The world's largest electronic consumer exhibition CES has opened, with Nvidia unveiling a new consumer grade chip for personal computers, achieving multiple AI business collaborations, and announcing the acquisition of several key Chinese automotive customers, driving investor optimism. At the close of the day, Nvidia's stock price surged 6.4% to a historic high of $522.53 per share.
In terms of hardware, Nvidia has released three RTX 40 SUPER series graphics card products for consumers, all based on TSMC's 4nm process technology and adopting the new Ada Lovelace architecture, which is a routine upgrade. However, Nvidia claims that the new products will help users prepare for the upcoming wave of generative AI applications on PCs. In addition, laptop manufacturers will also launch AI laptops based on Nvidia GPURTX to welcome the wave of AIPC.
The cooperation with multiple Chinese companies in the fields of gaming and autonomous driving is a highlight of Nvidia at CES. Nvidia announced at CES that Chinese game manufacturers such as MiHoYo, NetEase Games, ZhangQu Technology, and Tencent Games will cooperate with them in the field of AI technology applications and apply them in game development. According to Zhangqu Technology, Nvidia will provide AI technology support such as DLSS and Audio2Face to its AI game creation platform.
Nvidia's multiple AI technologies are expected to be applied in the game development process, which will change the form of game interaction. For example, NVIDIA's "ACE" (Avatar Cloud Engine) microservice, which creates virtual digital characters through generative AI models, bringing a realistic gaming interactive experience. In terms of specific implementation, in player games, the automatic speech recognition model will convert the player's voice into text, which will be imported into the AI model to form a real interactive response of the game characters. The dialogue of the game characters will also be synchronized with the real mouth shape through AI technology.
In the emerging autonomous driving market, Nvidia has also won several key Chinese automotive customers. On January 9th, at the same time as the release event before the CES exhibition, Nvidia announced that Ideal Automobile will use the Thor automotive chip platform in its next generation models. In addition, three manufacturers including Great Wall Motors, Jike, and Xiaomi have adopted Orin chips for the new generation of intelligent driving systems.
This series of collaborations is seen by Nvidia as a major breakthrough in its autonomous driving business. According to Wu Xinzhou, Vice President of NVIDIA's Automotive Division, Orin chips have become the preferred AI computing platform for smart cars today. As car manufacturers demand more advanced autonomous driving functions and AI performance, the next generation Thor chips will continue to be adopted by future car models.
According to Nvidia data, the highest computing power of the Orin chip is 254TOPS. In 2022, Nvidia released the next generation autonomous driving chip Thor, with a computing power of up to 2000 TOPS (1 TOPS is equivalent to 1 trillion calculations per second). It is expected to be mass-produced in 2025, making it the highest computing power intelligent driving chip at present.
Market research firm Counterpoint believes that achieving L2 level assisted driving requires 10 to 100 TOPS of chip computing power, L3 level (partially scenario autonomous driving) requires 150 to 200 TOPS, L4 level (mainly autonomous driving) requires 400 TOPS or more of computing power, and L5 level fully autonomous driving requires 1000 TOPS or more of computing power.
The automotive business is Nvidia's fourth largest sector independent of data centers, gaming, and professional visualization. Nvidia maintains a leading position in the field of autonomous driving, but its competitors are still catching up. According to data from the Intelligent Automotive Research Institute of Gaogong, in the first half of 2023, Nvidia accounted for 52.57%, almost half of the vehicles equipped with advanced driving assistance (NOA) functions.
At present, most high-end car models in China basically choose NVIDIA Orin chips for assisted driving, and some car manufacturers such as NIO have already equipped four Orin chips as standard on all lines. In terms of competitors, Mobileye, which was spun off by Intel, has partnered with Geely to enter the Geely brand of car models. The computing power of Horizon products, a domestic startup, is not as strong as Nvidia's. However, relying on close service and price advantages, it has entered the supply chains of Ideal Automobile, BYD, and others. Qualcomm is also trying to make efforts in the field of autonomous driving by launching related chip products.
Market news is still stirring investor sentiment. On January 9th, it was reported that Nvidia plans to mass produce the Chinese designed AI chip H20 in the second quarter to comply with US export regulations. The report stated that H20 was originally planned to be launched in November last year, but server manufacturers encountered difficulties in integrating the chip, and H20 will prioritize meeting the needs of major Chinese customers.