Seoul: The chief executive of artificial intelligence chip giant Nvidia Corp. emphasized the significance of its partnership with South Korea’s SK hynix Inc. during a technology conference. The collaboration focuses on the development of advanced AI accelerators, with an emphasis on the aggressive advancement of high bandwidth memory (HBM) chips.

According to Yonhap News Agency, Nvidia CEO Jensen Huang, speaking via video at the SK AI Summit 2024 hosted by SK Group, highlighted the need for more bandwidth with reduced energy consumption in HBM. He described SK hynix’s development plan for HBM as both super aggressive and necessary, given the evolution of AI from text generation to data retrieval involving images and videos.

SK hynix distinguished itself as the first chipmaker globally to supply fifth-generation HBM3E chips to Nvidia in March, enhancing graphic processing units crucial for AI computing. The company initiated mass production of 12-layer HBM3E chips last month, reinforcing its role in advancin
g AI technologies.

Huang, emphasizing Nvidia’s position as a computing platform company, acknowledged the crucial partnership with SK hynix in co-designing AI accelerators. He noted that SK hynix’s innovation in HBM and custom memory chips significantly contributed to Nvidia’s architectural work.

At the summit, OpenAI President Greg Brockman also addressed the audience, discussing South Korea’s potential competitive edge in the AI industry through government support. He drew parallels between South Korea’s successful broadband expansion in the 1990s and its current AI initiatives. In September, South Korea launched a presidential AI committee to formulate strategies for becoming a global leader in cutting-edge technologies.