Volvo's autonomous driving car model on display at an auto show in Los Angeles, the United States, in November 2015. [Photo/China Daily]

Yan Honghui, an analyst with Beijing-based Internet consultancy Analysys International, said: "Rather than making a fortune by simply selling smart cars, companies are betting big on tapping into people's various in-car demands.

"Take Baidu for example. If it knows where you go, what times you typically travel, and the music and entertainment that you use while driving, it can gain a much better understanding of your behavior," she said.

"The data it gathers from the pattern of your behaviors can help it better identify the service you need, and it can profit easily by selling a movie ticket and booking a restaurant."

The world is already on that path. Many vehicles boast features that assist drivers and eliminate errors, including blind-spot alerts, lane-departure warning systems and automatic emergency braking. But fully autonomous vehicles still need to overcome a wide range of barriers, from technology to regulations, before they could become commercially viable.

Wang Jin of Baidu said that to make a premium car such as Mercedes Benz perform all its functions flawlessly, engineers need to write 65 million lines of code.

But making cars smart will need more than 300 million lines of code. "And a fully autonomous car will be much more complicated than that," he said.

So much so, some of the engineers working with Baidu's autonomous driving vehicle unit write code when riding the prototype car.