SK Telecom puts SK hynix fabs into an NVIDIA Omniverse twin, following Samsung and TSMC

SK Telecom said on June 1 that it has put SK hynix’s semiconductor fabs into a digital twin built on NVIDIA Omniverse libraries, with commercialization to follow in stages under SK hynix’s “Autonomous Fab 2030” roadmap. SK hynix’s news is part of a trend involving Omniverse-based fab twins. SK hynix and SKT first surfaced the work at the APEC summit on October 30, 2025, and Samsung announced its own Omniverse fab digital twins a day later, inside a 50,000-GPU “AI Megafactory.”

The world’s largest dedicated semiconductor foundry, TSMC, disclosed its own Omniverse fab environment, FabTwin, for evaluating process-tool layouts.

The fab announcements extend a roster NVIDIA has been assembling for months. At GTC Washington, D.C. on October 28, 2025, NVIDIA named Belden, Caterpillar, Foxconn, Lucid Motors, Toyota, TSMC and Wistron as building Omniverse factory digital twins.

SKT says it will use “Agentic Digital Twin Modeling,” built on the NVIDIA Agent Toolkit to automate the data-conversion work of turning equipment and spatial data into a usable virtual environment.

The fab-twin push runs on the same AI demand that drives NVIDIA’s core business. SK hynix is a leading supplier of the high-bandwidth memory that are a part of NVIDIA’s AI accelerators. The surge in HBM orders has pushed memory makers to add capacity at a pace that makes layout and process errors expensive. WSJ reported that SK hynix announced a fivefold increase in quarterly net profit based on skyrocketing demand for its memory in first-quarter results announced in April.