Author | Pu Zixu

On April 28, 2024, a Gulfstream jet flew across the ocean and quietly landed in Beijing.

On this day, the plane door opened, and a tall man with an Eastern appearance dressed in a black suit slowly walked out: he looked haggard, with a beard covering his face, seemingly weighed down by worries. This was not the first time Musk had encountered trouble. Since his official visit to China in 2014, Tesla, which Musk controls, has encountered a series of issues such as sluggish sales, production hell, and market expansion challenges. However, every time he came to China, these problems seemed to be resolved effortlessly. But in the spring of 2024, simply promoting FSD (Full Self-Driving) in China might only address half of Musk's concerns. The other half lies in when Robotaxi, or self-driving taxis, will truly become a reality.

By October 10 of this year, with Tesla's continuous advancements in its no-high-precision-map + pure-vision solution (FSD) technology, Tesla's Robotaxi is set to be officially launched, ushering in the second growth phase for Tesla.

Coincidentally, on the other side of the ocean in China, enthusiasm for self-driving is also rapidly surging. In fact, China and the United States have undoubtedly been the leaders in the global Robotaxi industry.

As American companies ignite a new wave of autonomous driving in October, Chinese companies are quick to follow, bringing the flame of autonomous driving ignited in Silicon Valley to China in a more feasible manner, instantly intensifying the competition between China and the United States in the field of autonomous driving.

On October 9, just before the launch of Tesla's Robotaxi, Luobo Kuaipao actively pursued a global strategy, conducting in-depth communications with multiple international companies and planning to enter overseas markets. Additionally, with the upcoming release of Apollo's autonomous driving open platform version 10.0, there was a collective surge in the share prices of companies across the software development, semiconductor, and other related industries.

So, what is the magic of Robotaxi?

What are the obstacles faced by Robotaxi companies?

In the spring of 2024, Tesla's repeated delays in launching its mid-to-low-priced models sparked panic among the company's sales figures, pushing Musk into a corner and leading to widespread pessimism in the market. However, this is merely the surface of the problem. In reality, the valuation ceiling for traditional automakers is exemplified by Toyota: decades later, with a PE ratio of over ten times and a market capitalization of several hundred billion dollars, it has already reached its peak. Looking back at the years when Tesla's share price soared, it is evident that the capital market's valuation of over a trillion dollars for the company did not follow this logic. After years of growth, Tesla's market value was once again questioned in 2024, prompting Musk to address a crucial question: What kind of company is Tesla? Is it an automotive company or a high-tech firm?

The valuation logic in the capital market will vary significantly based on this distinction. "If you view Tesla as a traditional automaker, you have the wrong framework. What I mean is, if someone doesn't believe Tesla will solve the automation problem, I don't think they should invest in this company," Musk has repeatedly and forcefully stated on multiple occasions in response to criticisms from Wall Street. Over the past year, with the automotive business facing obstacles, Musk urgently needed to find a second growth path to steer Tesla through difficult times.

Among these options, self-driving taxis, or Robotaxi, undoubtedly became a key focus for Musk. In the biography of Musk, his grand vision is recorded: to create a Tesla vehicle capable of fully autonomous driving without any human intervention. In fact, as early as 2016, Tesla outlined the concept of shared taxis in its Master Plan 2: operating its own fleet, with Tesla owners also having the option to join, allowing their cars to generate income by driving autonomously during idle times. This ambitious vision has long served as a panacea for "rescuing" Tesla's market value: for instance, during the first-quarter earnings call in 2024, Musk announced that Robotaxi would be launched in August of that year; the news was updated again during the second-quarter earnings call. Both announcements sparked a frenzy in the capital market, leading to a swift rebound in Tesla's share price. In fact, with the rise of deep learning and computer vision technology, autonomous driving, with its immense potential to address traffic safety and efficiency issues, has mostly been warmly embraced by the capital market.

The allure of Robotaxi not only shines on Tesla but has also sparked interest from numerous companies worldwide. It can be said that Robotaxi, which integrates technologies such as artificial intelligence, communications, semiconductors, and automobiles, involves a long industrial chain and significant value creation space, making it a hotly contested arena for cross-border competition and cooperation between the automotive and technology industries of various countries. However, the path from concept to reality has also presented numerous obstacles for Robotaxi. For example, representatives of autonomous driving in Silicon Valley, including Zoox, Waymo, Cruise, Argo AI, and Aurora, have encountered various setbacks over the years in this field. Argo AI, despite burning through over $2 billion, ultimately failed; Zoox was acquired by Amazon at a 50% discount, valued at $1.3 billion; Aurora was not well-received by the capital market after its IPO, reporting a net loss of $792 million in 2023; and Cruise was suspended due to a serious traffic accident.

