Competition for the acquisition of Graphics Processing Units (GPUs) in the global automotive industry is intensifying.

Major automotive OEMs, led by Tesla, are stepping up technology development in the software area (mainly AI, self-driving, connected, etc.) and are investing heavily in building service platforms targeting B2C, as they seek to realize SDV, a hot topic these days. In addition to the existing in-vehicle software field, the Company has also been investing in the development of a service platform targeting B2C.
In addition to the existing in-vehicle software field, they are increasing investment in building infrastructure such as data centers and cloud computing to support this remotely. In addition to third-party collaboration, North America is deploying its own data centers for training AI through data collection and other purposes, In addition to third-party collaboration, the deployment of homegrown data centers for the purpose of training AI through data collection is gaining momentum, particularly in North America and China.

Tesla, which is considered the most advanced in this field, is said to have built a data center called Cortex in Gigafactory Texas in August 2024, where 50,000 NVIDIA H100 GPUs (water-cooled cluster) have been installed. The company plans to increase the number of GPUs to 85,000 by the end of 2024, add next-generation NVIDIA H200 GPUs, and increase the number of GPUs to 100,000, with plans to install its own semiconductors such as Dojo chips in the future.
Tesla is investing more than $1 billion a year to develop and train AI for its Full Self-Driving (FSD) technology and its Optimus humanoid robot. Tesla is investing in these areas because they are key to its dominance in the race for future leadership.
Meta (Facebook) already has 25,000 H100s in place for its own AI training, and Google, Microsoft, and others are reportedly in the process of securing GPUs of similar scale.

Among the key elements of SDV, the related industries, including the automotive industry, are now investing more and more in order to properly provide various services such as automated driving, user assistance by AI, and other personalization services.
SDV can only be realized when the right infrastructure and software support and backup systems are in place. Competition for high-end semiconductors is expected to intensify.

　Now, FOURIN plans to publish “Global Automakers’ SDV Strategies and Competition for Semiconductors” in April 2025.

　In this research report, we will investigate and analyze the global competition in the era of SDV, including policies and regulations in major countries around the world, as well as strategies and trends among automakers, component suppliers, and related stakeholders. The report also examines the status of software-related initiatives as well as hardware, and analyzes the expected competition for survival among industries as the shift to SDV progresses.