South Korea has announced plans to secure 10,000 high-performance graphics processing units (GPUs) by the end of this year to enhance its competitive position in the rapidly evolving global AI race. The government aims to secure these GPUs through a public-private partnership to support the early launch of services at its national AI computing center.
Acting President Choi Sang-mok emphasized the growing shift in the AI industry, where the competition is evolving from corporate rivalries to national innovation ecosystems competing on a global scale.
This move comes as the US government introduced new regulations aimed at controlling the flow of AI chips and technology, including GPUs, essential for cutting-edge AI applications. These specialized processors, originally designed for graphics rendering, are increasingly integral to AI model training. The US restrictions classify countries into different tiers. South Korea is among 18 countries largely exempt from the new rules, while 120 other nations will face limitations. Countries like China, Russia, and Iran are entirely barred from receiving these chips.
The South Korean government has not yet specified which GPU products will be procured, but budget, GPU models, and participating private companies are expected to be finalized by September 2024, according to officials from the Ministry of Science and ICT.
Notably, Nvidia, the dominant player in the global GPU market with about 80% market share, has seen soaring demand for its products, particularly from customers involved in generative AI and accelerated computing. It remains the leading supplier, far surpassing competitors like Intel and AMD.
We use cookies to ensure you get the best experience on our website. Read more...