China’s internet-search giant Baidu is planning to mass produce driverless electric cars in five years, bringing the fight to Google and Apple .
The company, which currently holds 80% of the internet-search market in China, will not manufacture the cars itself. Instead, it will outsource the actual production to an existing Chinese automaker, in the same way LeEco planning to outsource production of the LeSee. Baidu hasapproached several Chinese automakers but hasn’t made a final decision yet.
Currently, Baidu is testing its driverless technology with a fleet of BMW 3-Series GT. The tests started in December 2015 with test cars driving on public roads in Beijing and Wuhu. BMW will have no involvement in the eventual Baidu driverless production car.
In April Baidu announced the formation of a driverless-technology research center in Silicon Valley, and the company’s chief scientist said they want to start testing driverless cars in the United States “soon”.
Baidu’s first commercially available driverless production vehicle is scheduled for launch in China in 2018. It will be a shuttle bus designed to run on a pre-set loop line. Similar lines already exist in other countries, including Greece, The Netherlands, and China. Production of these shuttles will be outsourced as well.
The driverless car for public roads will launch in 2021. The car will be fully autonomous without any need of any input of the passengers, bar their destination. The operating system is called Baidu AutoBrain (百度汽车大脑).
The system was announced in September 2015. It combines all the key ingredients needed for driverless driving, plus Baidu’s MyCar car-connectivity software and Baidu’s CarLife synchronization software for mobile devices.
One of the most important elements of the system, or any autonomous driving system, is the map. And Baidu has one of the best, fully incorporated in their search software. BaiDu Maps is comparable to Google Maps, but at least in China it is more accurate and up to date, with fully photographed 3D surroundings of every street and lane in every Chinese city. And again like Google Maps, it is helped by the millions of users adding information and photos while using their mobile Baidu apps.
Another vital piece of the puzzle is the self-learning capability of the system’s software. China’s sometimes chaotic road conditions make this especially important. Baidu has said that the most difficult parts to master are the car’s responses to situations that cannot be easily scripted, like a policeman gesturing or cyclists breaking the rules. Self-learning enables the vehicle to learn how to react to these unexpected situations.
Will Baidu be able to pull it off? In 2015 the company had a revenue of $10.6 billion and a profit of $5.3 billion, market cap hovers around the $55 billion. For comparison; Apple is worth $512 billion and Google $466 billion. That leaves Baidu with far less money on the table than its American competitors, and every investment in the car business automatically means less investment in the core search business.
Baidu hopes to get around this deficit by outsourcing actual production of the vehicles, so they won’t need to build factories and a supply chain. This however also means Baidu won’t have full control over the production, which is a weakness especially in a new field like driverless cars. Furthermore, Baidu will still need to set up a distribution network and dealers for their cars, unless they plan to outsource that too.