This summary is sourced from the internet. If the organization believes it is inappropriate to disseminate, please leave a message in the backend of the 【Wenbagu Research Summary】 public account to delete the summary! We apologize for any inconvenience caused!
NVIDIA Jetson Thor will be launched tomorrow, with a focus on the direction of robotic brains.
On August 25, NVIDIA will release its “new robotic brain” product – the NVIDIA Jetson Thor platform.
The NVIDIA Jetson Thor platform is designed for physical AI and humanoid robots, supporting all mainstream AI frameworks and generative AI models.
The computing power of Jetson Thor is 7.5 to 10 times that of the previous generation Jetson Orin; performance per watt has improved by 3.5 times; and I/O throughput has increased by approximately 10 times, meeting high bandwidth perception requirements.
#Robotic computing power is the inflationary link in the robotics field.
As robots move towards practical applications, the complexity of robot data (visual, tactile, joint states, etc.), the complexity of application scenarios (home, industrial, special, etc.), and the real-time action generation will continuously increase the demand for data and computing power. The current computing power of the Orin chip is already under significant strain, and the Thor chip will alleviate the computing power bottleneck to some extent, but future demands will continue to rise.
With the catalysis of NVIDIA Jetson Thor, #we reiterate our optimism about the direction of robotic brains and suggest paying attention to:
【Tianzhun Technology】 has been a gold partner of the NVIDIA Jetson platform since 2020, and the two parties have a binding cooperation in chip + complete machine + scene implementation.
Tianzhun’s “Xingzhi 001” intelligent controller may be used in products like Zhiyuan’s “Lingxi X1”.
Based on the Thor platform, the solution price is expected to double to 35,000 to 40,000 yuan, with a significant value proportion;
【Zhiwei Intelligent】 is one of NVIDIA Jetson’s partners, providing controllers specifically for humanoid robots;
【Horizon Robotics】’s subsidiary, Digua Robotics, is a provider of general-purpose robot bases, with a complete product system from chips, algorithms to software;
【Mingyang Circuit】 provides board design and manufacturing based on the NVIDIA Jetson series, offering customized solutions for robot customers, with a value of thousands of yuan;
【Aobi Zhongguang】【Suteng Juchuang】 are partners in the sensor segment of NVIDIA Jetson Thor.
NVIDIA provides the brain for robots.
Full Text Summary
NVIDIA plays a key role in the humanoid robot industry chain, showcasing nine robots based on its AI technology at the 2024 GTC conference, utilizing diverse training data from both domestic and international sources. The AI platform provided by NVIDIA includes GPU computing power, asset lab simulation platforms, and large model training platforms, which are crucial for the advancement of the global robotics industry. A spokesperson emphasized that real-world data is indispensable for AI model iteration, and NVIDIA dominates the computing power field, discussing potential alternatives. Based on an analysis of AI and humanoid robot industries, investment recommendations were made, highlighting the core position of computing power and AI platforms in investments, and encouraging investors to engage deeply with relevant teams.
Chapter Overview
● 00:00 NVIDIA’s Core Position in the Humanoid Robot Industry ChainThe dialogue introduced the nine humanoid robots announced by NVIDIA at the 2024 GTC conference, and its core advantages in computing power, simulation platforms, and large model training. NVIDIA’s GPU computing power, asset lab simulation platform, and growth model training platform together constitute its important position in the humanoid robot industry chain. Currently, NVIDIA’s training programs cover most mainstream robots globally, significantly contributing to the advancement of global robotics technology.
● 03:27 Large Model Iteration and Computing Power Demand AnalysisThis section discusses the challenges of large model iteration in the context of scarce real-world data, emphasizing the importance of NVIDIA’s simulation platform and computing power, as well as the potential demand for alternative computing power chips from domestic companies.
● 07:07 NVIDIA and Tesla’s Roles in the Intelligent Industry Chain and Investment Outlook for Humanoid RobotsThe dialogue explores the different positions of NVIDIA and Tesla in the intelligent industry chain, pointing out that NVIDIA focuses on software solutions while Tesla emphasizes execution and integration technology. Based on this, it analyzes the future investment trends of humanoid robots as AGI products, highlighting the sequence of transition from computing power development to end applications, drawing parallels with the development process of the 4G industry, and predicting the investment transmission logic in the humanoid robot market.
