Hong Kong-based autonomous driving startup AutoX is in active discussions with local authorities in Southeast Asia to foray into the emerging market even as the self-driving industry back home in China still remains largely untapped and at an early development stage.
The three-year-old company, which is finalizing its Southeast Asia plan, expects to deploy 10 autonomous vehicles for early-stage experiments in one market to trial robotaxi and robo-trucking services, said AutoX founder and CEO Xiao Jianxiong, commonly known as Professor X, in a recent interview with DealStreetAsia.
The company aims to first build a successful model in one market with the initial batch of vehicles before it increases the car fleet to up to 100 and expands into more jurisdictions in Southeast Asia, said Xiao.
Fully autonomous vehicles in China
AutoX’s ambition in Southeast Asia comes as improved safety, growth of connected infrastructure, and transition from car ownership to mobility-as-a-service (Maas) are fueling the growth of the global autonomous vehicle market. The market is estimated to increase at a compound annual growth rate (CAGR) of 39.47 per cent, reaching $556.67 billion by 2026 from the estimated $54.23 billion in 2019, according to Allied Market Research.
The global autonomous vehicle development is now led by the United States, which is home to 163 autonomous driving companies, shows a ResearchAndMarkets.com report. While America already saw Alphabet’s self-driving division Waymo rolling out “completely driverless cars” – which literally means no safety driver behind the wheel – in the suburbs of Phoenix, Arizona this October, Chinese players are still conducting road tests and collecting data to improve their technologies.
“By the end of 2022, at least one city in China should be able to achieve what Waymo is doing in Arizona, i.e. charging passengers for completely self-driving services without trained safety drivers upfront,” said the 37-year-old founder, who was previously an assistant professor at Princeton University leading the Princeton Computer Vision and Robotics Lab.
“All top cities in China, including Shenzhen, Shanghai, Guangzhou and Beijing, have the potential to take the lead,” he said.
AutoX is not alone developing in the two crucial years ahead of the start of China’s adoption of fully autonomous vehicles. Other prominent self-driving players in the country include Pony.ai, which reached the unicorn valuation of around $1.7 billion after a $50 million funding round this April, and TuSimple, a self-driving truck unicorn with operations in both China and the United States, as well as Baidu, whose fleet travelled over 90 per cent of the total mileage in Beijing’s self-driving road tests in 2018 (as per a local government report).
Combined with China’s strong government support and booming sharing economy, these companies contribute to a domestic self-driving market that is projected to exceed 100 billion yuan ($14.22 billion) in 2019, according to the Chinese industry research platform Qianzhan.com.
“I’m quite sure that China will surpass the United States and lead the self-driving industry in the future. It is just a matter of time,” said Xiao. AutoX, which now has 100 vehicles in China, plans to expand the fleet, conduct more test rides, and venture into more cities in 2020.
Despite being optimistic about the autonomous vehicle development in the home country, Xiao considers now as the right time to take “first-hand advantages” in the emerging market, such as Southeast Asia, where the number of self-driving players is relatively small.
Amongst all jurisdictions in Southeast Asia, Singapore comes first in terms of embracing autonomous vehicles. The island-city state was ranked second worldwide as the best-prepared country for autonomous vehicles development – only after the Netherlands – based on factors including a measure of readiness for self-driving adoption, policy and legislation, technology and innovation, infrastructure and consumer acceptance, according to KPMG’s 2019 Autonomous Vehicles Readiness Index. The country targets to introduce autonomous buses in selected areas as early as 2022.
AutoX could potentially start its Southeast Asian journey in Singapore, or Jakarta, the capital city of Indonesia where the startup received an investment from a local venture capital firm, although not everyone is convinced the road conditions of Jakarta are quite ready for self-driving vehicles yet.
As part of its global business layout, the Chinese self-driving startup already formed a partnership with NEVS, a Swedish electric vehicle manufacturer, in July 2019 and is aiming to introduce a pilot robotaxi service in Europe by the end of 2020.
Besides its presences in China, Sweden and the United States as well as the upcoming foray into Southeast Asia, AutoX also looks at opportunities in the Middle East, Japan and South Korea.
