The Battle for Breakthroughs in Smart Driving Technology Among Traditional Automakers

Recently, Huawei's 1,000-person R&D team has been stationed in Chongqing, deeply involved in the development of the Changan Avita's intelligent driving, intelligent cockpit, and application systems, attracting industry attention. In building the Avita brand, Huawei will provide a full-stack intelligent driving solution. An industry insider likens it to cooking: "The vendor prepares the ingredients and seasonings, and you just need to stir-fry them."

This kind of in-depth collaboration between Huawei and Changan is becoming the norm in the automotive industry. The era of "software-defined cars" is approaching, and autonomous driving technology is undoubtedly at the core. The development of autonomous driving technology is inseparable from data; it can even be said that data determines success or failure. While traditional automakers possess vast amounts of driving data, their ability to effectively utilize and integrate it into intelligent technologies is currently relatively weak. Collaboration with companies like Huawei can quickly address these shortcomings in intelligent driving software technology.

However, the iron must be forged with strength. In the long run, traditional automakers will continue to need to develop their own intelligent driving solutions. Only by mastering core technologies can they remain viable in the future automotive competition. Otherwise, they risk becoming mere puppets of intelligent driving solution providers. When discussing the future development of the automotive industry, NIO co-founder Qin Lihong once stated, "The key to success in competition lies in intelligence, and the core of intelligence lies in autonomous driving." Internet-connected car companies, represented by Tesla and NIO, have overtaken competitors in the new energy vehicle market in just a few years. The key lies in seizing the development opportunities of intelligent and electrified technologies. Intelligence has become the defining characteristic that distinguishes new car manufacturers from traditional ones and is the foundation of their success.

Currently, there are two main technical approaches for autonomous driving: a purely visual perception solution based on cameras; and a multi-faceted approach that integrates cameras and sensors. The former requires extremely high data and algorithm requirements, and is currently only used by a few companies, such as Tesla. The latter, which utilizes a "camera + lidar/millimeter-wave radar" approach, is becoming mainstream.

In April 2019, Tesla unveiled its third-generation Autopilot system, powered by a self-developed FSD chip. The following year, Tesla rolled out the FSD Beta version, and this year is set to complete the V11 software update. Although Tesla has moved away from technologies like LiDAR, its advanced autonomous driving technology remains at the forefront, thanks to innovations such as the "HydraNet" multi-headed neural network architecture, an extensive dataset accumulated from millions of users, and unparalleled computing power—thanks to Dojo, the world's most powerful supercomputer entirely designed in-house by Tesla. This year, Tesla continues to strengthen its self-developed advantages across key areas, including chips (adopting the HW4.0 chip and introducing the D1 chip), sensors, and cutting-edge algorithms. Meanwhile, automakers like NIO, Xpeng, and Li Auto have chosen a diversified approach that integrates multiple technologies. Notably, Xpeng is developing its own intelligent driving solution—known as the XPilot system—which is expected to evolve to version 3.5 in the second half of this year.

Like XPeng, NIO and Li Auto have also regarded the development of autonomous driving technology as essential since their very beginnings. NIO has built the NIO Aquila super-sensing system and the NIO Adam supercomputing platform, featuring 33 high-performance sensing hardware components and four NVIDIA DRIVE Orin chips capable of delivering up to 1016 TOPS of computing power—ready for imminent product integration. Meanwhile, Li Auto’s vehicles delivered last year already come equipped with NOA navigation-assisted driving and comprehensive AEB (Advanced Emergency Braking) features, marking a significant step toward their full-stack, in-house R&D transformation in the realm of intelligent driving.
Looking back at traditional automakers, their presence in the autonomous driving space pales in comparison to that of new automotive forces like Tesla. Changan Automobile was among the first automakers to strategically invest in autonomous driving; however, its advanced driver-assistance technologies—currently focused on specific scenarios such as urban highways—are still in the pilot-production phase and are yet to be fully rolled out to the market.

Like Xpeng, NIO and Li Auto have considered the development of autonomous driving technology essential from the outset. NIO has built the NIO Aquila super-sensing system and the NIO Adam supercomputing platform, which include 33 high-performance sensing hardware components and four NVIDIA DRIVE Orin chips (with 1016 TOPs of computing power), and are about to see production applications. Last year, Li Auto delivered products equipped with NOA navigation-assisted driving and comprehensive AEB (active safety) functionality, demonstrating its transition to full-stack in-house development of intelligent driving.

Traditional automakers, by contrast, have a much smaller presence in the autonomous driving field than new entrants like Tesla. Changan Automobile was arguably one of the first companies to embrace autonomous driving, but its assisted driving technology for specific scenarios like urban highways is still in the pilot phase of mass production and will take time before it can be fully commercialized.

