Caitong Securities: NVIDIA Increases Investment in Robotics Track, Focuses on Domestic Companies That May Benefit from Breakthroughs in Embodied Intelligence Progress
NVIDIA has recently increased its investment in the embodied intelligence field by establishing the GEAR Embodied Intelligence Laboratory and launching the GR00T humanoid robot universal model. Caitong Securities recommends focusing on high-quality domestic companies that may benefit, such as Tianzhun Technology, Boke Technology, Eastern Precision, and Jiuhao Company. NVIDIA's layout in the intelligent robot field mainly revolves around large models, data, and development platforms, aiming to build the underlying development ecosystem for robots. In 2023, "Generative AI + Robots" will become the core direction of NVIDIA's robot technology. NVIDIA has established the GEAR Laboratory to lead the wave of embodied intelligence
According to the Zhitong Finance and Economics APP, Caitong Securities released a research report pointing out that embodied intelligence is the next wave of artificial intelligence. As the global leader in AI computing power, NVIDIA recently established the GEAR Embodied Intelligence Laboratory and unveiled the GR00T humanoid robot universal model at the GTC 2024 conference. By continuously increasing investment in intelligent robots, NVIDIA will lead advanced AI technology to deeply empower humanoid robots. It is recommended to pay attention to high-quality companies in China that may benefit from NVIDIA's progress in embodied intelligence, including Tianzhun Technology (688003.SH), Buoke Technology (688160.SH), Eastern Precision (002611.SZ), and Jiuhao Company (689009.SH).
Key points from Caitong Securities:
NVIDIA is the global leader in AI computing power, strategically positioning in the intelligent robot field
NVIDIA's layout in the intelligent robot field mainly revolves around three core areas: large models/data/development platforms, with the ultimate goal of building the underlying development ecosystem for robots. Starting in 2018, NVIDIA began its layout in the intelligent robot field, from chips at the bottom layer to computing platforms, training, and validation from 2019 to 2022, initiating a comprehensive layout in the intelligent robot field. In 2023, "Generative AI + Robots" became the core direction of NVIDIA's full-line product technology evolution and reshaping in the robot field.
NVIDIA officially establishes the GEAR Laboratory, leading the wave of embodied intelligence
NVIDIA continues to deepen its layout in the field of embodied intelligence and actively invests in external companies such as FigureAI, a manufacturer of embodied intelligent robots. The progress of AI + robots is smooth. In February 2024, NVIDIA established the General Embodied Intelligence Research (GEAR) team to promote the development of large models and other AI technologies from the virtual world to the real world. The team's research will mainly focus on four key areas: multimodal basic models/universal robot research/basic intelligent bodies in the virtual world/simulation and synthetic data technology.
Launch of the GR00T large model at GTC 2024, building the underlying development ecosystem for robots
At the GTC 2024 conference, NVIDIA CEO Huang Renxun launched the Project GR00T universal basic model for humanoid robots and released the accompanying new humanoid robot computer Jetson Thor, as well as made significant updates to the Isaac Lab. The GR00T project aims to further promote NVIDIA's breakthroughs in robots and embodied intelligence. The design of the Jetson Thor platform is based on the NVIDIAThor system-level chip and the latest Blackwell GPU architecture, providing 8 trillion 8-bit floating-point operation AI performance per second. In addition, Boston Dynamics announced a strong partnership with NVIDIA to accelerate the construction of the underlying development ecosystem for robots Risk Warning: Tesla humanoid robot mass production falls short of expectations, robot generalization capability upgrade falls short of expectations, domestic manufacturers' product R&D progress falls short of expectations, etc