"This morning, Lao Yu and I discussed the technical route. I believe that this kind of technical debate will benefit users in the end. We will also expand the AEB function through OTA. Thank you Lao Yu for your advice and generosity. Sometimes it is easier to become good friends after a misunderstanding."

On November 9, with a post by He Xiaopeng, the half-month-long war of words between Xiaopeng and Huawei came to an end. The issue of "AEB" swept the entire industry, and the intelligent driving behind this debate was once again pushed to the forefront. With the rapid increase in the penetration rate of intelligent driving, ordinary users' expectations for smart cars are also getting higher and higher.

However, for ordinary consumers, what kind of intelligent driving technology do they need? Under the current legal framework, the intelligent driving technology that has already passed the singularity point will usher in the industry ceiling. The intelligent driving technology that almost every car company has may not differ much in terms of user experience. What will be the key factor that affects their orders?

Autonomous driving transitions to L3, urban NOA involution

According to statistics from the Ministry of Industry and Information Technology, in the first half of this year, the market penetration rate of intelligent connected passenger cars equipped with assisted autonomous driving systems in my country has reached 42.4%, a significant increase from 34.5% at the end of 2022. It is expected that in the era of large models, the trend of accelerating automobile intelligence will become more obvious.

L3 is a key node in the transition from assisted driving to autonomous driving. The domestic industry is currently in an important stage of transition from L2 to L3. Representative functions of the L3 stage, such as highway and urban NOA (Navigate On Autopilot), have gradually begun to be implemented, and various car companies are also actively deploying high-end intelligent driving.

First of all, the new car-making forces are the vanguard of "NIO, Xiaoli, and Li-ion". NIO has started internal testing of the NAD system in June, and will equip the entire series with laser radars as standard to support urban NOA. In fact, NIO's NOA layout began in October 2020. This is the first high-speed NOA to be implemented in China, and it has been continuously upgraded since then. Starting from the NT2 platform, NIO began to provide users with NP software functions for free, but the NOP+ function adopts a paid model, which is also the first case in China.

Xiaopeng is in the first echelon in this regard. As early as 2021, the first version of Xiaopeng Motors' City DGP had been tested on closed roads. In September 2022, Xiaopeng Motors' City NGP landed in Guangzhou for the first time, and then expanded to five cities. On March 31, 2023, Xiaopeng fully pushed the XNGP (full-scenario assisted driving) 4.2.0 version, and all G9 MAX and P7i MAX users can get this software version. This system is an assisted driving system rewritten by Xiaopeng based on the BEV framework, and its capabilities are similar to Tesla FSD Beta. It is reported that Xiaopeng's goal this year is to land in 50+ cities.

At the third quarter earnings conference of Ideal Auto, Chairman Li Xiang said that the development of urban NOA on Ideal AD Max is progressing smoothly. Ideal plans to push the official version of AD Max 3.0 software to users at the end of the year, providing full-scenario NOA capabilities. At the same time, Ideal will push the official version of AD Pro 3.0 software to users in the first half of next year, and some of the algorithm capabilities of AD Max will be released on AD Pro. Li Xiang said that he is confident that he will become the first echelon of intelligent driving that has been verified by the market in the first half of next year.

The new M7 equipped with Huawei ASD realizes true de-mapping of highway and urban scenes. In December this year, NCA (Huawei's NOA) was implemented in all cities across the country. As a pilot assisted driving that does not require the assistance of high-precision maps, Huawei has put into practice the "heavy perception, weak map" direction that OEMs are doing for urban NOA.

In addition, different brands have different intelligent configurations for different target groups. How big are the specific differences among these key automakers?

When Xiaopeng's XNGP is running, it will compare the high-precision map, the vehicle positioning and the static network of visual output in real time. When the high-precision map is "wrong", it will switch seamlessly by real-time perception. The background map can run high-precision maps, or it can be mapless and rely only on SD navigation maps. Huawei's ASD integrates the GOD 2.0 network on the basis of BEV. The function of this system is to identify general obstacles, and the upgraded 2.0 system can more accurately identify special-shaped objects outside the whitelist.

