The Biden Administration seized the opportunity during its last days in office to introduce new regulations for artificial intelligence (AI) chip controls, categorizing China as a third-level control area with almost total prohibition.
This regulation raises questions about the potential impact on China's rapidly advancing high-end autonomous vehicle and robotaxi industries.
Supply chain experts point out that the development of China's robotaxi industry differs significantly from most countries due to strict government oversight over various networks, safety regulations, and firewalls.
From an external infrastructure perspective, insufficient advanced AI chips could hinder the computational power of data centers, potentially stalling the progress of autonomous vehicle development in China.
The Nvidia H20 chip is currently under scrutiny, as it remains uncertain whether it will be affected by the new regulations. This chip has become a focal point for China since it represents a special version provided by Nvidia for the Chinese market in response to US sanctions, replacing the banned A800 and H800 models.
The H20 serves as a seed for Nvidia's efforts in cultivating the autonomous driving ecosystem in China over the past six months, marking a significant breakthrough. Companies and automakers from both China and Taiwan are involved in this platform, allowing manufacturers to receive comprehensive support from infrastructure to application layers, thereby accelerating innovation in autonomous technology.
Regardless of whether Trump, who is set to take office on the 20th of January, will adhere to these bans, China has anticipated increasingly stringent US restrictions and has proactively prepared ample supplies of relevant chips. On the other hand, there is also a push to accelerate the development of domestic chips to avoid being constrained.
Thus, the extent of the impact of this ban on Nvidia's growth in the Chinese autonomous driving market becomes a key observation point.
China's importance to Nvidia
According to SemiAnalysis, Nvidia is expected to ship over one million new H20 chips to China in 2024, each costing between US$12,000 and US$13,000, generating approximately US$12 billion in revenue.
As for the vehicles themselves, given that China's connected vehicle network and related infrastructure for autonomous driving are relatively well-established, coupled with the government's commitment to maintaining cybersecurity against international hackers and terrorist threats, they have built a more robust protective barrier compared to Europe and the US.
Therefore, within this strongly protected environment, the vehicles may not require excessively high computing power. Instead, they can achieve their autonomous driving goals through features like vehicle networking.
Industry players indicate that beyond semiconductor chips, factors such as practical experience feedback, optimized data generation, and improved software continuously enhance the entire high-end autonomous vehicle operating system—areas where China holds a competitive advantage.
After all, China possesses the world's largest automotive market for real-world testing, and local brands dominate supply chains, enabling the Chinese government to make significant strides in autonomous driving development. The aim is to lead specifications while the market recognizes its developmental potential.
Moreover, autonomous vehicles represent one of the applications of AI robotics.
In reality, compared to humanoid robots or collaborative robots, autonomous vehicles present higher barriers due to the diverse and complex scenarios in which they operate, compounded by the element of "speed" and the need to ensure the safety of drivers and passengers.
As we move towards Level 3 advanced automation, effective vehicle networking must address concerns regarding hacking, which could turn vehicles into weapons for terror attacks, further jeopardizing societal and national security.
Article translated by Joseph Chen