So we can see that, as the new wave of automotive industry transformation increasingly shifts toward the second half—now centered on intelligent technologies—more and more traditional automakers are adopting multi-pronged strategies to position themselves in the autonomous driving space. On one hand, they are actively pursuing independent research and development of core technologies, either by establishing dedicated R&D teams internally or even setting up separate subsidiaries. At the same time, they are vigorously advancing collaborations with relevant companies while simultaneously ramping up investments in related fields.
For instance, while SAIC Group has invested in companies like Momenta, it has also incubated Zero-Bound Technology. Meanwhile, BYD is simultaneously developing its proprietary DiPilot advanced intelligent driving assistance system, while also collaborating with autonomous driving technology firms such as NVIDIA and Horizon Robotics. GAC Capital, part of the GAC Group, invested in autonomous driving company Hodo Technologies this past March and plans to launch multiple vehicle models equipped with Hodo’s self-driving system within the year. As for Great Wall Motor, it has expanded into the Robotaxi sector, with its subsidiary Weimo Intelligent Mobility aiming to deploy a fleet of Robotaxis powered by the HSD (HAOMO Self-Driving) system as early as 2023.

Some automakers have even chosen to form joint ventures with tech companies, leveraging each other's technological expertise and resource advantages to jointly drive the development of smart cars. For instance, SAIC has teamed up with CATL and Alibaba to create IM Motors, while Changan has partnered with CATL and Huawei’s Avatr brand. Notably, both Alibaba and Huawei are playing, to varying degrees, the role of providers of intelligent driving solutions. Clearly, traditional automakers are actively working to build robust ecosystems for smart driving, making leading companies like Baidu, Huawei, and Horizon Robotics highly sought-after partners. After all, as intelligence takes center stage in the second half of automotive competition, autonomous driving technology has emerged as a key differentiator. Back in 2019, McKinsey predicted that by 2030, autonomous driving technologies alone could generate up to $1.6 trillion in revenue—primarily within urban areas.
Analysts at the Gaishi Automotive Research Institute noted that in the past, chassis, powertrains, and styling design were considered the core elements. For automakers or component suppliers, ensuring top-notch systems and hardware would essentially help maintain healthy gross profit margins. However, now the entire industry has entered a transformative phase, where hardware, software, advanced algorithms, sensors, and even user services will become the key competitive factors for automakers and component companies moving forward. Therefore, in the long run, traditional automakers developing their own autonomous driving technologies and crafting proprietary, customer-centric autonomous driving solutions will be the critical weapon for sustaining revenue growth in the future automotive landscape. That said, it’s important to note: How should automakers define the boundaries of in-house R&D? Which technologies are best suited for collaboration with industry experts, which can be jointly developed, and which must be pursued exclusively through internal research—these are all crucial considerations for automakers as they navigate this evolving landscape.

Currently, as smart cars are moving toward hardware standardization, automakers are primarily focusing on strengthening their R&D capabilities in software and algorithms to build differentiated competitive advantages. For instance, CARIAD, Volkswagen Group's software company, has established a Chinese subsidiary and assembled a team to independently develop a brand-new software platform, advanced driver-assistance systems, autonomous driving technologies, and next-generation intelligent connectivity features. Meanwhile, Zero-Bound Technology, part of SAIC Motor Group, is developing the SOA software platform and a centralized electronic architecture, with the Galaxy Full-Stack 1.0 solution already completed earlier this year. Additionally, targeting intelligent driving solutions for commercial vehicles, SAIC Group also founded Shanghai Youdao Zhitu Technology Co., Ltd. last year.
Considering the critical importance of chips, in addition to new automotive forces like Tesla and NIO, traditional automakers represented by SAIC and Geely are also actively making strategic moves. Beyond investing in semiconductor companies such as Horizon Robotics and Black Sesame Technologies, SAIC Group is accelerating its efforts to achieve domestic production of high-performance computing chips and MCU chips. Meanwhile, Xinqing Technology, a subsidiary of Geely Holding, has designed the 7nm automotive-grade smart cockpit chip "Longying No.1," which is expected to enter mass production in the third quarter. Currently, traditional automakers have fully launched their battle to break through in the field of intelligent driving technology. By 2025, China's autonomous driving technology is poised for a significant breakthrough. Earlier this year, the National Development and Reform Commission released the "Strategic Plan for Innovative Development of Smart Cars," outlining that by 2025, conditionally automated smart cars will reach large-scale production, while highly automated smart cars will begin commercial applications in specific environments. With less than three years remaining, traditional automakers now face an increasingly tight race to catch up in the rapidly evolving autonomous driving arena.

Translated from Sina Auto

Email:
info@adtfm.com

Phone:
021-20532053

Address:
2nd Floor, No. 150 Jindian Road, Pudong New Area, Shanghai