Both of them first developed urban areas and then highways, and are in the first echelon of the industry, while Weilai first developed highways, and the ideal urban NOA is still in the research and development stage. In 2023, urban NOA has reached a new height. Smart driving has changed the concept of car owners for cars. It is no longer a "selling point" but a standard feature of the car. Who can seize the minds of users is another question.

With the advent of intelligent driving, cost-effectiveness becomes the key to competition

Relevant survey data show that the proportion of intelligent driving in the factors that are important in car purchase decisions has increased significantly, accounting for 10%. Although autonomous driving only accounts for 10% of all car purchase factors, it will become a key factor in the core car purchase decision-making of the automotive industry in the future. For consumers, in fact, many users do not know the classification standards of autonomous driving. Their core demand for autonomous driving is not how intelligent it is, but whether it can reduce the burden on drivers in a safe manner.

The survey found that the core demand of Chinese users for autonomous driving is to reduce driving fatigue, accounting for as high as 82.01% in 2021; the second is to free their hands and increase available time, accounting for 50.62%; the third is high technical content, accounting for 30.91%. At present, the proportion of autonomous driving in the R&D investment of automakers is increasing rapidly, and it has even become the core investment of some automakers. If automakers do not grasp the tone of autonomous driving R&D and blindly pursue extreme functions such as fully autonomous driving, they may gradually drift away from the actual needs of Chinese users.

In this context, all car companies have begun to compete on price, trying to attract consumers' attention through cost-effectiveness. How to reduce costs has become the key to the commercialization of intelligent driving. At the beginning of this year, Tesla set off a wave of price cuts, and brands such as Xiaopeng, Leapmotor, Nezha, and BYD all cut prices for promotions. In March, under the leadership of Hubei's "strongest car purchase discount season in history", all regions across the country, major car companies and even luxury brands BB A has joined the "price war".

In the consensus that "no car is not smart", the urban NOA mentioned above becomes a key factor for consumers to consider future car models. If the price is too high, consumers may also be "dissuaded".

Taking the 230,000 yuan Model 3 as a benchmark, when consumers buy a Model 3 with FSD function, the included urban NOA price is 21.8%. Currently, the urban NOA price inclusion rate of Chinese brands is controlled within 20%, and the prices are all lower than Tesla FSD level.

NIO currently only offers subscriptions to the NOP+ function, priced at 380/month. It plans to launch a fully functional version of NAD in the future, priced at 680/month, with price inclusion rates of approximately 11% and 18%; taking Xpeng's G6 as an example, the MAX version is priced 20,000 yuan higher than the PRO version (other items include leather seats), and the price inclusion rate of Xpeng City NOA is about 9%; Ideal's AD MAX function is also standard on the MAX model, with a price increase of 40,000 yuan over the previous level (including a screen + two speakers in the back row), with a corresponding price inclusion rate of approximately 11%; Huawei's smart driving costs are consistent with the pricing of the M5 Smart Driving Edition, Avita 11 and Alpha HI, with a subscription of 720/month, 7,200 yuan for an annual package, and 18,000 yuan for a buyout.

In addition to striving for price advantages in urban NOA, many car companies are compressing the production costs of the entire vehicle, trying to pursue higher cost-effectiveness by reducing costs in various links.

Yang Yuxin, Chief Marketing Officer of Black Sesame Intelligence, once said, "The 'price war' has been very clearly transmitted to upstream suppliers. Autonomous driving has bid farewell to the era of pursuing high performance. In the past, car manufacturers were pursuing high performance because high performance can bring selling points and traffic, but it may not necessarily bring commercial value in the short term. This year, it is obvious that car companies have changed from pursuing high indicators or high-performance autonomous driving functions to pursuing cost-effectiveness."