● 10:59 Humanoid Robot Industry and Computing Power Demand AnalysisThe dialogue delves into the development stage of the humanoid robot industry within the global AI industry, indicating that current computing power demand is a key factor driving industry growth. The strong performance of computing power suppliers like NVIDIA suggests the potential of future robot manufacturers, with investment recommendations focusing on the computing power end and the Tesla industry chain, emphasizing the importance of the supply chain.
Key Points Review
What role does NVIDIA play in the humanoid robot industry chain? What specific contributions does NVIDIA make in the industry chain?
NVIDIA plays a critical role in the humanoid robot industry chain, showcasing its integration of extensive domestic and international data and information from humanoid robot manufacturers through the nine robots announced at its GTC conference. Currently, most mainstream robot products globally may utilize NVIDIA’s brand or the computing power it provides, which is significant for the development of the entire industry chain. NVIDIA addresses challenges in the intelligent industry chain, such as AI model development, system integration, and high-performance computing needs, through its AI machine vision, cloud operation platform, and high-performance computing. Although there is currently no clear information indicating NVIDIA will directly participate in manufacturing, its overall position in the industry chain is crucial for promoting the development of the intelligent industry.
What components make up NVIDIA’s integrated platform?
NVIDIA’s integrated platform mainly consists of three aspects: first, the latest GPU computing power, although the name has undergone iterations; second, a simulation platform called asset lab, which allows robot companies to pre-train in a simulated environment, saving costs associated with acquiring real-world data, and the training results on this simulation platform are relatively ideal; finally, the growth model training platform based on large models. Together, these three components form NVIDIA’s relatively complete integrated humanoid robot plan.
What is the importance of simulation environments for humanoid robot manufacturers lacking large-scale real-world data?
For most manufacturers whose production volume has not yet reached thousands of units, NVIDIA’s simulation platform is particularly important as it allows them to utilize real data from leading companies for model iteration and training in a simulated environment, thus compensating for the lack of real-world data.
What is NVIDIA’s position in terms of computing power?
In terms of computing power, almost all iterations of humanoid robots globally rely on NVIDIA’s GPU computing support. Some companies use NVIDIA’s platform for iteration, while others, focusing on their own model updates and real data collection, do not depend on NVIDIA’s platform but still require NVIDIA’s powerful computing power for updates. Therefore, NVIDIA has formed a relative monopoly in the computing power sector, but there is also the possibility of domestic companies replacing its computing power chips, which has become a focus for future investments.
What is the core point that Tesla addresses in humanoid robots?
The core point that Tesla addresses is primarily based on its machine FSD model, where execution and integration technology are key components. If Tesla’s robot intelligence exceeds expectations, it will impact the analysis of the entire investment situation.
What is the positioning of humanoid robots in the era of artificial intelligence? What is the current development logic of the humanoid robot industry?Humanoid robots are considered fundamental AGI products in the era of artificial intelligence and may also be core applications of artificial intelligence. Their development process is similar to that of smartphones in the 4G industry, including the development of foundational computing power chips, commercialization of terminal carriers, and application development. Currently, the development of humanoid robots resembles the computing power development stage in the early phase of the 4G industry, starting from computing power and less mature applications, gradually transitioning to humanoid bodies, and ultimately achieving widespread application in real-world scenarios. In this process, computing power will develop first, subsequently driving small-scale applications and iterative upgrades of intelligent robot brains.
What is the current state of development in the AI industry?
Currently, in the AI industry, the stock prices of computing power suppliers like NVIDIA have risen significantly, closely related to the enormous demand for computing power brought by intelligent and supporting enterprises. In the robotics industry, although it may currently be in a similar first-stage development state, the significant increase in robot shipments in the future is expected to greatly enhance the valuation of manufacturers.
What are the investment recommendations for the current AI industry?
It is recommended to pay attention to companies in the computing power sector mapped to NVIDIA in China, as they can generate actual revenue. For companies with good iteration, consider investing in their bodies and supply chains. Additionally, suppliers within the Tesla industry chain are also important investment directions. If NVIDIA has further initiatives in the robotics field or plans to produce humanoid robots, its supply chain will become crucial.
Follow our public account for updates on the latest institutional research.
Click “Read the original text” to jump to the 【Wenbagu Research】 PC version.