According to “Technology Roadmap of Energy-saving and New Energy Vehicles”, a publication by the Society of Automative Engineers of China, the market share of driver-assist/partially-autonomous vehicles will reach 50% by 2020. By 2025, the market share of highly autnomous vehicles will reach approximately 15%, and by 2030, the market share of fully automated vehicles will be nearly 10%.
The development of autonomous vehicles triggered fierce competition in this wave of artificial intelligence craze. It is not only relevant to the traditional automobile industry, but also relevant to the interests of the emerging technology companies and the consumers. It also concerns core issues such as the iterations and upgrades of the automobile industry.
How have these emerging Chinese autonomous vehicle startups responded to such fierce rivalry ?
The Race of Self-driving Startups
On January 29th, the autonomous vehicle startup Pony.ai announced its trial operations in Nansha, Guangzhou. It opened a trial operation of its Level 4 unmanned vehicle to the public for the first time. During this time, people can hop on and off the vehicles to enjoy light shows on a designated route. Technical staff are also onboard to give explanations and presentations.
A day later, Jing Chi Corp (an AI-based smart travel company) held an opening ceremony for their self-driving trial operation on Guangzhou International Bioisland, and announced that from now on, the general public could try out the company’s Level 4 self-driving cars through open reservations. The trial operation will last at least three months, and will be expanded to other areas in the future.
It is safe to say that although these startups are not yet in a fight, they are at least in the middle of a race.
Let’s Talk about Level 4 and Autunomous Driving
At present, the classification of automobile intelligence technology comes from two sets of foreign standards. The first set of standard was developed by National Highway Traffic Safety Authority (NHSTA), a subdivision of the US Department of Transportation, and the second standard was set by SAE International, the Society of Automotive Engineers. On September 20, 2016, NHTSA announced that it would adopt the SAE International’s more detailed standard with six levels from Level 0 to Level 5.
According to the six-level classification, Level 4 is also called “highly automated driving”, but only limited to the vehicle’s operational design domain (ODD). It means that only under scenarios that the HAV (Highly Automated Vehicles) is designed can fully automated driving be achieved.
The operational design domain includes parameters such as road type, geographic interval, speed interval, environment, climate and so on. Level 5 basically refers to machines taking over the vehicles completely. This level requires that the vehicle must be able to deal with all conditions, even responding to extreme driving scenarios.
Chinese tech companies conducts modification and testing of self-driving technologies based on existing and mature vehicle platforms. No one dares to be as aggressive as Elon Musk, who experimented using the electric vehicles produced by Tesla. Although the self-driving system has repeatedly caused accidents or deaths, people have not lost affinity nor interest in Tesla, the pioneer of electric vehicles.
However, it’s a different story for the Chinese tech companies. Once the technology is commercialized, they cannot allow anything to go wrong. Auto manufacturers will not risk their own reputation accumulated over dozens of years. Traditional car manufacturers built cars based on the concept of zero casualty. They cannot afford to integrate any unsafe factors into their cars, and the same concept applies to the self-driving technology that will be used on these traditional vehicles.
How long until Level 4 self-driving can be commercialized and put into mass production? Do these startups have enough money to last until then? What kind of competition will they face tomorrow?
Baidu and their Startups
On October 17th of last year, Baidu signed a strategic cooperation agreement with Xiamen King Long United Automotive Industry Corporation. It plans to realize small-scale production and trial operation of commercial self-driving cars by July 2018. The main scenarios for the trial is in a factory area or in a park. Baidu aims to overturn the industry. It is trying to build not only an entire solution for the unmanned system, but also a common R&D and testing platform.
Baidu established the Intelligent Driving Group (IDG) through the integration of the smart cars (Level 3) and self-driving (Level 4) departments. The leader, Lu Qi, has been emphasizing that Apollo is a complete software, hardware and service system that combines vehicle, hardware, and software platforms as well as cloud data services. At the same time, the platform open sourced the tool relating to environmental awareness, path planning, vehicle control, and on-board operating system, while providing complete development and testing kits.
Jing Chi Corp and Pony.ai went straight to R&D of L4 technology and smart travels after leaving Baidu. The two companies are evenly matched in terms of technology, capital and talent, but the biggest problem still lies in pushing for mass production.
On April 3, 2017, Jing Chi Corp was established in Silicon Valley. It completed its first self-driving test on May 12th. In June, it obtained the testing license in California, and completed the open road test in the same month. On September 8, Jing Chi Corp completed the road commuting test during the rush hours in Silicon Valley. At present, it has raised $30 million angel round, $52 million in Pre-A round, and plans to start series A funding with a goal set to $100 million. The secret to Jing Chi Corp’s development seems to be fundraising and recruiting talent from other companies.
Comparatively, Pony.ai has kept a low profile. The company was established slightly earlier than Jing Chi Corp in December 2016, which should have given it a leg up in technical progress. Pony.ai seems to be more focused on developing the technology, as it has not disclosed much information about its plans in the smart travel market. The announcements of its Level 4 products ready for trial and its $112 million in series A both came out as surprises.
Yu Kai, the former director of Baidu Deep Learning Institute is more of a maverick. After leaving Baidu, he founded Horizon Robotics in July 2015. At present, his company focuses on several areas including self-driving and smart city.
Horizon Robotics is a local provider of embedded artificial intelligence core technology and system-level solutions. The company adheres to the solution of “algorithm+chip+cloud” and is designed to install “machine brains” onto numerous daily devices and products, making them capable of perception, interactions, understanding, and decision-making.
Yu Kai believes that it will take at least another 7 to 10 years before the prosperity of self-driving market, but there will be an explosive growth of assisted driving in the next three to five years. During this period, Horizon Robotics will focus on the semiconductor business.
Black Horses in the Race
Apart from these people from “the Whampoa Military Academy of AI” Baidu, there are also other self-driving developers in China, such as TuSimple, who focuses on Level 4 self-driving trucks mainly used on the highways. TuSimple is expected to realize the commercialization of 1,000 vehicles in the first quarter of 2019.
TuSimple and NVIDIA have been in close contact. TuSimple is also one of the earliest members of NVIDIA’s Inception Program. NVIDIA has taken over 3% of TuSimple’s shares during series B funding, and has raised $55 million in series C in November 2017.
Momenta was established in September 2016 and it received $46 million in series B round in July 2017. It is worth noting that this round of funding was led by NIO Capital with participation from car companies like Daimler Group (the parent company of Mercedes-Benz). Momenta’s main business includes creating solutions for all levels of self-driving systems, which is similar to other self-driving startups. The second part of their business which is unique to Momenta is providing big data products and services like high-precision maps and more.
Self-driving technology is the trend and traditional international car companies have all made commitments to the R&D of this technology. Although their progress might not be as fast as tech companies but their industry is relatively stable. All they need to do is to wait for the world’s most advanced and mature technological achievements to come out to avoid taking detours and risks.
On the path to self-driving, tech companies and traditional car manufacturers are all competing for speed and progress. This competition has reached a feverish pace and let us anticipate the future!