AutoX closed a $100 million Series A round led by Chinese state-owned automaker Dongfeng Motor this September, with participation from investors like Alibaba Entrepreneurs Fund and Silicon Valley-based incubator Plug and Play’s China fund. Hong Kong-based venture capital firm Capital Today, Taiwanese chip giant MediaTek and another Chinese state-owned automaker SAIC Motor were also investors in its previous rounds.
Prior to AutoX, Xiao Jianxiong served as an associate assistant professor at Princeton University and founded the Princeton Computer Vision and Robotics Lab in 2013. He has over fifteen years of research and engineering experience in autonomous driving, computer vision and robotics.
Xiao spoke to DealStreetAsia about AutoX’s upcoming foray into Southeast Asia, self-driving technologies, China’s strong government support, and industry outlook in China. Below are the edited excerpts:
Does AutoX have any recent plans of venturing into new markets?
We are in active discussion about entering into Southeast Asia, but there is no finalized plan yet.
Firstly, Southeast Asia represents a huge market potential and is geographically close to China. AutoX, started in China and now leading the domestic market, will naturally consider overseas opportunities as we are doing very well in the home country.
As a starter, AutoX will experiment and try to build a successful model in one market, before the company expands into other countries in the fragmented Southeast Asian market. We have an idea about where to start our SEA journey, but I cannot share at this stage, because we are still in discussion with the local government about more details including self-driving commissions and discounts.
But the Chinese self-driving market is yet to be developed…
AutoX will continue to develop in the Chinese market. It is not conflicting but parallel with our overseas plan because we deploy local teams to operate in each market.
We now prioritize the Chinese market while having presences in the United States and Sweden. Besides Southeast Asia, AutoX also looks at opportunities in the Middle East, Japan, South Korea and Europe.
Why is it important for AutoX to foray into the overseas market when the domestic market in China is still developing rapidly at the early stage?
The self-driving business is very different from other industries such as app development and the sales of products like iPhone. Self-driving companies need time to work with local governments to obtain licenses, decide parking locations, build high-definition maps, and cooperate with local businesses like taxi operators and logistics companies.
We also need to drive for enough mileages and collect enough data to prove to the local government that the technology is mature and safe. Moreover, self-driving companies will take time to educate local consumers and make them feel comfortable with using autonomous vehicles.
Due to the long duration required to get the market ready, the entry barrier of the self-driving industry becomes so high that early starters can always have first-mover advantages even confronting with strong competitors like Waymo. Late-comers, despite having abundant capital and strong capabilities, will have to start everything from the very beginning.
There are not many self-driving companies in the market. NuTonomy, an MIT technology spin-off that makes software to build self-driving cars and autonomous mobile robots, might be one of a few players to have launched robotaxi services in 2016 in Singapore which allows the public to ride self-driving cars with trained safety drivers behind the wheel.
(Editor’s Note: NuTonomy was acquired by Delphi in 2017 for $450 million. It then became part of Aptiv after its spin-off from Delphi was complete. Aptiv is also in discussions with potential partners for mapping and commercial deployment of Aptiv’s vehicles in China, according to a TechCrunch report.)
Business aside, is it easy to bring a self-driving product that has proven to be successful in China to a new market and quickly scale?
Self-driving technology is essentially very similar regardless of the market. If a self-driving car can drive very well in China – where it has collected much more complicated data and has adopted to various road conditions, the car should also be able to drive very well in most other countries like Japan and South Korea.
What will be AutoX’s offerings in Southeast Asia?
There will be two major businesses, which are the same as our offerings in China and the United States: robotaxi and robotrucks, which are basically trucks equipped with the same pack of self-driving systems used by their robotaxi.
In your observation, do you think local governments in Southeast Asia are supportive of the development of the self-driving industry, and in what regards?
Not only Southeast Asia but also governments around the world strongly support self-driving development. Firstly, safety is one of the critical concerns of local governments. Human drivers could be emotional, distracted or tired, while self-driving cars and compatible computers are fully dedicated to nothing but driving. Many countries in Southeast Asia, such as Thailand, have some of the highest fatality rates of traffic accidents worldwide.
