NVIDIA and LG Group are expanding their AI infrastructure work into a broad "AI factory" plan that connects robotics, mobility, data centers, model development, and enterprise AI agents.

The announcement is more specific than a simple chip-supply partnership, though it is still mostly a roadmap rather than a finished deployment. NVIDIA says the factory is meant to support LG's robotics, autonomous driving, data-center technologies, and GPU cloud services. LG's own release frames the work around mobility, AI infrastructure, and physical AI.

The robotics track centers on LG Electronics and LG CNS. NVIDIA says LG can use Isaac Sim and Isaac Lab to simulate and validate home cobots, while also exploring Isaac GR00T for home robots and modular robotics platforms. LG CNS is tying NVIDIA robotics software, Cosmos world models, and Isaac GR00T into PhysicalWorks, its industrial robot platform for manufacturing and logistics sites.

The data-center side involves several LG units. NVIDIA says LG Uplus plans to build a large AI data center capable of hosting current NVIDIA GPUs, while LG Energy Solution is looking at 800 volt direct-current data-center energy systems aligned with NVIDIA DSX guidance. LG CNS is also positioned around GPU-powered AI factories.

For developers, the sovereign-model piece may be the most concrete. LG AI Research is using Blackwell GPUs, NeMo, Nemotron datasets, and TensorRT-LLM for EXAONE development and deployment, while LG plans broader use of EXAONE-based agents such as ChatEXAONE across the group.

The conservative read: this is an infrastructure collaboration, not a product launch. But it shows how NVIDIA is packaging physical AI as a stack that spans robots, simulations, GPUs, power systems, and enterprise agents.