Equal rights in intelligent driving, Bosch proposes infrastructure "open strategy"

Author | Chai Xuchen

Editor | Zhou Zhiyu

"Five years from now, the full-stack self-developed intelligent driving systems that car companies today spend huge sums on and take pride in will become as ordinary as airbags." This is Bosch's ultimate judgment on the "involution" battle in the Chinese automotive industry.

On July 26, at the World Artificial Intelligence Conference (WAIC), when Eddie Wu, President of Bosch Intelligent Driving Control in China, took the stage, the entire industry was focused on how this century-old Tier 1 giant would prescribe new remedies for the extremely competitive Chinese car market.

His speech outlined a clear and grand path: currently, using leading intelligent driving capabilities to help car companies quickly address their shortcomings and alleviate anxiety, navigating through the intense competition cycle to keep pace with industry races; in the long term, committing to becoming a "infrastructure" supplier for the era of intelligent vehicles, laying the foundation for the future of the industry.

Bosch's "sunshine strategy" is gradually coming to light; it aims to become the Nvidia or Qualcomm of the intelligent automotive era, playing an indispensable core role at the foundational level. This is the starting point of its future strategy and a key battle to break the deadlock of price wars in the automotive circle.

Transforming into Intelligent Driving "New Infrastructure"

"In the future, as technology converges, intelligent driving will inevitably become standard equipment, just like today's seat belts and airbags."

As the competition for "intelligent driving equality" in the automotive circle heats up, Eddie Wu threw out radical remarks at the July 26 WAIC forum—intelligent driving will eventually become a standard feature, and car companies that lag behind should invest their energy and resources in the intelligent cockpit track, which can provide emotional value to users, rather than obsessing over self-development.

In Eddie Wu's view, car companies need not feel anxious because their intelligent driving systems have not kept pace with the situation. His thinking is clear; he believes that the country has already introduced many strict standardization measures for the intelligent driving industry, and even if manufacturers invest a lot of manpower and resources in standardization, it will not bring differentiated competitive advantages in product sales.

"Intelligent driving will not provide emotional value in the future; getting from point A to point B safely, comfortably, and on time will not become a highlight of differentiation for manufacturers selling cars." Eddie Wu even asserted, "In the future, manufacturers may abandon full-stack self-development in four to five years."

To be fair, car companies are already well aware of the importance of intelligent driving. From Audi, Mercedes-Benz to BMW, and even more and more overseas automotive giants like Toyota and Volkswagen are marrying intelligent driving suppliers in the Chinese market to quickly address their shortcomings.

This also means that competition in intelligent vehicles is developing towards "ecological integration."

Eddie Wu stated that in the long term, whether for mid-to-high-level intelligent driving, it will gradually become standard equipment for vehicles. In this foreseeable future, those car companies still investing huge manpower and resources in the "algorithm black hole" of intelligent driving will face significant strategic risks.

Industry insiders pointed out to Wall Street Journal that manufacturers need to invest a team of one to two thousand people and hundreds of millions of funds annually for self-developed intelligent driving, but they can only adapt it to a dozen of their own models, while suppliers can cover dozens of models with similar resources; More importantly, it is a matter of time; it takes three years for automakers to develop a full-stack self-research process to connect various workflows.

Thus, Wu Yongqiao prescribed the right remedy, throwing an enticing proposal to car manufacturers who are deeply mired in competitive pressure and investment risks—Bosch would provide a high-starting point, rapidly mass-producible mature solution. Car manufacturers would marry this solution to quickly fill the gaps in intelligent driving, directing valuable R&D resources toward areas that can truly create differentiated value, thereby breaking through in this "involution" battle.

Clearly, this automotive supplier giant wants to become an "infrastructure builder" akin to "water, electricity, and coal" in the industry, thereby supporting car manufacturers in the Chinese market.

In fact, Wu Yongqiao's idea is gradually becoming a consensus among some suppliers.

At the China Electric Vehicle 100 Forum in March this year, Horizon founder Yu Kai regarded the intelligent driving high ground fiercely contested by car manufacturers as a basic function that will be standard in future vehicles, "not providing emotional value," and will not define brands.

He bluntly stated that this seems to be a counter-consensus thought in the industry, but if we look back at the past 30 years of mobile phones, mobile phone manufacturers have gone from self-researching communication basebands to the emergence of Texas Instruments and Qualcomm, resulting in an 80% reliance on suppliers.