It can be said that factors such as unprofitable spending, immature business models, the limited maturity of autonomous driving technology, and safety concerns have long been "roadblocks" for many Robotaxi companies. Nowadays, after experiencing a burst of capital, the bursting of the bubble, and industry consolidation, the market has returned to calm. At this juncture, finding the key to advancing Robotaxi and achieving genuine real-world applications has become paramount in this gentle slope of value growth.

The Comprehensive Battle for the True Realization of Robotaxi

Faced with these challenges, Tesla's Robotaxi and Google's Waymo, the two US autonomous driving giants, have prescribed advancements in autonomous driving technology as the solution. Roland Berger previously proposed five core elements for the realization of Robotaxi, including government regulation, technology, cost, operation and service, and market acceptance. Among these, advancements in autonomous driving technology play the most critical role. Only vehicles with mature technology can adapt to various complex environments, enabling long-term and safe autonomous driving. On the other hand, technological advancements in autonomous driving significantly reduce costs and human input, thereby closing the loop for real-world applications. Based on this, policy incentives and user acceptance of new technologies will naturally follow. US companies with a head start in autonomous driving technology research and development have begun a collective leap forward this October. For instance, with Tesla's continuous advancements in its no-high-precision-map + pure-vision solution (FSD) technology, Tesla's Robotaxi was officially launched on October 10 of this year, bringing Musk closer to discovering Tesla's second growth curve. In fact, Tesla advocates for an "end-to-end" FSD approach, utilizing neural networks for vehicle control, which requires significant data accumulation and feedback to make autonomous driving "smarter." However, with Musk's 2024 visit to China solidifying the comprehensive rollout of FSD in the country, Tesla's data advantage is now even more pronounced, significantly propelling its autonomous driving progress.

In fact, as of July 2024, Tesla's FSD had accumulated over 1.6 billion miles of driving. Meanwhile, Tesla plans to launch FSD V13 in October, with a necessary takeover mileage increased sixfold. In terms of computing power, Tesla's AI capabilities continue to strengthen, reaching nearly the equivalent of 90,000 H100 chips in 2024 to support autonomous driving training. Apart from Tesla's rapid progress, on October 2, another major US autonomous driving giant, Google's Waymo, also celebrated a "harvest" with its "single-vehicle intelligence" model: the company announced that its self-driving taxi service in Austin, Texas, would soon be open to the public. The service will initially be offered to select users through the Waymo One app and will transition to the Uber app in early 2025.

Furthermore, Waymo is collaborating with China's Geely Group subsidiary Zeekr to develop new vehicles specifically designed for autonomous driving. Currently, both parties are advancing tests of the sixth-generation Waymo Driver in San Francisco. Waymo, which has been operating since 2009, has long been recognized as one of the leaders in the autonomous driving industry. In July of this year, Alphabet announced an additional $5 billion investment in Waymo, meaning that since 2020, Waymo has received over $10 billion in funding, providing ample resources for the next phase of competition in autonomous driving.

It can be said that the recent rapid progress and application of US companies in autonomous driving have undoubtedly put greater pressure on autonomous driving enterprises worldwide.

The "Foreign Carrot" and "Local Carrot" of Autonomous Driving

It is noteworthy that as American companies ignite a new wave of autonomous driving in October, Chinese companies are quick to follow, bringing the flame of autonomous driving ignited in Silicon Valley to China in a more feasible manner, instantly intensifying the competition between China and the United States in this field.

One day before Tesla's Robotaxi announcement, on October 9, Baidu's autonomous driving mobility service platform "Luobo Kuaipao" began to contemplate a global strategy, engaging in communications with multiple international companies and planning to enter overseas markets. Meanwhile, the upcoming release of Apollo's autonomous driving open platform version 10.0, equipped with Baidu's latest autonomous driving large model ADFM, will significantly enhance the platform's safety, intelligence, and ease of use.

