CloudMinds, a Chinese startup company, has raised $100 million in Series A venture funding with investors including Softbank China Capital, Keystone Ventures, and Walden Venture Investments. This is after raising an impressive $30 million seed round in May 2016 from Softbank and Foxconn. CloudMinds is focused on delivering an AI-based cloud intelligence solutions for robots, using secure low-latency communications technology.
Blockchain is part of the authentication mechanism being used, which is currently patent pending. Blockchain technology is a cryptographically secure, open ledger system that records every transaction taking place for an asset. In the case of Bitcoin, the asset is currency and in the cloud robotics case, the asset is the authentication mechanism itself where end nodes or robots can connect to the cloud and exchange credentials before the robot gets access to the intelligence tasks and trained models. The use of blockchain in the cloud robotics context is new and should give CloudMinds an edge in terms of scaling its authentication system in a low-cost and efficient way.
CloudMinds is the brainchild of Bill Huang, who served as the General Manager for China Mobile Research Institute between 2007 and 2015, and was responsible for one of the first cloud-based radio access network (RAN) deployments for China Mobile, among many other achievements. He was also CTO and co-founder at UTStarcom, a telecommunications equipment provider and innovator in mobile soft switches. With Mr. Huang’s extensive telecom background, CloudMinds brings a fresh approach to cloud robotics, suggesting that cloud robotics is as much as a networking problem as it is a robotics and artificial intelligence (AI) problem. CloudMinds is also involved in the Cloud Robot program being pursued by the GTI, which is the industry consortium to expand TD-LTE and 5G. The use of 5G in cloud robotics makes sense as 5G brings high-bandwidth, low-latency, and high-throughput connectivity.
Although we have seen cloud robotics as a research area being pursued by companies like Toyota, Google and OpenAI, CloudMinds is possibly one of the first commercial attempts at creating a cloud-based intelligence platform for robots. In one of my earlier blog posts, I explored the use of training environments to speed up the testing burden of robots, and to benefit cloud robotics. Rather than training one robot at a time, open-source environments like Universe can be used to train a large sample of household robots by having a cloud-based intelligence watch a video feed of humans performing household chores, and then have household robots be instantly updated in the field with new and improved skills. CloudMinds is taking that architecture one step further, adding enterprise-grade connectivity and allowing for robotics manufacturers to have access to cloud-based intelligence. With most traditional robot manufacturers being new to AI, CloudMinds can provide an AI engine that runs in the cloud, whether it is for vision-based intelligence, language-based intelligence, or any other AI feature set. It can also be extended to other capabilities like virtual reality (VR) and augmented reality (AR). CloudMinds has already announced a partnership with mobile device maker LEIA to offer 3D holographic experiences using its Mobile-Intranet Cloud Services (MCS) platform. Another partnership that has been announced is with Immerex Inc., a VR headset manufacturer that will leverage CloudMinds Human Augmented Robot Intelligence (HARI).
While the mobile and VR device partnerships are good low-hanging fruit, the real test of CloudMinds is in convincing consumer robot, enterprise robot, drone, and autonomous vehicle manufacturers to trust its technology. Being a Chinese company with investors like Foxconn gives them unprecedented access to emerging hardware and robotics startups and the latest innovation from the world’s hardware hub, Shenzhen. CloudMinds has a Silicon Valley and Japanese presence, which suggests that it isn’t limiting itself to the Chinese market, although it will be surprising not to see U.S.-based or European competition emerge over time. A few years back, one would have expected such innovation to only emerge out of Silicon Valley or Japan, but with AI and robotics we have seen a levelling of the playing field as China brings a unique combination of technical knowhow and abundant financial resources.
With CloudMinds, we might be entering a cloud platform-based era for robotics, as the value moves away from hardware into software and, more importantly, into the intelligence layer as that gets embodied in the cloud. It will be interesting to see how the cloud versus device/edge-based intelligence plays out for the different robotics and machine sectors like automotive, drones, consumer, and enterprise robots, and IoT in general.