Joey’s Perspective

It has been one year since Baidu’s foray into the automotive industry.

On this day last year (March 2nd), Jidu, the new automotive brand co-created by Baidu and Geely, was officially established. In contrast to other market newcomers that are obsessed with staying relevant amid the immense competition, Jidu appears to be the type of brand that does not attempt to manipulate their volume intentionally.

Nevertheless, precisely because of this, it seems that they are worth everyone’s attention. This is because Jidu is a new automotive brand that doesn’t do much publicity. If they don’t have the money, they must be quite wealthy (money is undoubtedly important to new automotive brands). And Jidu is quite wealthy.

Firstly, Jidu’s financial stream is not a problem because it’s a joint venture between Baidu and Geely. Therefore, they don’t need to amass a stable market volume before car design and production even began to ensure the smooth flow of funds. In fact, Jidu easily achieved its A-round financing of 400 million USD in January of this year, with investments coming only from Baidu and Geely.

Additionally, Jidu’s road map for car manufacturing is more clear-cut than others. After all, Geely, which has already climbed to the top-tier of Chinese car brands, has unquestionably robust design, production, and manufacturing capabilities in the Chinese market. They chose to collaborate with Baidu because the roles and responsibilities between the two companies are very distinct within Jidu.

As early as 2017, Baidu proposed the strategic concept of “All in AI,” which aims to shift the company’s focus fully toward artificial intelligence. It is this goal that enabled the two parties to believe that they can generate fantastic chemistry between them. Geely is responsible for automobile manufacturing, while Baidu heavily focuses on the research and development of AI and other artificial intelligence technologies. If each party does well in its own sector, Jidu’s product and prospective success are pretty much guaranteed.

Compared to other emerging car manufacturers, it may be challenging to perceive Jidu’s existence in this first year without intentionally inquiring about it. It is like a hardworking student in a class quietly making progress and keeping a low profile.

It’s Time to Step Up

If Jidu was in the process of accumulating a team, constructing a framework, and establishing a supply chain in the first half of 2021, this year-end and early 2022 phase is the first wave of Jidu’s brand display achievement.

The speed of Jidu’s development is undoubtedly very fast. In June of last year, just over three months after its establishment, Jidu’s first car model clay model was exposed. In July, the car’s styling was confirmed and entered the engineering development stage. In August, the supply chain system was completed, and Jidu formed strategic cooperation with CATL, Continental, and Hesai Technology. In September, Jidu’s SIMU Car (Software Integrated Mockup Car) road test was exposed, and the intelligent cockpit entered into research and development.

In January of this year, Jidu released the part of the product style of its car robot concept car and officially released a series of related product supporting contents for its tailor-made brand visual system, brand logo (Pixel-J). In short, Jidu’s car-making seems to be from scratch, but such a fast development speed clearly tells everyone that they have extremely profound car-making heritage and strength.

Of course, just looking at these progress, although the speed is fast, what you see is still some “conventional” car-making operations. According to Robin Li’s idea, Jidu is not going to make traditional new energy vehicles, but travel robots. So, at what stage is the development of intelligent driving?

In 2021, there were three events that can be called important milestones in the development of intelligent cars. The first is Tesla’s launch of FSD (just recently, they also officially pushed out the Beta version in Canada); the second is IM’s 40-minute unmanned driving assistance system test on Shanghai’s bustling streets; the third is the jointly launched Alpha S Huawei HI version by Huawei and Jihou, which can also achieve very advanced intelligent driving assistance.

Obviously, judging from the existing information, the technology these brands have, in terms of leadership, is considered to be the head of the intelligent driving assistance brands (stability and reliability may still need verification from time and the market). But in fact, Baidu had also reached cooperation with Jihou much earlier.

In the month before Jidu was established, in June 2021, Baidu Apollo cooperated with Jihou to launch a Robotaxi model. And in the torch relay process of the Beijing Winter Olympics that has just ended, Baidu’s Robotaxi also participated. In the form of an unmanned car, it completed an 800-meter torch relay ceremony.

Although from the perspective of Robotaxi, it seems out of touch with the car’s appearance that ordinary consumers want to buy, with a large laser radar on top. But don’t forget that this is to deal with unmanned driving scenarios, and the peripheral hardware required at this stage.

In the daily intelligent driving auxiliary scenario of ordinary mass-produced cars, a more exquisite design and miniature laser radar device has already emerged. Therefore, we can expect advanced intelligent driving assistance technology without worrying about it being different from the appearance of ordinary electric intelligent family cars.

After all, from the beginning, Baidu’s goal was not to build a simply electric intelligent car, but an intelligent robot with L4 automatic driving capability. After all, since the end of 2020, Li Yanhong has redefined Baidu as an “AI ecosystem company.”