The latest moves by Chinese autonomous driving companies, represented by Luobo Kuaipao, have also triggered a sustained rise in the share prices of related companies across the industrial chain, reigniting enthusiasm in China's autonomous driving market. In fact, domestic autonomous driving companies have developed rapidly in recent years, with a large number of players. Generally speaking, domestic Robotaxi companies can be broadly classified into autonomous driving technology companies and traditional automakers, with technology companies currently leading the way.

From the perspective of cities where services are offered and testing mileage, technology companies such as Luobo Kuaipao, WeRide, and Pony.ai are relatively advanced. Traditional automakers, though starting later with road tests, are actively seeking partnerships with autonomous driving solution providers to advance the commercialization of Robotaxi. Notably, compared to Tesla's Robotaxi and Google's Waymo, Baidu's Luobo Kuaipao boasts technological differences, earlier and broader service implementations, and robust safety features, effectively competing head-to-head with the two giants and forming a three-way race.

In terms of technology, which has garnered significant attention, Luobo Kuaipao has not blindly followed the two major US autonomous driving giants in recent years but has instead independently developed a distinct technical approach. For instance, while Google continues to focus on "single-vehicle intelligence," Luobo Kuaipao actively promotes a "single-vehicle intelligence + vehicle-road coordination" solution."At the 2024 ApolloDay conference, Luobo Kuaipao unveiled Apollo ADFM, the world's first large model supporting L4 autonomous driving. This model can reconstruct autonomous driving based on large model technology, ensuring safety that is over ten times higher than human drivers, covering complex urban scenarios, and enhancing data processing efficiency through self-annotation. Furthermore, the sixth-generation driverless vehicle from Luobo Kuaipao is equipped with a battery-swapping solution and integrates Baidu Apollo's sixth-generation autonomous driving system solution, fully utilizing the "Baidu Apollo ADFM Large Model + Hardware Products + Security Architecture" to ensure vehicle stability and reliability through ten levels of safety redundancy and six levels of MRC safety strategies, with a mass-production price below $30,000."From the first to the sixth generation, Luobo Kuaipao has collaborated with automakers such as BMW, Chery, FAW, Beijing Automotive Group, and Jiangling Motors, gaining market validation. Meanwhile, compared to the "foreign carrots" of autonomous driving from across the ocean, China's "local carrots" have accumulated more experience in application practices and safety. Co-founded with Waymo, Baidu's Luobo Kuaipao has provided over 7 million safe rides across 11 cities in China as of July 28 this year, with testing and operating mileage reaching 100 million kilometers. Since 2021, Luobo Kuaipao has opened manned tests in 11 Chinese cities, including Beijing, Shanghai, Guangzhou, Shenzhen, Chongqing, Wuhan, Chengdu, Changsha, Hefei, Yangquan, and Wuzhen."In terms of cost control, which significantly impacts commercialization, Tesla adopts a "Lego-like" approach to producing Robotaxi, significantly reducing production costs. Luobo Kuaipao, on the other hand, focuses on operational efficiency and charging speed. Currently, the purchase price of Luobo Kuaipao's sixth-generation driverless vehicle is RMB 204,600, a 60% decrease from the previous generation, with an expected service life of five years. In terms of operational costs, Luobo Kuaipao has built an autonomous operation network for driverless vehicles, enabling full-lifecycle service automation, including one-click cloud commands to automatically wake up vehicles, self-inspection, automatic dispatch, and autonomous parking, all without human intervention.Currently, Luobo Kuaipao has over 30 fully automated management units. When scaled up to the entire fleet, the reduction in operating costs from these automations is significant. Additionally, Luobo Kuaipao's sixth-generation driverless vehicles will adopt a battery-swapping model. As a result, all vehicles will be automatically dispatched to battery-swapping stations based on their service status, location, and battery level, with the entire process taking only three minutes. The automation of the battery-swapping process significantly reduces labor costs."From this perspective, the intense updates from Tesla's Robotaxi, Google's Waymo, and Baidu's Luobo Kuaipao in October are undoubtedly a battle for the global autonomous driving flag, reflecting the intense competition between China and the United States in this field. With rapid advancements in deep learning and computer vision technology in recent years, the competition in autonomous driving has evolved beyond a simple technology race or corporate rivalry. Instead, it encompasses a comprehensive "war" across industries such as artificial intelligence, communications, semiconductors, and automobiles. While Chinese companies like Luobo Kuaipao have made rapid progress, it is undeniable that US companies have an unparalleled first-mover advantage in autonomous driving technology research and development compared to the rest of the world."For instance, the United States boasts advanced integrated circuit technology and has maintained a leading position in high-end chip design, laying a solid foundation for the development of high-performance automotive chips. Whether it's the "Waymo faction" excelling in single-vehicle intelligence or the "Tesla faction" driving hardware with software, both rely on core capabilities in artificial intelligence algorithms and decision-making chips, which are advantages that the United States is unlikely to relinquish in the short term. Facing the aggressive US companies, Chinese enterprises undoubtedly need to accelerate their pursuit with greater urgency."The competition between the two giants: The outcome of the Robotaxi race remains uncertain

