Intel Joins Forces with AI Alliance to Challenge Nvidia

The Naver-Intel-KAIST AI Joint Research Center (NIK AI Research Center) was officially unveiled at the KAIST main campus in Daejeon, South Korea, on April 30. This collaboration marks a significant step towards advancing artificial intelligence research and development in the region.

One of the key objectives of this partnership is to leverage the Gaudi series of AI training and inference processors from Intel. The aim is to position these processors as viable alternatives to the GPUs supplied by Nvidia for AI-related tasks. By harnessing the capabilities of the Gaudi processors, the research center hopes to drive innovation and enhance the performance of AI systems.

Naver, a prominent South Korean Internet conglomerate known for its search engine and Line messaging application, will play a crucial role in this collaboration. The company will grant access to its Naver Cloud platform, which is powered by the Gaudi2 processor from Intel. Researchers at KAIST will utilize this cutting-edge technology for their studies and are expected to publish research papers annually.

Moreover, the NIK AI Research Center plans to facilitate knowledge exchange and collaboration between industry and academia. The center aims to host between 20 and 30 industrial-academic research projects, fostering a dynamic environment for innovation and discovery. By sharing infrastructure facilities and promoting personnel exchanges, the center seeks to drive interdisciplinary research in the field of artificial intelligence.

Heading the NIK AI Research Center are Professor Kim Jeong-ho from KAIST and Lee Dong-Su, an expert in AI semiconductor design and AI software from Naver Cloud. Together, they will lead a team of approximately 20 faculty experts and over 100 master's and doctoral students who will contribute as researchers. The center is slated to operate for an initial period of three years, with the possibility of extension based on the outcomes and impact of its research initiatives.