Trade tensions and a global economic slowdown, particularly in automotive manufacturing, have affected demand in the Chinese robotics market. However, interest in supply chain automation and political support of domestic innovation could encourage growth in 2020.
This is Part 2 of The Robot Report‘s Q&A with Georg Stieler, managing director for Asia at international consulting firm STM Stieler. In Part 1, he discussed the state of the robotics market in China, looking at causes for the current slowdown and what types of robots are in demand. Here, Stieler continues his analysis with a look ahead.
Room for optimism?
We’ve seen some optimistic descriptions of the robotics market in China — are they wrong, or are they looking only at certain areas?
Stieler: The International Federation of Robotics (IFR) receives its numbers from the China Robot Industry Alliance (CRIA), which in April still said it expected China to be the main driver for a 14% annual average growth of the global robotics market until 2021. The organizations lowered their expectations when they presented their numbers for 2018 in July.
As I previously mentioned, we do not expect this to be a mere cyclical downturn, but more the beginning of a phase of slower growth in China.
Speaking of the IFR, nobody at the large robot manufacturers we are talking to believes in the 12% global growth that was forecast for next year.
Developers and AI in the Chinese robotics market
What should robotics developers consider when designing or selling systems into the Chinese robotics market?
Stieler: Reducing cost is certainly a major issue. Another major issue is delivery speed.
With the state as a main economic actor, demand in China can be very volatile. When the economy picks up, buyers expect vendors to be able to supply quickly. Balancing this act is a major challenge for operations in China.
How important are technologies such as artificial intelligence and machine learning, the Internet of Things, and 5G networks to developers of robots for China?
Stieler: There are some fundamental misconceptions about China‘s strengths in AI. Yes, the country is leading in facial recognition technology and is strong in speech recognition, but it is lagging in autonomous cars, for instance.
Waymo, the technology leader in this field, can drive over 10 times the distance without disengagement than Pony.ai and over 50 times than Baidu.
However, all this is been built on foreign technology. The two dominant deep learning frameworks, also in China, are Google’s TensorFlow and Facebook’s PyTorch. The Chinese alternative, PaddlePaddle by Baidu, has a much lower presence.
What about computing hardware?
Stieler: The cloud computing centers for the training of neural networks are running primarily on NVIDIA chips. Even though there are now alternatives in form of sophisticated chip designs from Chinese companies such as Huawei’s HiSilicon, these firms are still dependent from U.S. chip design software from Cadence Design Systems or Snyopsis.
Also, chip foundries in mainland China are a couple of years behind when it comes to manufacturing technologies. The most sophisticated chips made by TSMC [Taiwan Semiconductor Manufacturing Co.], Samsung, and Intel are using 7 nm technology. In comparison, SMIC [Semiconductor Manufacturing International Corp.], the most sophisticated foundry in mainland China, is able to master 14 nm. For the production of these chips, you need photolithographic systems, where Dutch ASML is the leading supplier.
In industrial applications, only a few high-end enterprises have truly so-called digital factories. Foxconn, BOE, Tianma, Comau, or SAIC have production lines where the management and processing workflow has been opened up and linked together. These generate multi-dimensional data that can be used for statistical pattern analysis.
The vast majority of robotics companies are still making normal industrial robots. Many of them are not even using offline programming tools at a noteworthy extend, yet.
Robotics market segments worth watching
Where are investors supporting robotics development?
Stieler: Worth mentioning are Dorabot and Mech-Mind, which are both combining computer vision, motion planning, and deep learning for robots, primarily in logistics. In cloud robotics, there is currently nothing comparable to AWS RoboMaker in China.
The next closest might be Geek+’s infrastructure. It claims to have 7,000 robots installed at clients worldwide. CloudMinds, which has filed for an IPO in the U.S. to raise $500 million, still has to deliver its first robots to customers and show that it can live up to its promises.
Ubtech has been outfitted by investors led by Tencent with $800 million to develop humanoid robots. Yet, while Ubtech’s Walker is completing its first steps, Boston Dynamics’ Atlas is already doing somersaults.
There is a lot of talk about 5G as Huawei is trying to position itself as an enabler for the factory of the future with cloud, IoT, AI, etc. I have the impression that it wants to learn as much about manufacturing from its clients as its clients still have to learn about cloud and these other technologies.
Stieler: Sales markets with positive dynamics included the food and beverage industry, which is less affected by the trade war and has still comparatively low-hanging fruit in terms of applications. Driven by subsidies, photovoltaics had good year, and manufacturers have invested in replacing human labor with robots. We also saw growth in the medical industry, albeit on a low level.
Influenced by China’s tech hype of the past three years, some people might expect AGVs or drones in last-mile delivery, but in reality, it’s a human driver with a smartphone and an electric scooter who makes a bit more than $20 a day.
There were reports of big data, drones, and first field robots being used in Chinese agriculture, but we don’t see this as a mass phenomenon yet. The average farm size in China is 1.4 hectares [3.4 acres], which is less than 1% of the average farm size in the U.S. Their purchasing power is correspondingly low.
The Robot Report is launching the Healthcare Robotics Engineering Forum, which will be on Dec. 9-10 in Santa Clara, Calif. The conference and expo will focus on improving the design, development, and manufacture of next-generation healthcare robots. Learn more about the Healthcare Robotics Engineering Forum, and registration is now open.
Geopolitics and robotics
Does uncertainty in Europe around Brexit and in the U.S. around upcoming elections create an opportunity for Chinese or foreign companies?
Stieler: In the U.S. in particular, Chinese companies are having a hard time winning trust these days.
In Europe, the situation is still a bit different. There is still the idea that Chinese investors had endless amounts of money because they have paid high premiums in the past. Estun’s recent acquisition of Cloos showed that this is still somewhat true if interests are aligned with those of the government.
In general, we are seeing that financing at home has become more difficult, and Chinese the government is approving less cross-border transactions this year. China is perceived by many as this stable monolith whose rise was a historical necessity. If you look closer, it has a lot of headaches of its own.
Looking ahead into next year, what is most likely to change? Do you think the situation improve or get worse before it gets better?
Stieler: Orders received among all major robot manufacturers continued to fall, with double-digit declines in Q3 2019, darkening the outlook for Q4 2019.
In the beginning of this month, the open quotations of many systems integrators are on an all-time high. Due to the ongoing economic uncertainty, the majority of their customers keep postponing their investment plans further. Only few clients are pushing ahead now to do something in 2020.
How might the political and macroeconomic environment affect the robotics market?
Stieler: Over the past year, the Chinese leader Xi Jinping has repeatedly warned Communist Party officials to steel themselves for “struggle” and hazards such as possible economic turbulence, rising debt levels linked to local governments, technological competition, and sparks of social discontent spread across the Internet. The wording from the latest leadership meeting in early November suggests that Xi sees no easing in those risks.
Further into 2020, we expect a stronger emphasis on manufacturing for domestic demand. There is still potential for the automation in the production of solar panels. The demand for semiconductors for servers and cloud computing is also expected to grow further.
With the change to 5G, new production lines for smartphones will be needed. Besides electric vehicles, we might also see more hybrid and hydrogen vehicles with a demand for corresponding production lines.