2025 Outlook: Cars Without Intelligent Driving to Become "Outliers"

Today we're delving into the 2025 outlook, focusing on intelligent driving, the endpoint of the next mobility era.

In 2015, autonomous driving was a rare sight in sci-fi films, but a decade later, we've entered the age of assisted driving, where intelligent driving has become a key measure of a carmaker's technical prowess.

In 2025, high-level intelligent driving will shift from the mid-to-high-end market of the past five years to the broadest and most potential market of 100,000 - 200,000 RMB vehicles. In short, intelligent driving will become a standard feature for most cars. This is similar to mobile phones gaining camera functions around 2000.

Currently, Huawei Vehicle BU CEO Jin Yuzhi believes that "2024 is the first year of large-scale commercialization of intelligent driving."

From 2024 to 2025, the tangible progress of intelligent driving includes BYD's 2025 models being fully intelligent, XPeng bringing high-level intelligent driving to vehicles under 200,000 RMB, and Zhuoyu Technology's intelligent driving solution costing as low as about 5,000 RMB.

In 2025, intelligent driving will be marked by scale and popularization. It has become the "camera function" of today's vehicles, except for micro cars priced at 50,000 RMB.

Looking to 2025, will a family car without intelligent driving become an "outsider"? How will high-level intelligent driving permeate every automotive market segment?

The Greatest Shift in 2024

The biggest change in the field of intelligent driving in 2024 was the shift from lidar to vision - based systems, led by China's top players. This change is backed by lower hardware costs, better software, and a leap in computing power.

The rise in computing power is especially significant. In less than ten years, the computing power of a single intelligent - driving chip has grown from 2. 5TOPS (Mobileye EyeQ4) to 254TOPS (Nvidia Orin X). Horizon Robotics' J6P even reaches 560TOPS. Orin X is now a key part of intelligent connected vehicles, with over one million units shipped yearly in China. Mid - to - high - end models are particularly reliant on Nvidia chips.

Cost - cutting by automakers in a competitive market creates new opportunities for other chip makers.

Price is key.

Per media reports, top intelligent - driving suppliers say Nvidia Orin X costs around $500 (about 3,642 RMB), with "very high gross margin".

Most suppliers need two Orin X chips for advanced intelligent - driving functions. Some automakers now use one Orin X, like Le Dao L60, which offers full - scene navigation assistance with a single Orin X.

Beyond Orin X, the new - energy vehicle industry, aiming to popularize intelligent - driving functions, is seeking more cost - effective solutions.

In recent years, chips like TI TDA4 VH, Horizon Robotics' Journey 5, and Black Sesame's Huashan series have gained automakers' favor.

In 2023, high - compute - power SoC chip shipments in China and globally were 1.5 million and 1.6 million units respectively. Horizon and Black Sesame ranked second and third. Although Nvidia still has an absolute advantage, the challenge from latecomers is evident.

Chinese firms are good at using low - prices to target high - end markets. To gain more market share, seeking technological progress and cost - effectiveness is crucial.

Horizon Robotics' Journey 6 series is a star product, with J6E, J6M, and J6P offering 80TOPS, 128TOPS, and 560TOPS respectively.

It's reported that the autonomous - driving domain controller cost for J6E is expected to be 2,000 - 3,000 RMB. Zhixing Technology says its iDC510 system based on J6E can be kept at around 5,000 RMB.

Hardware costs in the intelligent - driving field have declined since 2024 and will continue into 2025.

Even lidar, not used in vision - based solutions, has seen its cost drop from 5,000 RMB to 1,000 - 3,000 RMB due to scale, and the cost of millimeter - wave radar has fallen from 50 RMB to 20 RMB.

If Moore's Law rapidly boosts the compute power and experience of intelligent - driving chips, then for high - level intelligent - driving hardware, it's like Momenta CEO Cao Xudong said, "The BOM will halve every two years."

Tongji University's Professor Zhu Xichan noted that NOA costs have dropped from 70,000 RMB, going through prices like 35,000 RMB, 20,000 RMB, and 8,000 RMB. He expects that by 2025, the industry will launch NOA products with a BOM cost of under 3,000 RMB.

Entry of New Supply Chains

One of the major events in the automotive supply chain in 2024 was the large-scale layoffs by traditional tier-one suppliers like Bosch and ZF Friedrichshafen.

