South Korea has announced plans to acquire 10,000 high-performance GPUs by the end of 2025 february, as part of its strategy to stay competitive in the global AI race. Acting President Choi Sang-mok emphasized that AI competition is no longer just between companies but has escalated into a battle between national innovation ecosystems.
The government aims to secure these GPUs through public-private partnerships to facilitate the early launch of services at South Korea’s national AI computing centre.
Global AI Chip Regulations & Impact
Last month, the U.S. government introduced new regulations restricting the export of advanced AI chips, categorizing nations into different tiers:
- South Korea is among 18 exempt countries with unrestricted access to U.S. AI chips.
- 120 other nations face export caps, while Iran, China, and Russia are completely barred.
South Korea has not yet finalized which GPU models or brands it will procure, but decisions on budget, products, and private-sector involvement are expected by September 2024, according to the Ministry of Science and ICT.
Nvidia’s Market Dominance & AI Industry Shifts
Nvidia, which dominates 80% of the global GPU market, is the leading supplier of AI chips, far ahead of competitors Intel and AMD. However, tech giants like Microsoft-backed OpenAI are working to reduce dependence on Nvidia by developing their own in-house AI chips.
OpenAI’s first in-house AI chip is set to be finalized in the coming months and will be fabricated by Taiwan Semiconductor Manufacturing Co (TSMC), according to Reuters.
Meanwhile, China-based AI startup DeepSeek is taking a different approach—developing AI models optimized for computational efficiency rather than relying solely on high-powered GPUs, potentially narrowing the gap between Chinese AI processors and U.S. counterparts.
With South Korea accelerating its AI infrastructure, the global AI arms race continues to intensify, with nations and companies racing to secure hardware and computing power for next-generation AI advancements.
