Summary of Key Points
NVIDIA recently partnered with five major South Korean technology companies (SK Hynix, SK Telecom, Doosan, LG, and Naver) to focus on advancing two main areas: AI factories and physical AI. An AI factory is a smart production facility that utilizes electricity to generate "AI tokens"—units of information processed by NVIDIA. Physical AI enables devices such as robots and autonomous vehicles to perceive and perform tasks in the real world. This collaboration not only strengthens NVIDIA's transition from a chip manufacturer to a leader in AI infrastructure but also drives up the prices of A-share stocks related to physical AI in South Korea, indicating high market expectations for this field. By providing a full-stack solution (hardware + software + platform), NVIDIA is working with Korean companies to secure a position in the core sectors of the upcoming industrial revolution.
Detailed Explanation
1. What are AI factories and physical AI, explained in simple terms?
- AI factories: Jensen Huang described them as facilities that use electricity as raw material to produce tokens, which can be understood as units of information processed by AI systems. These factories utilize NVIDIA's technologies (such as GPUs and software platforms) to efficiently handle AI tasks like training and inference.
- Physical AI: This refers to the technology that allows machines to act autonomously in the real world, such as robots understanding their environment and performing actions. For example, robots need physical AI to recognize objects or avoid obstacles. NVIDIA's official website defines physical AI as "autonomous systems that can perceive, understand, and execute complex actions," integrating vision, motion prediction, and physical laws to handle uncertainties in the real world.
- NVIDIA has also released tools such as the DSX platform (the "blueprint" for building AI factories) and the Cosmos 3 model (the world's first multimodal physical AI model that can process various types of information).
2. How do the Korean companies collaborate?
The five companies each play a specific role in creating a complete ecosystem for AI factories:
- SK Hynix: Supplies high-performance memory (HBM) required for processing large amounts of data in AI applications, serving as a key hardware provider for NVIDIA.
- SK Telecom: Builds AI cloud infrastructure using the DSX platform, aiming to operate a gigawatt-level AI cloud by 2027 to provide training and inference services for Korean companies.
- Doosan Group: Develops intelligent robot systems and provides power solutions and electronic materials needed for data centers, as well as home robots.
- LG Group: Utilizes NVIDIA's digital twin technology to create autonomous manufacturing ecosystems and develops home robots like CLoiD, which are trained using NVIDIA's simulation tools in virtual environments.
- NAVER: Lays the foundation for AI infrastructure, helping Korean companies expand their AI capabilities.
3. Why did A-share stocks related to physical AI suddenly rise?
On the day of the announcement, companies such as Zhongwang Software and Fantuo Shuchuang saw significant price increases, with Suochen Technology rising by over 10%. The reason is simple:
- NVIDIA is a leader in the AI field, and its investment in physical AI indicates potential for this technology.
- These A-share companies are related to the technology: Zhongwang Software develops industrial design software related to digital twins; Fantuo Shuchuang focuses on digital twins; Tianyu Shuke is involved in AI simulation; and Dengke Technology works in industrial internet applications, all of which have connections to physical AI and AI factories.
- The market expects physical AI to be the next big trend, as Jensen Huang hinted that a breakthrough in this technology is imminent, potentially benefiting related companies.
4. What are NVIDIA's strategic goals?
NVIDIA's shift from being a chip seller to an AI infrastructure provider includes:
- Full-stack solutions: Offering a comprehensive set of technologies, including software (like the Isaac robot framework), platforms (such as DSX), and models (like Cosmos). For example, Doosan uses NVIDIA's Isaac framework for robot training, and LG uses its simulation tools. Once companies adopt NVIDIA's technology, they are less likely to switch to competitors.
- Ecosystem integration: Collaborating with Korean companies helps NVIDIA spread its technology across industries such as manufacturing and robotics. For instance, the humanoid robot platform being developed with Yutu will combine NVIDIA's software and computing power with Yutu's hardware to promote its ecosystem.
- Preparing for the industrial revolution: NVIDIA sees AI factories as the engine of the next industrial revolution, with physical AI at its core. By investing early, it aims to set the standards for this transformation.
5. Is physical AI still far from reality?
Physical AI applications are already in progress:
- Home robots: LG's CLoiD robot can perform household tasks and is trained using NVIDIA's simulation tools to ensure reliability in real-life use.
- Industrial robots: Doosan's robots use the Isaac Lab framework for complex industrial tasks like assembly and handling.
- Humanoid robots: Yutu's H2 Plus humanoid robot, equipped with the Jetson Thor computing module from NVIDIA, will be available by the end of the year, allowing researchers to develop various capabilities (such as walking and object manipulation).
These examples show that physical AI is not just a concept but is being implemented in practical applications, which may change our lives and work styles in the future.
Conclusion
The collaboration between NVIDIA and Korean companies represents a strategic battle for AI infrastructure. NVIDIA provides technology and an ecosystem, while Korean companies offer hardware, infrastructure, and use cases to advance the development of AI factories and physical AI. For investors, this suggests that fields like physical AI, robotics, and digital twins may become hot investment areas in the future. For the industry, this partnership could accelerate the industrial revolution, making machines more intelligent and integrated into our daily lives.