Partnership shifts toward real-world deployment across manufacturing, biotech
LG and Nvidia have agreed to expand their artificial intelligence partnership, stepping up efforts to develop next-generation models and broaden the AI ecosystem as competition intensifies globally.
Executives from LG AI Research and Nvidia met in Seoul on Tuesday to discuss closer cooperation on model development and ecosystem building, LG said Wednesday.
Attendees include LG AI Research co-head Lim Woo-hyung and Exaone Lab head Lee Jin-sik, alongside Bryan Catanzaro, vice president of Applied Deep Learning Research at Nvidia and Jeong So-young representing Nvidia's Korea office.
“Nvidia has been a key technology partner in developing Exaone,” Lim said. “We will expand this collaboration across the broader R&D ecosystem to deliver tangible sovereign AI outcomes in real-world industries.”
The companies plan to combine LG’s Exaone foundation model with Nvidia’s Nemotron ecosystem to develop domain-specific AI systems for industries such as manufacturing and biotechnology — a shift toward commercial deployment rather than standalone model upgrades.
Nemotron bundles open models, datasets and tools designed to help customers build AI systems. LG intends to incorporate those datasets — drawn in part from Nvidia’s global GPU ecosystem — to improve the performance of its models.
Analysts view the collaboration reflects a realistic strategy for Korea’s AI industry. Rather than the costly and resource-intensive route of building fully independent models, companies can boost competitiveness with local data while leveraging Nvidia’s infrastructure.
Catanzaro said the company has worked closely with LG AI Research as a core partner to help position Exaone os Korea's leading AI model.
He added that specialized models built on country- and industry-specific cultural and language data have strong potential to maximize AI capabilities, and that, combining LG's Exaone with Nvidia's Nemotron ecosystem would help advance sovereign AI and expand the broader ecosystem.
The partnership builds on an existing collaboration that dates back to Exaone 3.0 and now extends to the latest multimodal version 4.5. LG said it has used Nvidia’s stack, including GPUs, the NeMo development framework and TensorRT-LLM inference software, to optimize training and deployment.
The companies said the cooperation has already yielded results. In the latest AI Index Report from Stanford Institute for Human-Centered Artificial Intelligence, South Korea ranked third globally in notable AI models, with four of the country’s five selected systems coming from LG’s Exaone lineup, including Exaone Deep, Path 2.0, Exaone 4.0 and K-Exaone.
LG said it will continue to expand partnerships with global technology firms to accelerate its AI push. LG Group Chairman Koo Kwang-mo recently met Alex Karp of Palantir Technologies and Skild AI co-founders in Silicon Valley as part of that effort.
herim@heraldcorp.com