Last November, Bosch laid off 5,000 people and announced plans to cut 8,250 jobs in its automotive department over the next few years. ZF Friedrichshafen plans to lay off 12,000 people by 2028.

The layoffs at Bosch and ZF Friedrichshafen reflect the difficulties faced by traditional automotive suppliers, especially in Europe. Continental and Schaeffler, among other traditional suppliers, have also announced layoff plans of at least four digits.

The advent of new - energy vehicles has upended the traditional automotive supply chain.

The traditional automotive supply chain, pyramid - structured, had clear - cut hierarchical separations between suppliers and automakers, with T1 suppliers maintaining high independence and pricing power due to core technology barriers.

However, electrification and, more so, intelligentization have not only shifted demand but also transformed the linear supplier - automaker relationship into an integrated, networked one.

This integration requires suppliers to closely collaborate with automakers in the fast - paced new - energy vehicle era, engaging more openly in product design, R&D, and manufacturing. They must speed up their responses and iterations.

A typical example of this close relationship is the emergence of intelligent - driving suppliers. In the software - defined vehicle age, these suppliers play a key role in the new automotive supply chain.

Among new supply chain participants, Mobileye's story from 2020 - 2022 is a classic example of adapting to China's market changes and personalized needs.

Before 2019, Mobileye's Mobileye EyeQ4 visual solution, economical and efficient, won most automakers' favor, even giving it a 70% global market share.

As intelligent driving evolved, automakers demanded more functions. Mobileye's closed algorithms couldn't keep up, losing share to Nvidia, Horizon Robotics, and Qualcomm.

Later, Mobileye launched the EyeQ Kit, offering personalized designs based on automakers' needs and application scenarios. Though it couldn't fully regain its former glory, it still secured orders from VW, SAIC, Geely, and Great Wall.

Mobileye's earlier missteps highlight the importance of an open attitude for suppliers today. Being insightful about demand changes and adjusting promptly are essential skills for the new supply chain.

Automakers are competing in higher - level tech, and intelligent - driving solution suppliers sensitive to market changes are also adjusting.

Zhuoyu Technology's CEO Shen Shaojie revealed at the China EV 100 Forum on March 17, 2024, that the Chengxing platform equipped with inertial navigation binoculars and the TDA4 VH chip has a total cost of about 5,000 RMB for the 7V + 32TOPS solution, which can achieve urban memory navigation and driving. The overall hardware cost for the 7V + 100TOPS solution with urban driving capabilities is 7,000 RMB.

Momenta, working with BYD, Aion, and others, is bringing urban NOA to vehicles in the 200,000 RMB segment. Its CEO said that by the end of 2025 or early 2026, the cost of high-level intelligent driving is expected to drop to around 5,000 RMB, making urban NOA a potential standard feature.

Huawei Vehicle BU has also deeply participated in the intelligent empowerment of the new energy vehicle era through models like HI, HarmonyOS SmartDrive, and supplier partnerships, securing orders from automakers like Sibü, Avita, and岚图.and suppliers is more flexible now, and it's too early to determine the winners.

When selecting Huawei Vehicle BU's intelligent driving solutions, automakers can choose products with different capabilities and price points based on their situation and needs, such as the ADS SE, which retains features like highway and urban fast driving and smart parking, and the full - featured ADS3.0.

The new supply chain injects vitality into the automotive supply chain's transformation. Currently, some automakers excel in intelligent driving while others lag. In this situation, in - house development and procurement of intelligent driving solutions coexist. The diverse intelligent driving suppliers offer automakers more choices.

The relationship between automakers and suppliers is more flexible now, and it's too early to determine the winners.

Change the Statement

As intelligent driving has risen in consumers' decision-making criteria from 13th to 5th place, and with Huawei and Xiaomi entering the market to boost consumer acceptance of intelligent driving, especially during the growth period of highway NOA, the adoption of mid - level intelligent driving capabilities became a major 2024 event.

In 2024, car brands' "retraction" on intelligent driving was the most common behavior we saw.

Among the many attitude changes dubbed "boomerangs" by netizens, traditional Chinese automakers' shift stood out.