For example, the Feifan R7 directly cut off the hardware of high-end intelligent driving. When it was launched, the price was adjusted to an increase of 20,000 yuan for the top configuration, with the addition of one Orin (originally one) and one laser radar. The prices of other models were reduced by 20,000 yuan. Fatigue monitoring, automatic lane change, and 4D radar were changed from standard to optional. The Nvidia Orin chip was downgraded to Mobileye Q4H, and the 5G network was downgraded to 4G.

With the support of SEPA 2.0 Fuyao architecture, Xiaopeng Motors is trying to achieve intelligent platform cost reduction. It is reported that the development cycle of Xiaopeng Motors' new models in the future will be shortened by 20%, and the commonality rate of parts based on the architecture can reach up to 80%. The development and popularization cost of the XNGP intelligent assisted driving system is the highest. With the help of the intelligent platform, the comprehensive development efficiency of intelligent driving is improved by 30%, and the adaptation cost is reduced by 70%.

The front-end experience of consumers may just be a difference in numbers, but for car companies, it requires efforts from all links from hardware suppliers to software solution suppliers.

Too much is as bad as too little, keep your original intention

If we want to achieve the same development speed as electrification, it is imperative to reduce costs through intelligence. When intelligence becomes a standard feature that “everyone” can afford, what will consumer demand become?

Having a price advantage will naturally attract more consumers' attention, but beyond price, cars as a means of transportation still need to grasp the advantages of their performance. After all, no matter how smart they are, they still need to return to the driving experience and performance.

Among the many autonomous driving functions, Chinese users are most concerned about pedestrian collision warning (PCW), accounting for as high as 79.1%. In China, where jaywalking is rampant, avoiding collision with pedestrians is still the core demand of Chinese users. The second is automatic emergency braking (AEB), accounting for 71.96%, and then blind spot monitoring (BSD), accounting for 63.2%. These three functions are closely related to safety and are very practical, not flashy smart stunts.

This also reminds major automakers that in the process of intelligent transformation, they do not need to overemphasize "what others do not have, but what others have, but I am better at". BYD, the leader of new energy vehicles, has obvious shortcomings in the second half of intelligentization, and for this reason it has gradually been labeled a "traditional automaker".

In fact, BYD's "tradition" is more reflected in product strength from the user's point of view. For example, most of its products are equipped with three-year-old lithium iron phosphate blade batteries. Although they are indeed cheaper and safer, they still have defects in battery life and charging time. There are also deficiencies in power. Take BYD's latest Song L as an example. Except for the top configuration of 280,000 yuan equipped with dual-motor four-wheel drive, the rest are single-motor rear-wheel drive.

This is why BYD Song L has sparked controversy over "high price and low configuration" since its launch. The focus of the controversy is not on the level of intelligence, but on basic aspects such as performance, configuration, and battery life. This reflects that users do not have obvious differences in their micro-perception of the current level of intelligence. More users only require "having it". When intelligence becomes standard, they will still care more about the performance, battery life and other functions of the car itself.

Car companies must also grasp user needs on the road to intelligent transformation, and not lose sight of the main goal in pursuit of intelligence. This is like a smart TV. When the TV is smart enough, it will trigger users' calls for traditional TV. As a means of transportation, performance is always the key that car companies cannot give up.

As a winner of Toutiao's Qingyun Plan and Baijiahao's Bai+ Plan, the 2019 Baidu Digital Author of the Year, the Baijiahao's Most Popular Author in the Technology Field, the 2019 Sogou Technology and Culture Author, and the 2021 Baijiahao Quarterly Influential Creator, he has won many awards, including the 2013 Sohu Best Industry Media Person, the 2015 China New Media Entrepreneurship Competition Beijing Third Place, the 2015 Guangmang Experience Award, the 2015 China New Media Entrepreneurship Competition Finals Third Place, and the 2018 Baidu Dynamic Annual Powerful Celebrity.