While the likes of NVIDIA, Alphabet's Waymo, and Intel have been making waves in self-driving vehicles in the West, Baidu (BIDU 5.11%) is charting its own course in China. The internet giant had promised a few years ago that it would make a driverless vehicle by 2018 and start mass production by 2020.
True to its promise, Baidu recently rolled out the 100th self-driving bus off its production line at a facility in the province of Fujian. The company plans to start operating these buses across several Chinese cities, including Beijing, Xiongan, and Shenzhen, as well as international markets such as Japan. But will such a move actually help Baidu make a dent in the autonomous vehicle space, or are the buses just a marketing gimmick? Let's find out.

Image source: Baidu.
Not a big deal right now
Baidu has developed the 14-seater "Apolong" autonomous buses in collaboration with Chinese bus manufacturer King Long. These buses are capable of Level 4 autonomy, which is considered a "high automation" level: They can mostly drive on their own and need human intervention only under some circumstances when the special maps they rely on aren't available.
So Baidu's autonomous buses will be a good fit for driving in selected areas such as airports, industrial complexes, parks, or tourist spots, which are basically geofenced areas. But there's a question mark about their ability to be deployed in mainstream traffic right now, given the restrictions they face.
First, the buses can ferry only 14 passengers, so deploying them in congested cities such as Beijing doesn't make any sense, even though they can drive on their own. Additionally, the low passenger volume will probably make it uneconomical for fleet operators to deploy them, while the fact that they would still need a human driver in certain situations means that there would be no savings on labor costs either.
So the new Baidu buses are unlikely to become a means of mass transportation. But there are a couple of positive takeaways from this development that could help Baidu's autonomous-driving prospects in the long run.
The bigger picture
Baidu has been very clear about the way it's approaching self-driving cars. The company will focus on developing and perfecting the technology, and use it to power vehicles from its numerous partners. The Apolong self-driving buses will help Baidu conduct a real-life trial of its autonomous driving platform, Apollo, enabling the company to collect data about the roads its buses travel and to perfect the algorithm.
Once that's done, it can easily sell Apollo's services to the 116 partners that have signed up to use this platform, including the likes of NVIDIA, Hyundai, Ford, and Honda.
Another opportunity that Baidu could attack with these self-driving minibuses is the shared-mobility market. After ride-sharing cabs, operators are now busy exploring the feasibility of buses ferrying passengers using an app. Ford, for example, recently received a permit to test its Chariot minibus service in London after doing so in five cities in the U.S.
Late last year, German automaker Volkswagen announced that it's planning to start a ride-sharing service using 200 electric minibuses on the streets of Hamburg. Users will be able to book seats on Volkswagen's service through an app, which will match customers looking to get to similar destinations.
Such applications are going to boost demand for self-driving buses, which is why Technavio analysts estimate that the global autonomous-bus market could grow at a CAGR (compound annual growth rate) of 38% over the next four years.
An eye on the future
Baidu's self-driving buses will ensure that it doesn't miss out on the autonomous shared-mobility market. This is in addition to the opportunity that Baidu has to hone its self-driving technology in the real world, by deploying the buses in select areas before it goes on to make a play for the bigger prize in mass public transit.
Although the Apolong buses might not be of much help to Baidu right now from a financial perspective, they will pave its way to making a mark in the autonomous driving space in due course.