In 2021, Chen Hong, then chairman of SAIC Motor, said:

"SAIC can't accept collaborating with Huawei - like third parties on autonomous driving. That would make them the soul and us the body. We can't allow it; we must keep the soul in our own hands."

The "soul" statement made SAIC Motor a focal point. Yet, by late 2024, rumors of a new cooperation model with Huawei, beyond the supplier, HI, and HarmonyOS SmartDrive modes, sparked market speculation.

SAIC Motor's three - year attitude shift is closely linked to industry and its own development pace.

In 2024, SAIC Motor's total vehicle sales fell by 20.07% year - on - year to 4.913 million units.

SAIC Motor's 9.9% new - energy - vehicle sales growth in 2024 couldn't offset its overseas - market and domestic - passenger - vehicle - market woes. Rising overseas barriers have made boosting domestic demand a key way out for SAIC Motor.

Though details of SAIC Motor and Huawei's cooperation aren't out yet, Huawei's tech and consumer - acceptance edges in the intelligent field might become a major growth driver for SAIC Motor's new - energy business.

The absolute purity of the "soul" has room for maneuver in the face of fierce competition.

As SAIC Motor's biggest challenger, BYD's chairman once criticized unmanned driving as "nonsense" and "a fraud" like "The Emperor's New Clothes" during a 2023 financial meeting.

But about a year later, at BYD's Dream Day on January 16, 2024, he clarified that "unmanned driving is not yet ripe, but intelligent driving is here."

Since then, there have been frequent changes in BYD's intelligent driving team. By the end of 2024, the MIIT's rule allowing new cars to optionally install an interior rearview mirror base means that BYD's three-camera inertial navigation visual perception system, or "Eye of the Gods," will be widely installed in its models.

Great Wall is also becoming more intelligent. At the Guangzhou Auto Show in November 2024, it displayed 23 new WEY Blue Mountain models in the Cyan color.

Great Wall Vice President Liu Yanzhao said the new Blue Mountain is “the culmination of Great Wall’s intelligent technologies” and is key to its intelligent transformation. “As a latecomer in intelligent driving, we must show our determination and offer products that surpass previous ones,” he noted.

Great Wall Chairman Wei Jianjun, who rarely appears in public, tested the Coffee Pilot Ultra system on a live broadcast in April last year, showing his determination.

Outside the three aforementioned automakers, Geely is integrating intelligent - driving resources and finalizing R & D plans for its brand and Zeekr. Chery, after launching the Da Zhu intelligent - driving system in April last year, announced in October that advanced intelligent - driving would be installed in over ten models for mass production by 2025.

In this fiercely competitive phase, keeping pace with intelligent - driving market development is the top priority, rather than just discussing changes in stance.

A New Era

At the end of 2019, we first experienced "Navigation Assist" on the Model 3, getting a feel for the "computer driver's" capabilities.

By the end of 2020, Chinese new - car makers started rolling out highway navigation functions. Then, in the 2022 urban navigation battle, they caught up with Tesla.

At the end of 2024, Tesla and Chinese new - car makers, in their home markets, declared "parking spot - to - parking spot" driving, aiming for nationwide and eventually global usability.

More importantly, this intelligent - driving trend is spreading from the mid - to - high - end market to the "silent majority."

According to data from GaoGong Intelligent Automotive Research Institute, from January to September 2024, the proportion of models in the 200,000 - 300,000 RMB price range with factory - installed standard NOA was 41.25%, and by September, the delivery volume of models in this price range equipped with NOA had surpassed that of models priced above 300,000 RMB.

After the delivery of the XPeng P7 +, Le Dao L60, and Zeekr 7X, the data for models in the 150,000 - 300,000 RMB price range is expected to be even more remarkable. This is not to mention the huge impact that will be generated by the continued efforts of Huawei ADS SE, BYD's Eye of the Gods, and other companies such as XPeng and Li Auto.

The introduction of intelligent - driving functions to models in the 100,000 RMB price range, making it a standard feature for more and more new models, marks the gradual realization of "technological egalitarianism" in the field of intelligent driving. It is finally becoming a tangible reality and entering the homes of ordinary people.

The new phase of intelligent driving brings fresh competition. The December 2024 drop in new-energy vehicle penetration rate to below 50% suggests the game is far from over.

With more strong players entering and making moves, what will the intelligent - driving landscape look like in 2025?