Fortunately, due to Robotaxi's unique nature, technology is not the sole criterion for evaluation. To effectively and sustainably deploy a large fleet of driverless vehicles for commercial use, a comprehensive balance must be struck across safety, cost, profitability, and other factors. Moreover, the level of national infrastructure development, the effectiveness and accuracy of policy implementation, and market size all play crucial roles in this competition. In fact, in recent years, China and the United States have led the world in the large-scale application of Robotaxi due to their relatively advanced technological capabilities and early policy deregulation. Cities such as Wuhan, Shanghai, San Francisco, and Phoenix have already implemented Robotaxi services. On this basis, a preliminary two-power dynamic between China and the United States in Robotaxi has emerged. Although China's autonomous driving research started later than that of the United States, its broader market, stronger infrastructure development, and rapid progress in full-scenario technologies suggest that China still has the potential to catch up and overtake. The final outcome of the autonomous driving race between the two countries remains uncertain.

For example, in terms of supporting industries, relying on China's vigorous promotion of new infrastructure construction such as 5G, satellite Internet, data centers, and intelligent transportation, the 5G LTE-V2X technical standard is promoted in road renovation to support the smooth evolution of LTE-V2X to 5G-v2X. In addition, China places greater emphasis on the synchronized development of intelligence and networking, and constructs an overall solution of "person-vehicle-road-cloud" with networking capabilities. Undoubtedly, China's autonomous driving has embarked on a "high-speed road". Moreover, China's vast automobile market has also given Robotaxi more room for maneuvering.

According to the Ministry of Public Security, as of July 2024, the number of motor vehicles in China has reached 440 million, of which 345 million are cars. With the maturity of technology and the reduction of costs, Robotaxi is expected to gradually replace the traditional taxi market. By 2030, the domestic market size is expected to reach 488.8 billion yuan. At the same time, China's autonomous driving perception layer, including high-precision maps, LiDAR, and vehicle-mounted computing chips in the control layer, has made great progress in recent years. Diversified application scenarios, good road conditions, and rapidly developing communication technologies have provided fertile ground for the development of China's autonomous driving industry.

In addition, China's Robotaxi policy is top-down and unified nationwide, with unparalleled policy cohesion and execution capabilities compared to other countries. To date, the General Office of the State Council has initiated revisions to the Road Traffic Safety Law and the Regulations on Road Transport. More than ten cities across the country have promoted local legislation on autonomous driving, with legislative documents already issued in six places, including Shenzhen, Pudong, Yangquan, Wuxi, Suzhou, and Hangzhou. Wuhan, Beijing, Hefei, and Guangzhou are accelerating the legislative process.

According to public information statistics, 54 provinces, municipalities, and autonomous regions across the country have issued normative documents for open roads and demonstration applications. Among them, 28 cities support commercial exploration of autonomous driving, and 24 cities support driverless testing of autonomous driving.

In the most important area of safety, taking Baidu as an example, the intelligent driving system powered by a large model can achieve more than ten times the safety of human drivers, providing users with safe travel services. Baidu Apollo has purchased 5 million insurance policies for each driverless car and passenger. It is worth noting that in the past two years, the actual vehicle accident rate has been only 1/14 of that of human drivers. In this regard, as domestic technology matures, safety becomes controllable, policies tilt in favor of autonomous driving, and landing applications are gradually realized, Robotaxi, once considered a fanciful dream, may become an even more intense battlefield of competition between the two major powers across the Pacific Ocean.