The Robotics Startup Scene In China
摘要： Robots are going to be a major driving force that will transform the world of mobile Internet. Many people might have noticed that robots nowadays can be easily found in hotels and hospitals as helpers in different fields. And the Bank of America recently estimated in its 300-page-long report that in 2025 robots and other forms of artificial intelligence would shape the world we have right now into something we would barely recognize.
Robots are going to be a major driving force that will transform the world of mobile Internet. Many people might have noticed that robots nowadays can be easily found in hotels and hospitals as helpers in different fields. And the Bank of America recently estimated in its 300-page-long report that in 2025 robots and other forms of artificial intelligence would shape the world we have right now into something we would barely recognize.
The development of the robotics industry is taking up further into the post-mobile-Internet era. At least in China, the robotics industry and the field of artificial intelligence are constantly renewing people’s perception on technologies. Additionally, these two areas are also driving the startup scene in the country as well.
Just like what you have seen in many sci-fi movies, any form of robotics technologies is based on artificial intelligence. Kevin Kelly once predicted that the AI technology, as the next significant application of cloud computing technologies, would be a disruptive innovation. In the future, a majority of innovative products would include AI technologies in them. For instance, products like automobiles, toys, and appliances etc. would possess AI features and become smarter. “The next startups that are about to emerge would all do something like this: take X；add AI. ”
Jack Ma believes that in 30 years the best CEO would be a robot thanks to the rapidly developing information world. In the future, the adoption of robots will become as ordinary as cars and planes. They might even become part of our family.
Without a doubt, many countries around the world are getting ready for a robot competition in which they want to have an upper hand. In some developed countries there are even plans of reviving the manufacturing industry based on robots, such countries include Germany, Japan, and the U.S.
For a long time people in China only know about robots from reports from foreign media outlets. Discussions about robotics technologies and the industry are also fragmented and rare. In June this year, Alibaba Group announced jointly with Foxconn in Japan to have invested 14.5 billion Yen in Softbank’s robotics company SBRH. After the launch of SBRH’s innovative robot model Pepper, which is priced at 100,000 RMB, the first two group of merchandise, around 2000 robots, were bought out by Japanese consumers. This even then sparkled a heated discussion about the field of robots and has won much attention in China.
However, even though China’s robotics technologies are obviously lagging behind in comparison with developed countries, the demand for industrial robots is high and growing. But still, most robotics companies in China focus more about service robots, companion robots, and smart appliances etc., and therefore there’s still a huge black space waiting to be filled in. But many might ask: Just how big is the robot market? According to reports from International Federation of Robotics, the global service robot market, which mainly aims at hotels, old people’s homes, and the medical sector, will attain a market share of more than 50 billion dollars by 2017.
In October this year, the Plateno Hotel in Asian Games Village, Beijing, purchased a group of service robots for the hotel, basically realizing the goal of automatic delivery within the hotel. The robots Plateno Hotel bought were from a startup company called Yunji Technologies.
Service robots are entering the functional stage in different fields. In August last year, American company Savioke rolled out its robot series SaviOne and had it tested and later officially adopted for actual uses in the Aloft Hotel. It’s considered as the first case in which service robots for hotels were adopted for actual business uses in foreign countries. Coincidently, in July this year a hotel named Henn-na in Japan featuring robot services was opened. The hotel’s service crew consists of robots that can speak multiple languages and whose responsibilities include serving at the front desk, delivering food, and cleaning the hotel.
However, industry insiders had revealed to TMTpost that although service robots had been put in to business uses, they were still just machines instead of robots since the robots we have today were programmed, without actual sensitivity and thought. If we were compare them with humans, they would be kids around 2 or 3 years old, far from the standard of adults.
“The lack of mature technologies, advanced algorithm, and the support of big data makes service robots the best entry point into this sector,” an insider explained. He didn't deny the huge demand for service robots in China for the fact that the Chinese society is aging and losing labor force, and the cost of human labor is rising. It’s natural for Chinese companies to seek alternative to replace human workers with robots.
Most startup companies start with developing software and applications and later begin to develop robots of different uses. For example, Ju Shang Information Technology who has just launched its Small Zhi Robot model started with voice recognition technology. This company became famous for developing 360’s voice interaction function. It wasn’t until January 2015 that Small Zhi launched its Small Zhi speaker that featured remote voice interaction function, targeting the smart appliance market.
