Chip Stocks Retreat After SK Hynix Warns of Demand Challenges

Image Credit: Reuters

Chip stocks fell sharply on Thursday after a key Nvidia (NVDA) supplier signaled uncertainty about semiconductor demand for the year during its earnings call.  

South Korea’s SK Hynix (000660.KS), which produces memory chips for Nvidia’s GPUs used in AI data centers, exceeded analysts’ expectations with its fourth-quarter earnings. However, SK Hynix’s head of finance, Woo-Hyun Kim, warned of a murky 2025 outlook for memory demand. Kim cited inventory adjustments by PC and smartphone manufacturers, as well as geopolitical risks and protective trade policies, as factors contributing to the uncertainty.  

Following these comments, Nvidia initially dropped 2% before recovering to close flat. Other chip stocks saw declines, with British chip designer Arm (ARM) falling over 9% and SK Hynix competitor Micron (MU) losing 4%. These losses followed a rally earlier in the week after U.S. President Donald Trump announced "Stargate," a major AI infrastructure project funded by OpenAI, SoftBank (SFTBY), Oracle (ORCL), and UAE-based MGX.  

Kim’s remarks highlighted the contrasting trends in the semiconductor market. Chips for consumer products like PCs and smartphones are struggling, while those for AI data centers see strong demand.  

Needham analysts noted in a January 13 investor update that 2024 revealed sharp differences in semiconductor performance based on end markets. Suppliers tied to PCs, smartphones, industrial, and automotive sectors faced challenges from weak demand and inventory overhang. Conversely, AI-related semiconductors benefited from robust demand due to the rapid expansion of AI infrastructure.  

Kim emphasized the evolving dynamics in the memory chip market, stating that it is shifting from a commodity-based model focused on price and volume to a customized approach emphasizing high performance and quality. AI-driven memory demand is expected to remain strong as major tech firms continue to invest heavily in advanced AI training and inference capabilities.  

However, Needham analysts cautioned that the gap between non-AI and AI chip performance may narrow. They foresee growth in AI-related revenues slowing in 2025, as companies like Google and Microsoft recently indicated that their extensive AI spending is expected to decelerate in the near future.  

Paraphrasing text from "Reuters" all rights reserved by the original author.