"When functional value leaps forward with technological advancements, manufacturers either self-develop or use suppliers, but the unchanging core is to outpace the slow and outperform the low. Those who cannot keep up with the speed, like Nokia and Motorola, will be overturned by the times," Yu Kai said.

Perhaps, new variables in the intelligent driving industry are quietly emerging. With the push from Bosch and others, cars are stepping into the AI robot era, and the automotive industry will welcome its most exciting moment of transformation.

However, what courage does Bosch have to become a disruptor in the automotive circle? It is important to note that leading players like Tesla, Huawei, and XPeng have been able to establish themselves in the intelligent driving arena due to early predictions and continuous efforts and accumulation.

At this moment, Wu Yongqiao played his trump card.

The Confidence of Tier Giants

"In the era of rules, it is difficult for us to catch up with players like Huawei; they have spent over a decade and hundreds of billions accumulating over 100,000 corner cases," Wu Yongqiao stated. However, he quickly shifted his tone, "Now in the end-to-end era, with money, computing power, good algorithms, and high-quality data, continuous training will definitely catch up with leading players."

He pointed out that Li Auto is a representative of overtaking on a curve. After launching AI, the gap between the industry and Huawei has shrunk from a year to three or four months. Wu Yongqiao confidently stated that after launching a one-stage end-to-end solution in the second half of 2025, Bosch has the opportunity to achieve a leap in experience.

However, on the road to the pinnacle, Bosch is not reckless. As an industry infrastructure builder, it regards delivery as its top priority.

Currently, VLA (Vision-Language-Action) and world models are gaining fame, seen by many as the ultimate solution to general artificial intelligence. But from the perspective of a seasoned engineering professional, Wu Yongqiao pointed out the core obstacles to its short-term implementation.

"Many people are discussing world models, VLA, and one-stage end-to-end solutions; explaining these terms is very easy. However, how to engineer and deliver complex technologies is extremely challenging," Wu Yongqiao stated. "Firstly, the multi-modal feature alignment of VLA is very difficult; "The acquisition and training of multimodal data is very difficult."

He further pointed out the most fatal engineering bottleneck: "If the VLA model needs to be deployed on a chip, achieving true driving safety and a highly humanized driving experience must reach a scale of 7B or 10B parameters. But currently, all intelligent driving chips are not designed for large models."

He explained that this mismatch of computing power and bandwidth leads to the large model running on existing chips, where "it is very difficult to achieve a frequency of 10 Hz," which means the model "cannot control the vehicle in real time."

This is a cold engineering reality: no matter how advanced the algorithm, if it cannot meet the real-time high-response vehicle control requirements, it can only remain on a PowerPoint presentation. Wu Yongqiao judged that the true implementation of the VLA large model may need to wait "3 to 5 years" for the emergence of a new generation of chips.

Wu Yongqiao's remarks articulated a common pain point in the industry; he even described the difficulty of implementing advanced intelligent driving as "the most complex systematic engineering in automotive history, without exception." Countless ambitious projects ultimately falter at the threshold of large-scale production and engineering delivery. The complexity and agony of this process resonate deeply with him as the project leader.

He pointed out bluntly: "Therefore, engineering delivery and implementation are the core factors determining whether intelligent driving can move towards the future, rather than the technology itself." Faced with the "distant water" of future technology, Bosch chose a pragmatic route that could quench the "near thirst."

"We have always firmly pursued a one-stage end-to-end approach," Wu Yongqiao stated. This solution integrates multiple links such as perception and regulation into a single model to achieve more human-like driving decisions.

Its "one-stage end-to-end" solution is not just a set of algorithms but a complete "turnkey" project that includes hardware, software, testing, validation, and production. This capability makes it not just a technology supplier but a reliable "infrastructure" builder.

Bosch is not just talking the talk. Wu Yongqiao announced at the conference that Bosch is collaborating with local autonomous driving company WeRide to develop a "one-stage end-to-end" intelligent driving solution based on NVIDIA's Orin-X chip, which will be implemented in high-end models of the Xingtu brand by the end of August this year.

It can be said that this is the core value of Bosch as a top global Tier 1 supplier. While new forces and technology companies are still struggling with the "pain of delivery," Bosch has already internalized the capability for large-scale, high-quality engineering delivery into the DNA of the enterprise.