Small Zhi Robot’s CEO He Yon has always had the idea of making humans robots, and therefore he started with voice assistive technologies. However, due to the lack of fund and mature computer and biotechnologies, he had no choice but to start making machines intelligent. That’s why he used voice assistive technologies as an entry point.
But here’s the problem. Voice assistive product could generally be popular for a short period and they lose users fast. And it happened that in 2012 the smart alliance sector started to rise. So what did they do to turn the tide? At that time, He Yong chose to lead his team and go find cooperation opportunities with appliance makers in hopes of fusing his software with hardware. It was definitely a good move to enter the smart appliance sector.
“The company was positioned to the direction of smart appliances. We were not limited to smoke detectors, alarm systems and such, but real smart appliances. The fusion of humans and machines, machines and machines, is the tendency,” He Yong told TMTpost. After Small Zhi launched its speaker, Internet giants also started to act, including Alibaba and Philips who had teamed up together to launch Xiao Fei smart speaker, and JD, who later rolled out Ding Dong speaker. And the dominating advantage that Internet giants have makes it extremely difficult for newly-emerged brands like Small Zhi to make their names.
At present, Small Zhi Robots not only has launched its robot hardware product, but also rolled out its robot operating system named S OS and SmartKit open source platform that targets smart appliances.
According to He Yong, S OS is a operating system made available to all robot makers. With S OS, enterprises can easily make their own customized robots. It’s expected that the company will roll out S OS’s developer tools which developers can make apps for robots.
SmartKi is a platform for smart appliances. Through easy connecting protocols, smart appliance makers can build their own smart devices.
Small Zhi’s goal is establish an ecosystem of smart appliances that integrated platforms, software, and hardware. This is He Yong’s third entrepreneurial attempt and according to him it has been quite a journey. But still, thanks to the hardship, he and his team were able to grow faster and better. In China, making robots is not an easy thing to do. For instance, Small Zhi Robots has encountered the following obstacles:
- Educational robot products rely heavily on offline promotions since consumers prefer to observe the real products themselves;
- The cost of online channels is not as low as people expected;
- In the business of intelligent robots entrepreneurs need to create their own fortress by offering unique features, remote voice interacting function and speech recognition function;
- It’s hard for small and mid-sized companies to order components for their products since factories in general only accept orders above the scale of million or even tens of million;
- The continuous low pricing strategy will lead the company into a pit where the fund will run out and eventually making the company die down in the hole.
In He Yong’s opinion, the current Internet giants in the world such as Apple and Microsoft will not be the leading force in the robotics industry. “The future lies in the hands of startup companies,” He said.
Turing Robot and Small Zhi Robots actually share a lot in common in terms of their developmental history. Both companies started with developing voice recognition technologies and had cooperated with IFLYTEK before. The biggest difference between these two is that Turing Robot hasn’t launched any robot product yet. Instead, Turing Robot teams up with its partners to develop hardware, and focus on developing and refining its open-source platform.
That being said, Turing’s robots are the actual robots we have in mind, but rather an intelligent platform. Before Turing Robot’ establishment, this team, which was famous for their technical talents, once launched a voice assistive product named Wormhole.
Turing OS was recently launched earlier this month. According to Turing Robot’s founder Yu Zhichen, compared to traditional operating systems that provide users with mouse, keyboards, and other peripheral devices to achieve human-machine interaction, Turing OS enables robots the capability to be highly humanoid. Simply put, mouse and keyboards will be replaced by fingers, voice control, gestures, and even the movement of our eyeballs can achieve the function of human-machine interaction.
When asked about the reason for developing robot operating system, Yu Zhichen said:
“More and more tech giants start to agree that the developmental path of the robotics industry will be similar to that of the PC industry. And what we can really learn from the PC industry is that operating system that makes operating a computer easier and offers better user experience is what really took the PC sector forward.”
