SK Group Chairman Chey Tae-won forecasts that the global semiconductor wafer shortage could persist until 2030, driven by the soaring demand for artificial intelligence (AI) technologies.
Speaking to reporters at Nvidia’s annual graphic processing unit (GPU) Technology Conference (GTC) event in San Jose, California on Monday, Chey explained that AI requires vast amounts of high-bandwidth memory (HBM), and HBM production consumes a significant number of wafers.
He elaborated that ramping up wafer supply would take at least 4 to 5 years, suggesting that the current shortage might continue until 2030, with an anticipated deficit exceeding 20%.
SK Hynix, a crucial HBM supplier for Nvidia, dominates the HBM market with a 57% share, according to Counterpoint Research. The company also holds the second position globally in the dynamic random access memory (DRAM) market, commanding a 32% share.
When asked about DRAM price stabilization, Chey remained cautious, stating that while he can’t provide specifics at this time, he anticipates the Chief Executive Officer (CEO) will announce new strategies soon.
Addressing the potential listing of SK Hynix’s American Depositary Receipts (ADRs), Chey noted that such a move would broaden the company’s global investor base and boost its international profile. He pointed out that ADRs would allow U.S. investors to trade SK Hynix shares without a direct listing on the New York Stock Exchange.
Regarding expansion of U.S. production, Chey highlighted several challenges, including securing power, water, suitable construction conditions, and skilled engineers. He admitted that responding immediately to demand isn’t straightforward, adding that the company’s current focus remains on production in South Korea.
Chey also touched on the deteriorating business environment due to rising energy prices, exacerbated by tensions in the Middle East. He emphasized the growing importance of securing alternative energy sources in light of these developments.