The landmark Namsan Seoul Tower.
Jung Yeon-je | Afp | Getty Images
South Korea’s dominance in the memory chip market and a robust artificial intelligence ecosystem gives it an advantage in the global AI chip race, said industry observers.
“South Korea is very strong in memory chips. AI does require a lot of memory. South Korea dominating in the memory market is definitely an advantage,” said James Lim, senior research analyst at Dalton Investments.
South Korea is aiming to become one of the world’s top three AI powerhouses by 2027, following closely behind the U.S. and China, according to the nation’s “digital strategy.”
The country’s minister for science and information and communications technology, Jong-ho Lee, told CNBC the country “aims to maintain its leading position in the memory semiconductor field.”
“South Korea seeks to emerge as a prominent player in rapidly growing and promising areas such as AI semiconductors,” said Lee.
Large language models such as ChatGPT — which caused global AI adoption to explode in recent months — are increasingly in need of high-performance memory chips. Such chips enable generative AI models to remember details from past conversations and user preferences in order to generate humanlike responses.
Generative AI is a type of artificial intelligence that can generate content such as text, images, code and more.
“In order for the use of AI, including ultra-large language models, a significant number of semiconductor chips are required to operate, and global companies are competing fiercely to create high-performance and low-power AI semiconductors optimized for AI computation,” Lee said.
Chip giants Samsung, SK Hynix
South Korean firms Samsung Electronics and SK Hynix are two of the world’s largest dynamic random-access memory chipmakers and have been actively investing in AI research and development to bolster their capabilities.
Samsung in March said that it plans to invest 300 trillion Korean won ($228 billion) in a new semiconductor facility in South Korea.
Samsung is “spending and spending and spending,” Dylan Patel of research and consulting firm SemiAnalysis told CNBC last month. “And why is that? So they can catch up on technology, so they can continue to maintain their leadership position.”
We will spare no effort to help Korea secure world-class AI semiconductor technology by leveraging our memory semiconductor capabilities to advance AI semiconductors…
Jong-ho Lee
Minister for Science and ICT
Data from research firm TrendForce showed that Samsung held a market share of 40.7% and SK Hynix held 28.8% in the same period in the fourth quarter of 2022, followed by Micron in third place at 26.4%.
“South Korea has a robust local AI ecosystem, capable of competing with global tech giants,” said Sung Nako, executive for large scale AI development at South Korean internet giant Naver.
ChatGPT maker OpenAI’s CEO Sam Altman had urged South Korea to lead AI chip production during his meeting with South Korean President Yoon Suk-yeol in June. Altman also expressed interest in investing in South Korean startups and partnering with major chipmakers like Samsung Electronics.
“U.S. chip giants Nvidia, Intel — they are not involved in the memory business. They don’t have any exposure in the memory space,” said Dalton’s Lim, adding that this would give South Korea an advantage.
Samsung is the supplier of high bandwidth memory chips to Nvidia, which fit into the U.S. chipmaker’s latest A100 graphics processing units that train ChatGPT.
Geoffrey Cain, author of the 2020 book “Samsung Rising,” told CNBC last month that he sees Samsung “diving deeper into the logic chip segment. So, [that’s] the AI chips, the future applications for semiconductor technology.”
An ‘upper hand’
The South Korean government is investing heavily in AI.
In 2022, the MSIT said it will be deploying 1.02 trillion won ($786 million) of funding for AI semiconductor research and development over the next five years.
“AI not only drives the growth of digital industries such as cloud computing and metaverse but also serves as a key factor in dramatically improving productivity in traditional industries such as manufacturing and logistics,” Lee told CNBC.
“With AI being applied across various domains, even greater economic ripple effects can now be anticipated,” he said.
South Korea will also allocate 826.2 billion won through 2030 to build high-end chips through new data centers and working with startups.
In a press release last month, “The economic and industrial value of AI semiconductor will continue to improve and Korea has the upper hand in the memory chip [sector] and foundry.”
“We will spare no effort to help Korea secure world-class AI semiconductor technology by leveraging our memory semiconductor capabilities to advance AI semiconductors in stages by 2030, developing additional to apply them to data centers, and fostering AI semiconductor experts,” he said in the release.
In a bid to challenge to U.S. chip giants, South Korean AI chip design startup Rebellions claimed its new chip surpassed performance standards, outperforming Nvidia’s equivalent GPUs by more than three times.
“In terms of AI workload, we have much better energy efficiency, cost efficiency … sometimes better performance,” Rebellions co-founder and CEO Park Sung-hyun told CNBC in May.
Rebellions is reportedly racing to win government contracts as Seoul aims to bolster its local companies.
“I see a lot of — thanks to OpenAI’s ChatGPT — founders starting companies in the region, and also a lot of investors, with the support from the government, showing a high interest in backing these startups,” said JP Lee, CEO and managing partner at SoftBank Ventures Asia, on CNBC’s “Street Signs Asia.”
— CNBC’s Katie Tarasov contributed to this report.