"US Profiting from Korean Data"
Major Corporations Move to Reduce Dependence on Specific Platforms

Concerns about dependence on Nvidia are not limited to semiconductors. As physical AI becomes integrated into the entire Korean manufacturing industry—including sectors such as automobiles and robotics—a structure is rapidly spreading in which core software powering factories and robots relies on American big tech companies. If the current system, where Korea provides data as a raw material and the United States captures the added value, becomes entrenched, there are warnings that Korean manufacturing could devolve into a technological colony.


According to the industry on July 6, outside of semiconductors, the most active field for utilizing Nvidia's digital twin technology in Korea is physical AI. Boston Dynamics, a subsidiary of Hyundai Motor Group, is using Nvidia's robot simulator Isaac Sim to train its humanoid robot Atlas in a virtual environment. LG Electronics is also incorporating Nvidia solutions throughout the entire cycle of robotics advancement—including the training, simulation, and chipset of its home robot CLOi. Among domestic robot startups, there is also a trend to participate in accelerator programs that allow them to use Nvidia's robotics platform, aiming to shorten product development periods.

[The Two Faces of Digital Twins]② Korea Pays for the Materials, America Reaps the Value... 'Technological Colony' Warning Lights for K-Manufacturing View original image

The issue is that all these processes run on Nvidia infrastructure. Digital twins in automotive and robotics manufacturing environments must calculate complex physical laws—such as airflow, friction, and torque—in real time, making high-performance computing essential, which in turn leads to reliance on Nvidia graphics processing units (GPUs). If global hegemonic competition or supply chain disruptions cause problems in securing GPUs, the advancement of physical AI by domestic companies will be hampered. Ultimately, this means that control over the pace of innovation in Korean manufacturing could lie in the hands of foreign companies.


Choi Byungho, a professor at the Korea University AI Research Institute, warned, "No one can deny the risk of domestic companies being locked into the Nvidia platform. In a situation where simulation through digital twins inevitably relies on Nvidia models, if access to the platform is ever restricted, the entire industry could come to a halt."


An industry official also stated, "If this situation solidifies, Korean manufacturing could end up operating highly advanced intelligent factories but, in reality, fail to internalize the core underlying technologies that drive those factories, effectively turning into a technological colony."



Amid these concerns, some efforts are emerging to reduce dependence on specific platforms. LG Electronics has been selected as the lead agency for the government-led project to localize physical AI infrastructure and has begun developing its own proprietary robot foundation model. Doosan Robotics is also focusing on internalizing task-specialized intelligence to perform unstructured skilled work in manufacturing environments at a high level of proficiency.


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing