Nvidia CEO Jensen Huang. AP-Newsis
Nvidia and SK Group are deepening their collaboration in the field of physical artificial intelligence (AI). The plan is to accelerate the realization of an autonomous factory by applying Nvidia’s digital twin technology to SK hynix’s (000660.KS) semiconductor fabrication plants (fabs). SK Telecom (017670.KS) is also participating in the project to carry out technology optimization.
According to SK Telecom on Tuesday, Nvidia CEO Jensen Huang mentioned SK Telecom as a key partner in the manufacturing physical AI field during his keynote speech at the “GTC Taipei” event held in Taiwan that day. He introduced the case in which SK Telecom and SK hynix adopted Omniverse in the semiconductor manufacturing process. Omniverse, developed by Nvidia, is a collaboration platform for digital twins and 3D simulation.
The project is one of the steps SK hynix is taking to achieve its “Autonomous Fab 2030” goal. SK hynix and SK Telecom completed technology verification for applying digital twins to semiconductor fabs last year and plan to commercialize it in stages. Digital twin is a technology that replicates actual factories and equipment in a virtual space, allowing the impact of process changes and equipment layouts to be verified in advance through simulation. Because various scenarios can be tested in a virtual environment, it is drawing attention as a core physical AI technology that reduces trial and error and supports data-driven decision-making.
This is not the first time SK Telecom has appeared in CEO Jensen Huang’s keynote. At the GTC 2026 keynote held in San Jose, U.S., last March, the company was also introduced as a strategic partner of Nvidia alongside Boston Dynamics and others.
SK Telecom plans to implement a stable and scalable digital twin environment even in complex, large-scale data manufacturing environments such as semiconductor fabs. To this end, the company has developed “agentic digital twin modeling” technology, which uses Nvidia’s agent toolkit to automate and intelligently process diverse data—such as equipment and spatial structures on the manufacturing floor—to fit a digital twin environment. The technology is expected to improve the efficiency of data conversion, scene optimization, and performance enhancement work that occurs during the construction and operation of digital twins on the manufacturing floor.
“A semiconductor fab is one of the most demanding manufacturing environments, combining large-scale 3D data, complex equipment structures, and highly optimized requirements,” said Mike Geyer, head of industrial digital twin at Nvidia. “SK Telecom has demonstrated a high level of technical capability to apply and verify the Nvidia Omniverse agent toolkit at actual industrial sites in such environments.” Cho Ik-hwan, head of physical AI at SK Telecom, stressed, “Through our collaboration with Nvidia, we have confirmed that manufacturing digital twins can evolve beyond simple 3D visualization into a physical AI platform where AI understands and optimizes large-scale 3D data on the manufacturing floor.”