omniture

DIGITIMES Asia: Tesla and Nvidia set to go head-to-head in high-end autonomous driving

2024-07-31 15:30 1454

Nuying Huang, Taipei; Jack Wu, DIGITIMES Asia Wednesday 31 July 2024

TAIPEI, July 31, 2024 /PRNewswire/ -- In the realm of end-to-end advanced autonomous driving solutions, both Nvidia and Tesla are dominant forces.

Credit: AFP
Credit: AFP

Their continued development in this area means their paths are bound to cross. However, they will likely face a common challenge: the variables in the Chinese market.

According to the news report from the technology-focused media DIGITIMES Asia, supply chain sources noted that Nvidia collaborates with numerous Tier 1 suppliers and automakers, continuously adjusting and enhancing its platform based on customer feedback and demands. Unlike Tesla, which focuses primarily on cameras, Nvidia's platform is compatible with various sensory components and systems. Nvidia also has ample computational resources at its disposal, ready to capitalize on the autonomous driving boom whenever it occurs.

In L2 to L3 autonomy, Nvidia held approximately 50% market share in 2023 for advanced driver assistance systems (ADAS)/Navigation on Autopilot (NOA) in China's passenger car market. Horizon Robotics had about 35%, Huawei HiSilicon around 8%, and Texas Instruments (TI) about 4%.

Market estimations suggest that Nvidia has a substantial database due to its existing platforms having accumulated data for a while. Tesla, meanwhile, is a major automaker itself.

While Tesla couldn't find a licensing partner for its Full Self-Driving (FSD) technology, it has been regularly collecting and optimizing data for some time. Nvidia, on the other hand, may need to rely on feedback from various customers.

Both end-to-end and generative AI require high-end GPU computing power support, making Nvidia's GPUs highly sought after and constantly in short supply. Tesla CEO Elon Musk anticipates that Nvidia GPU prices will continue to rise, repeatedly testing the limits of market acceptance.

In response, Tesla plans to address this issue with its in-house AI chips, the HW5 or AI 5, expected to enter mass production in 2025. These chips are comparable to Nvidia's B200 in inference capabilities, indicating a direct competition with Nvidia. Additionally, Tesla's AI is not limited to autonomous vehicles but extends to its upcoming humanoid robots, meaning that AI inference will extend from vehicles to homes and factories.

Thus, Tesla and Nvidia are set for a showdown in high-end autonomous driving. In particular, MediaTek's collaboration with Nvidia to enhance smart cockpits and Qualcomm's expansion from smart cockpits to high-end autonomous driving highlight the intense competition between MediaTek and Qualcomm, a rivalry that has gone from smartphones to future vehicles.

Variables in the Chinese market

However, geopolitical factors could pose challenges for these companies in the Chinese market. As China has successfully advanced in the automotive sector, the development of smart driving vehicles is booming, with major cities having already launched Robotaxi demonstrations. This indicates China's development vision requires effectively controlling all software, hardware, and data.

This is also why the domestic brand Huawei is so highly regarded. Its comprehensive capabilities, from semiconductor manufacturing to smartphones, vehicles, and smart homes, allow it to generate and connect data across different sectors.

In 2024, the Chinese car market is expected to be driven by smart driving. As Tier 1, Huawei holds about 30% market share in smart driving solutions. Other significant players, Momenta and DJI Automotive, each hold about 20%. Most Chinese automakers with notable market presence are currently evaluating potential partners.

New Chinese automakers such as NIO, Li Auto, and Xpeng, also have strong platform-building capabilities, although their production and sales scales still need to grow. Additionally, Xiaomi's EVs are selling well, indicating the competitive landscape is continuously evolving.

Source: DIGITIMES ASIA
collection