SEOUL, April 22 (AJP) - LG AI Research and U.S. chipmaker NVIDIA will deepen their technology alliance to co-develop specialized artificial intelligence models.
Executives from both companies, including LG AI Research Co-head Lim Woo-hyung and NVIDIA’s Vice President of Applied Deep Learning Research Bryan Catanzaro, met at LG's research headquarters in Seoul on Tuesday to discuss joint AI ecosystem strategies.
Under the agreement, the companies will combine LG’s EXAONE AI model with NVIDIA’s Nemotron open ecosystem to build domain-specific models.
The two firms have collaborated closely from the development of EXAONE 3.0 to the recently unveiled multimodal model, EXAONE 4.5. LG utilized Nemotron open datasets to ensure training data quality, while NVIDIA supplied its latest Blackwell GPUs, NeMo Framework, and TensorRT-LLM software to optimize the models' learning capabilities and inference efficiency.
A recent Stanford University AI Index Report ranked South Korea third globally, behind the U.S. and China, for the number of notable AI models. Four of the five recognized Korean models belonged to LG's EXAONE series, including EXAONE Deep, EXAONE Path 2.0, EXAONE 4.0, and K-EXAONE.
Catanzaro said the integration of EXAONE and Nemotron will help lead the development of "sovereign AI" and expand the broader ecosystem. Lim added the partnership aims to produce tangible sovereign AI results that can be applied directly to industrial sites.
The NVIDIA tie-up follows a series of meetings earlier this month by LG Corp Chairman Koo Kwang-mo with Silicon Valley tech leaders, including Palantir CEO Alex Karp and Skild AI co-founders Deepak Pathak and Abhinav Gupta, as the conglomerate moves to accelerate its AI initiatives.
Copyright ⓒ Aju Press All rights reserved.