Industry "Breaker"

At this point, Bosch's "sunshine strategy" has been revealed; its ambition is to persuade most car companies with "intelligent driving anxiety" to use its solutions to navigate through the intense competition and keep pace with the industry's race.

In fact, Bosch, which aims to be the "savior" of OEMs, is also trying to push the entire industry to escape the narrative of internal competition and return to a growth trajectory.

In his speech at WAIC, Wu Yongqiao candidly stated that the current Chinese automotive industry is facing a severe paradox.

From January to May this year, the revenue of the Chinese automotive industry grew by 7%, but profits astonishingly fell by 11.9% year-on-year. This phenomenon of "increased revenue without increased profit" is the most intuitive manifestation of the industry's "internal competition." Wu Yongqiao admitted that in this brutal competition, with very few exceptions, the vast majority of suppliers, including Bosch, are facing tremendous price pressure.

"Originally, there should have been hundreds of billions in profits in the Chinese automotive circle, but after the automakers conveyed a 'chill' to the supply chain, those hundreds of billions vanished into thin air." In Wu Yongqiao's view, the price war in the automotive sector has caused irreversible damage to the entire supply chain. "You will find that more and more supply chain suppliers will go bankrupt and shut down. I believe that the price war in 2025 will definitely cool down and will not continue in such a disorderly manner."

Therefore, after taking the helm of Bosch's intelligent driving control business last year, Wu Yongqiao began to make decisive decisions, with the key being to inject "wolf culture" into Bosch.

"If everyone doesn't compete, there will be no opportunities. Bosch is the only manufacturer left at the table doing high-level intelligent driving; companies like ZF, Aptiv, and Denso have all exited high-level intelligent driving. If we don't compete and don't engage, there will be no opportunities, and it will be impossible to have a chance to compete with the giants. I require the ADAS team to further promote the 'wolf culture' to better adapt to the survival mode of intelligent driving in China."

Bosch's ambitions, clearly, go beyond merely becoming a "standard component" supplier for intelligent driving. In the blueprint drawn by Wu Yongqiao, as intelligent driving gradually becomes homogenized, the main battlefield of automotive competition will undergo a dramatic shift.

He believes that when the era of "0 takeover in 100 kilometers" of intelligent driving arrives, "the cockpit will usher in a moment of flourishing and brilliance." Unlike the functional attributes of intelligent driving, "the cockpit can bring emotional value to people." Wu Yongqiao criticized the current cockpit experience as "highly homogenized," merely staying at a superficial level of interaction such as voice, navigation, and music, which is passive and cannot reflect true AI capabilities.

Bosch is exploring with OEMs an AI intelligent cockpit with computing power of up to 300 TOPS. He described a scenario: when you enter the car, the system perceives through a full-body scan that you are "in a bad mood or heartbroken, and the car will immediately play a song... to soothe your body and mind"; and when you "make hundreds of thousands in stocks, the car system will immediately play 'Borrowing Five Hundred Years from Heaven Again,' full of ambition and passion."

This is the true intelligent cockpit—one that "becomes your soul mate through perception and big data accumulation." Such a cockpit that can provide deep emotional value will become the most core differentiating selling point for automakers in the future.

This also leads to Bosch's thinking about the ultimate form of automotive intelligence. From intelligent driving to the cockpit, it will ultimately converge into an integrated cabin and driving experience, eventually integrating into a powerful central computing platform.

In the future, this "brain" may have computing power of up to 1000 or even 2000 TOPS, integrating the strategies of all controllers in the vehicle, "truly achieving unified coordination of the entire vehicle layout, and genuinely becoming the soul of the vehicle driven by a central brain."

The "one-stop end-to-end" intelligent driving capability and AI intelligent cockpit capability that Bosch is building are the core AI foundations that constitute this "central brain." From providing pragmatic "one-stop end-to-end" intelligent driving solutions to help car companies "break the deadlock," to laying out AI cockpits that provide emotional value, and finally building the vehicle's "central brain," Bosch's strategy progresses layer by layer, clearly demonstrating how it leverages its deep engineering capabilities and forward-looking AI technology to evolve from a system supplier to a "new infrastructure" role in future smart vehicles.

A new variable in the automotive industry's intelligence is quietly emerging. With Bosch's efforts, the automotive industry is set to welcome the most exciting moment of transformation.

Risk Warning and Disclaimer

The market has risks, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article align with their specific circumstances. Investment based on this is at one's own risk