However, challenges still exist for Jidu. Needless to say, just the three companies with good strength in the field of intelligent driving that I mentioned earlier can become visible competitors with strong competitive strength in the direct confrontation with Jidu.

So, what are the challenges Jidu faces?

Challenges Faced by Jidu

What will the future intelligent car look like?

In Jidu’s view, the future intelligent car will be characterized by “free mobility, natural communication, and self-growth”. Among them, having L4 automatic driving capability is one aspect, and the rest is that the vehicle can accurately recognize users’ instructions, understand their needs and emotions. At the same time, it realizes self-learning and iteration based on the first two functions.

Undoubtedly, what I mentioned above is not only Jidu’s understanding of car intelligence, but also the consensus of the entire industry on the development direction of future intelligent cars except for Jidu. Therefore, what everyone is doing is the same.

At present, there are not a few companies fully committed to research and development in this field. For example, the Alibaba Dharma Academy, which also specializes in AI artificial intelligence, almost cooperates with Jidujilan in the same period to cooperate with Ziji. From the perspective of known information, their R&D progress can be considered very good.

At the same time, Huawei’s full-stack intelligent solution for cars is also trying to solve this problem. On the AITO Queshi M5, which I recently test drove, you can find the most advanced design and application scenarios in human-machine interaction almost at the moment. Whether it is the design of the interface, the interaction of the voice, or the recognition of voice semantics, as well as the self-learning and self-evolution scenes driven by big data.

For now, it seems that the intelligent car machine and human-computer interaction have encountered a small bottleneck. How to break through is not just a matter of Jidu, but a problem that almost all car manufacturers who are walking this path will face.

Does Jidu have its own advantages?

First is the problem of the intelligent architecture of smart cars. Actually, many manufacturers are developing this kind of architecture, including SAIC’s SOA architecture, GAC’s Xingling architecture, and so on. The primary issue to be solved by these architectures is the integration of function blocks that were originally unrelated and had different system languages, which is followed by the creation of a brand-new smart cockpit and vehicle system.

As everyone is currently developing this technology, Jidu naturally cannot be left behind, after all, this kind of ecological architecture platform is a must for the intelligentization of vehicles. Fortunately, Geely has already developed the Hohm architecture. The function undertaken by this set of architecture is to achieve the cross-domain integration of various function blocks. Therefore, for Jidu, this task has already been completed.

In addition to the platform architecture, Jidu’s first requirement for intelligent robots is to achieve free movement. Among the components of free movement, one is the driving perception and sensing technology, and the other is the capability of relying on big data to recognize the route. In this regard, compared to new car brands that need to layout in the high-precision map field in advance, Baidu has a natural advantage.

For now, Baidu Maps is the only one that can contend with the “leader” of the navigation map field, Gaode. What’s more important is that compared to its 2C business, Baidu Maps has a much larger market size in the 2B field. That is to say, Baidu pays more attention to the monetization model in the 2B field, which not only involves a choice between different business models but also reflects the applicability and leadership of its technology to some extent.

Undoubtedly, what the intelligent vehicle autonomous driving field needs most is the leadership of maps in the commercial field. At the same time, Baidu Maps’ high daily activity data can help it upgrade more efficiently, frequently and accurately, and ultimately achieve the goal of truly helping vehicles achieve better autonomous driving.

Regarding the safety of autonomous driving, high adaptability and high fault tolerance are necessary. Just recently, Jidu has reached a strategic partnership with the world’s top automotive parts supplier, ZF. The purpose of this cooperation is to hope that the two sides can work together to create a set of intelligent chassis technology that is more suitable for automotive robots.

With the technology of this intelligent chassis, the goal that Jidu Motors wants to achieve is not only to upgrade the driving experience but also to upgrade the control of the chassis, and even achieve connectivity from the chassis to the intelligent cabin, including upgrading the core issue of autonomous driving – “safety redundancy”.

For Jidu Motors, a new car brand aiming to launch a mass-produced car on the market, with L4 level autonomous driving, by 2023, its first year of establishment could even be described as “speeding ahead”. Of course, challenges will inevitably still exist because autonomous driving technology is the ultimate goal, and the factors that hinder its implementation are also very complex.

However, a more trustworthy point is that Jidu Motors is at least promoting the development rapidly and solidly without worrying about funds. Such a car company is easier to have far-reaching ideals and less likely to be influenced by external factors that change its original intention.

For those who truly want to see, touch, and drive L4 level intelligent driving cars, Jidu Motors is trustworthy and the waiting time will not be very long. From now on, it’s just a short one-year wait.

This article is a translation by ChatGPT of a Chinese report from 42HOW. If you have any questions about it, please email bd@42how.com.