Secondly, the government wants to improve public transportation and make it more convenient for the public. Buses are not a much-welcomed mode of transportation in the tropical region, while taxis, to meet the massive demand, will have to recruit a large number of qualified taxi drivers yet still bear potential safety loopholes like what happened to Didi Chuxing.
Thirdly, the aging population is a major problem. There is a population of 1.4 billion in China right now, but it will shrink to 0.4 billion in 100 years – by 2120 – based on the current birthrate. The shortage of young labour will become more obvious in the next one or two decades, especially in second and lower-tier cities.
Actually, we don’t see many challenges for our venture into Southeast Asia. Some countries, Singapore in particular, are even more friendly than other jurisdictions because of their well-established regulations.
How many vehicles does AutoX plan to deploy at the initial stage of the development in Southeast Asia?
We usually deploy a batch of 10 vehicles for early-stage experiments and then increase to up to 100.
What are the major trends you’re seeing in the self-driving market in China?
A small number of top-tier players in China, including AutoX, Pony.ai, and Baidu, are doing quite well, while second-tier companies are starting to disappear due to financial pressure and other obstacles. I think there will be only a few top players left, even before self-driving vehicles can be put into the market without the presence of safety drivers. The competition will come to an end before the industry takes off.
This is different from the United States where there are almost 10 players developing simultaneously in the market.
AutoX uses cameras to help self-driving cars “see” the surrounding environment instead of lidars, which are currently one of the most popular hardware in the industry. Can you illustrate the major benefits of using cameras?
Cameras are much more accurate and safer than lidars.
Firstly, the two-megapixel resolution of cameras is so high that the details of every single object can be recognized. Cameras generate high-frequency signals, which contain more information than low-frequency signals that are created by lidars. High-frequency signals can help recognize the categories of objects, like humans, dogs, cats, cars, and houses, more accurately than lidars. Cameras can also provide complementary information when lidars are reflected away by windows and glasses.
In comparison, many lidars cannot tell objects very well when they cluster together. For example, if a couple walks together holding hands, lidars could mistake them as one person or simply a moving object. If the couple suddenly splits out, lidars might not be able to recognize and process the information well.
Meanwhile, cameras are also much cheaper than lidars. The price of one camera is usually around $200, and no more than $500 even for a high-end one. A mass-produced radar costs $200. In comparison, the minimum cost of a mechanical lidar is around $20,000, while the price of a solid-state lidar varies from brand to brand. The lidars we’re using, developed by Chinese drone maker DJI, cost $500 each. But lidars made by many other companies are still very expensive.
In general, AutoX takes about $20,000 to upgrade a normal vehicle into a self-driving car using radars, cameras and solid-state lidars.
AutoX currently cooperates with top lidar makers in China, including DJI, whose founder and CEO Frank Wang is my close friend and classmate at HKUST; Shanghai-based Hesai Technology, and RoboSense, which develops micro-electro-mechanical systems (MEMS)-based smart lidar sensors for autonomous driving. The founder of RoboSense is my classmate at high school, while the founder Hesai is the high school classmate of our co-founder Dr. Jewel Li.
We will release our first self-driving vehicle powered by a solid-state lidar at CES in January 2020 in Las Vegas.
Different from AutoX which specializes in level 4 autonomous driving technology, some automakers started by empowering vehicles with level 2 and level 3 technologies and claimed to gradually upgrade them into level 4 in the future. What do you think of this approach?
This is a pure business-driven decision. L4 and L2&3 are two completely different types of technologies, similar to the difference between smartphones and old-version personal handy-phone systems. Automakers simply want to use L2&L3 as a cool feature to help them sell more vehicles.
What is the major business plan of AutoX in the next 12 months?
AutoX now has 100 vehicles in China and is expected to expand the fleet in 2020. We plan to keep polishing the self-driving technology in the next 12 months, increasing the mileage, conducting more test rides, and collecting more data. The company will also expand into more cities with more vehicles.
There are many discussions about the commercialization of self-driving vehicles. When do you think self-driving vehicles will become universally available to the public?
The universal adoption of self-driving vehicles has already started in the United States. Waymo, the self-driving division of Alphabet, already launched commercial ride-hailing service in December 2018 with safety drivers in the vehicles. The company also rolled out “completely driverless Waymo cars” this October in the suburbs of Phoenix, Arizona.
