SK Telecom (SKT) partnered up with Nvidia to use the chip company’s digital twin technology for semiconductor manufacturing environments operated by SK Hynix, pushing industrial AI deployments to achieve more automated factory operations.

The operator said it used Nvidia “Omniverse libraries” to adapt digital twins for complex, large-scale manufacturing environments, following a proof-of-concept completed last year at a SK Hynix semiconductor manufacturing site. It plans to commercialise the technology in stages as SK Hynix works to establish autonomous fab operations by 2030.

Using Nvidia’s Agent Toolkit, SKT developed Agentic Digital Twin Modelling technology to automate data processing, including site equipment and spatial structures, for use in digital twin systems. It also integrated Nvidia Omniverse libraries to make large 3D factory scenes load faster, run more smoothly and use GPU and memory resources more efficiently.

The set up aims to improve data conversion, scene optimisation and performance tasks required to build and run digital twins at scale.

SKT explained digital twins act as working replicas of real factories and equipment. In semiconductor plants, they can be used to test changes to processes or equipment layouts in advance, helping reduce costly trial and error in highly complex production sites.

Mike Geyer, head of industrial digital twins at Nvidia, said semiconductor fabs are “among the most challenging manufacturing environments”, citing “massive amounts of 3D data, complex equipment structures, and the need for high-level optimisation”.

Cho Ik-hwan, head of physical AI at SKT, added the collaboration demonstrates manufacturing digital twins can move “beyond simple 3D visualisation” into systems capable of “understanding and optimising large-scale 3D manufacturing data”.

SKT added the move bolsters its plans to expand its enterprise and public sector business with AI offerings spanning infrastructure, models and services.

The post SKT puts Nvidia digital twins to work in chip fabs appeared first on Mobile World Live.

Nvidia CEO Jensen Huang positioned the adoption of humanoid robots in industry as opening a multitrillion-dollar economic opportunity, as it announced a model for academics using hardware from Unitree and Sharpa intended to accelerate advances.

In an announcement made at Computex 2026 in Taipei, Taiwan, the executive backed humanoid robotics to “bring physical AI to the world’s largest industries” but indicated there were barriers to academic work to this end, which it aims to resolve by introduction of the “reference robot”.

The machine uses Nvidia compute systems and Isaac GR00T development platform, a Unitree H2 body standing at almost 6 feet tall and weighing 50 pounds in weight, and Sharpa Wave tactile five-finger hands.

“Nvidia Isaac GR00T Reference Humanoid Robot gives researchers a single, open platform to make breakthrough discoveries toward general-purpose physical intelligence,” Huang added.

During his keynote at the event Huang explained “we built this for higher education and university researchers, because for them to build this is insanely hard to do”, pointing to the complexities and expense of starting from scratch in every project.

Nvidia noted by using its “compute and open software stack” at the core “the reference design gives research teams a more unified, secure foundation for advancing humanoid robotics”.

Discussing Sharpa’s role founder David Li said “partnering with Nvidia on a humanoid robot reference design and end-to-end development solution is a meaningful step toward deploying robots that can perform real work, in real settings”.

The executive added its “vision is to make robots genuinely productive – by advancing fine manipulation skills through dexterous, tactile hardware and the AI models that power them”.

The post Nvidia chief pushes industrial humanoid robot opportunity appeared first on Mobile World Live.