Connecting robots to artificial intelligence in the cloud can reduce onboard processing and power demands. CloudMinds Technology Inc., which offers cloud-based systems for robotics, last week filed with the U.S. Securities and Exchange Commission for an initial public offering of stock worth $500 million. The company has headquarters in Beijing and Santa Clara, Calif.
In its SEC filing, CloudMinds claimed to be the world’s first company “to commercialize products and services related to the end-to-end cloud robot system.” It cited a Frost & Sullivan report estimating that the global market for cloud robotics will grow from $75.5 billion last year to $103 billion in 2023.
CloudMinds connects brains for service robots
CloudMinds said it “makes robots smarter providing the secure virtual backbone network coupled with a real-time, multi-modal, human-in-the-loop cloud brain platform to connect different types of service robots.”
The “cloud brain,” or Human Augmented Robotics Intelligence with Extreme Reality (HARIX) platform, is capable of operating millions of connected robots conducting different tasks simultaneously, said CloudMinds. The AI system includes capabilities such as computer vision, natural language processing, and vision-controlled manipulation, aided by human supervision and input.
HARIX is connected to smart devices and robots via CloudMinds’ Virtual Backbone Networks (VBNs). The “nerve network” is built on 4G and emerging 5G network infrastructure, primarily in China, the U.S., Japan, and Europe.
Onboard the robot, the Data A1 smartphone, which serves as a robot control unit (RCU), processes data locally before transmission to the cloud and receipt of commands. Smart Compliant Actuators (SCAs) from INNFOS Technologies Co. include embedded AI processors and multimodal sensors for human-like motion, said CloudMinds.
Honing humanoid service robots
Sales of service robots last year rose by 39% to $6.6 billion, said the International Federation of Robotics.
The need for service robots is growing because of aging populations, rising labor costs, and ongoing demands for greater productivity, said Frost & Sullivan. It predicted that there will be 42 million service robots by the end of 2020 and that personal robots will be a $19 billion market opportunity by 2020.
CloudMinds’ portfolio includes CloudPepper, a humanoid robot from SoftBank Robotics using its cloud-based AI. Other robots include cloud-based vending machines, security robots, and the Cloudia virtual AI robot.
In February, CloudMinds combined its technologies in the XR-1 humanoid service robot. At Mobile World Congress 2019 in Barcelona, co-founder and CEO Bill Huang described the general-purpose robot as a step toward toward general-purpose “family nanny robots by the year 2025.”
“Based on a set of advanced technologies, CloudMinds will develop a number of intelligent compliant service robots, from wheeled to two-legged,” Huang said. “Leveraging the 5G network’s high bandwidth and low latency, our cloud AI platform will rapidly grow its intelligence as well as its deployment size. We will usher in a new era of service-oriented robotics.”
In April, CloudMinds released the XR Robotics Development Kit to enable programmers, developers, and end users to create skills for the XR-1 and subsequent robots in the series. The Robotics Development Environment includes a HARIX digital twin and uses deep learning to acquire and test skills that can later be transferred back to physical XR-1 robots.
Banking on robot brains
CloudMinds was founded in 2015 and has close to 700 full-time employees. Citigroup, JP Morgan, and IBS Investment Bank are bookrunners in the IPO, said Xinhua News.
This past spring, CloudMinds closed its $186 million Series B fund-raising round, according to Crunchbase, and it raised a total of $300 million with investment from the SoftBank Vision Fund. Other investors include Foxconn, Keytone Ventures, and Walden International.
Filed Under: AI • machine learning, The Robot Report, Wireless, Robotics • robotic grippers • end effectors