By the end of 2022, at least one city in China should be able to achieve what Waymo is doing in Arizona, i.e. charging passengers for completely self-driving services without trained safety drivers upfront. But after the first city, it is hard to predict how fast self-driving vehicles will be adopted across the whole nation. That heavily depends on regulations.
All top cities in China, including Shenzhen, Shanghai, Guangzhou and the suburbs of Beijing, have the potential to take the lead.
Are there any major breakthroughs we should see in the next two years, in order to realize the commercialization of self-driving services in 2022?
I don’t think any breakthroughs will be needed in the next two years. It is more about expanding into more cities and conducting more test rides to persuade the local government that self-driving cars work.
We understand that the United States is now taking the lead in developing autonomous driving technologies. What is China’s position in the industry?
China ranks second in self-driving development. China and the United States are far more advanced in autonomous driving than any other country in the world. I’m not even sure which country comes third.
Most self-driving companies are in the United States due to historical reasons, including the [more advanced] technological and academic background. Much fewer players were born in China, and even fewer solid players come from other regions including Southeast Asia.
AutoX, Pony.ai and Baidu are the three largest self-driving companies in mainland China.
What are the advantages of developing autonomous driving in China?
Firstly, the road condition is more complicated in China, which is good for conducting stress tests. The more complicated the road is, the more valuable the data will be. You might drive for miles in the United States yet nothing interesting happens to help stress-test the vehicle.
Another benefit is about cost-efficiency. It takes a lot more to do road tests in the United States than China because the labour costs are much higher in the country – about 10 times more – so American players have to burn billions of U.S. dollars.
A lot of technology companies are enjoying many preferential government policies including subsidies and cheap office rents. Is it the same for AutoX?
Yes, it is.
All technology companies enjoy preferential government policies, such as tax reduction, as a default so we don’t need to go through any application procedures. The local government in Shenzhen does not just give startups billions of dollars for nothing. They provide favourable policies to make it cheaper for technology companies to develop.
For example, the rent of our current office [in Shenzhen Bay Science and Technology Ecological Park] is 88 yuan ($12.5) per square metre every month. Even so, this is not the cheapest in the area. We’re moving into a new office building nearby, where the floor we plan to rent is 45 yuan ($6.39) per square metre per month. [Chinese artificial intelligence (AI) giant] SenseTime is also in the same building.
Subsidies from the government are actually not a big deal compared to the amount of money raised by top players. What is more important is the support of the local government. Honestly speaking, the Chinese way of doing things – in which the relationship is critical – is very different from the western style. Self-driving companies, in particular, need to build mutual trust with the government in order to obtain licenses and conduct test rides in the city.
What are the major differences you have observed in terms of government participation in the development of autonomous vehicles in China and the United States?
In the United States, the government will carefully set up rules and introduce regulations, but nothing afterward because they don’t take a stance to support or against the industry.
In China, local governments welcome self-driving companies, and they want to actively promote the industry. If a taxi company is not interested in self-driving vehicles, the government will even call them and persuade them to reconsider it. They want to adopt whatever advanced technologies that are available.
Based on China’s strong government support, a large and dense population, as well as the prosperity of its sharing economy, I’m quite sure that China will surpass the United States and lead the self-driving industry in the future. It is just a matter of time.
Which verticals and use cases do you think will generate the bulk of revenues for self-driving companies in the future?
I think robotaxi and robo-trucking will become the two major use cases in China.
Eighteen per cent of Chinese GDP is generated by the logistics sector, and more than half of the 18 per cent is created by human labour. That means self-driving vehicles have the potential to contribute to 9 per cent of the country’s GDP in the future. This is the robo-trucking industry alone.
Are you seeing the domestic self-driving industry being influenced by China’s economic slowdown, and the Sino-US trade war that has splattered on the technology sector since many tech companies are blacklisted by the United States?
Non-top players will find it much harder to survive because the domestic capital market is colder than before.
In terms of the trade war, I don’t think it is really influencing the Chinese self-driving industry because most components we’re using are made in China, including lidars, sensors, cameras, and even AI chips. We work with many local chip companies, such as Horizon Robotics.