There’s also another reason for developing an operating system and it’s deeply linked to the current state that the robotics industry is in. Yu Zhichen told TMTpost that the robotics industry in China was still lagging behind in terms of innovation, and that the industry had been overly eager. He believed that in many cases a company not only needed to do both hardware and software, but also integration. And because of that, only a few companies in China actually have the R&D capability to come up with innovation. “If we can make a great operating system, it can lower the R&D cost and give us more time to think about how to define a good product. This is what robots for domestic use should be about,” Yu said.
Yu Zhichen also mentioned that many startup companies were following Apple’s concept in developing smart hardware, meaning they tend to integrate software and hardware. From where he is standing, in the field of robots for domestic uses there should companies like Apple, Microsoft, and Google. Yu Zhichen’s goal is simple. He wants to use Turing OS to build an ecosystem like Pepper has done.
However, Turing Robot didn’t make Turing OS available to everyone, but only a few key cooperating partners. What’s more, Turing OS lacks functions and currently focuses on child services. On the press conference of Turing OS, Turing Robot together with GDALPHA announced a companion robot model for children that ran Turing OS.
Yu Zhichen believed robots were made to solve people’s problems instead of replacing people. But robots so far can only be seen as tools.
Furthermore, the Chinese language makes it harder for robots comprehend Chinese people since Chinese relies in syntax to express meanings and sometimes it’s very hard to tell the meanings if the context is not given. To solve this problem, Yu needs to create a better algorithm and continue to improve it in practice. Yu Zhichen’s view matches Small Zhi Robots’ CEO He Yong’s. Hence, voice recognition and interaction technologies are the common field that robot makers should pay attention to.
As a matter of fact, whether it’s Yuji Technology, Small Zhi Robots, or Turing Robot, their products are powerful because of having a cloud-based brain that’s able to learn. Voice recognition and interaction technologies are just a tip of the iceberg of robotics technologies and artificial intelligence technologies.
This explains why Internet giants around the world are learning everything they can to know about deep learning technology.
Google’s method is use open-source machine learning system TensorFlow to transform the field of AI technology. It works this way: using multiple computers as nerve cells to form a network, which doesn’t need researchers’ help to tell the differences between things if relevant data is given. Simply put, the data is the learning material. This technology allows machine to learn to understand data and use data.
In fact, at the end of 2014, Google had successfully given machines the logic through RNN. In the end the machine was able to make simple description about a picture with one sentence. In this case, the machine possesses the ability to distinguish different items.
So, just how important is it for machines to have deep learning ability? Google’s researcher Greg Corrado explained that machine learning technologies were just like the rocket booster and the sea of data is the fuel.
Baidu’s first scientist Andrew Ng also sated on the BIG TALK Conference in January this year that combining deep learning algorithm with big data would make AI better, and that the future AI would be able to create a circle.
Chinese startup companies have been working on this field as well. For example, DeepGlint is a Beijing-based company that commercializes 3D computer vision and deep learning technologies, and it intends to explore the value AI technologies have in the field of public security, finance, and in the legal world.
DeepGlint’s co-founder He Bofei’s wish is turn every little detail that’s happening in our lives into data that computers can understand. And with that, He Bofei would be able to bridge the gap between the physical world and the virtual world.
He Bofei also stated that the ultimate goal for humans of building robots or let’s say, artificial intelligence, is to liberate humans, which is part of human nature. In the future, robots might not be in human form or animal form. Instead, anything can has to ability to sense and act is regarded as a robot.
He Bofei’s understanding is actually a reflection of how the industry thinks about robots:
- In the future, should we treat robots as actual intelligent being or just purely tools?
- After robots acquire the ability to learn by themselves, would they turn rogue against humans?
So far there’s no answer to these two questions. However, it’s for sure that the robotics industry would disrupt the traditional business model just like KK predicted. In accordance with McKinsey’s estimates, by 2025 robot production for the manufacturing sector and service sector would reach a volume of 1.7 trillion dollars to 4.5 trillion dollars. In this case, even though the robotics industry in China is facing obstacles such as the lack of attention and mature technologies needed for development, high cost, and competition from international heavyweights, nothing is going to stop China from becoming the largest robot market in the world.
[The article is published and edited with authorization from the author @Song Changle, please note source and hyperlink when reproduce.]
Translated by Garrett Lee (Senior Translator at ECHO), working for TMTpost.