasks, solve each one with a new subroutine, push it out with an update, and keep going until the robot slave is capable of performing most household tasks. That kind of AI is never going to turn into skynet on us, we’ll never be able to hold a conversation with it, but it would at least be practically useful.

  7. If they aren’t going to do anything useful with the body, why not just have a projector and throw a personalized 3d modelled image of your home assistant on the wall? Anything from Rosie the Robot to a movie star, cartoon or favorite animal friend. Add a store that sells templates of famous people (appearance, voice, mannerisms, personality, etc), with subscription options for services that make your copy track the real person – wearing the new clothese they bought that day, talking about events that happened to them, etc. The more real-time and detailed, the more expensive.

  8. While cool, domestic robots still are very limited by battery power and abilities. I think the next big robotic platform are SDCs, which aren’t as limited and improve into something that has an actual social role already in our world: cars. Our cities are made for them and they will be able to freely move around them, whenever law abides.

    The other big robotic market with big strides into our lives, are domestic voice assistants Those integrated with/into SDCs and probably drones, will bring the biggest market synergies and opportunities IMO.

    I’m sure Waymo’s cars will have very good integration with Google Now. To the point of it being ‘the’ voice and emotional equivalent of a personality of the car for many users.

    But it won’t be the only one. Alexa’s and Siri’s owners are surely looking for partnerships with other SDC developers right now.

Comments are closed.