Home > IT/기술 > HBM Market Forecasted for Stable Growth Amid Debates Over End of Semiconductor Super Cycle

HBM Market Forecasted for Stable Growth Amid Debates Over End of Semiconductor Super Cycle

IT_기술 2026-03-27 IT/기술

Analysis of the Background Behind the Stock Price Decline of Samsung Electronics and SK Hynix Recently, the stock prices of Samsung Electronics and SK Hynix have fallen by approximately 20% within a month, delivering a significant shock to the South Korean semiconductor industry. This decline, coupled with debates over whether the global semiconductor market's 'Super Cycle' is ending, has created tension across the industry. Notably, this downturn occurred after NVIDIA announced its new AI-centric GPU, 'Blackwell,' leading to a readjustment of investor expectations for the AI semiconductor market. The sharp drop in stock prices is interpreted as reflecting market anxiety that the growth momentum of the AI semiconductor market might slow down compared to investors' initial expectations. To understand the semiconductor market, it's necessary to look back at industry trends over the past few years. A semiconductor super cycle typically refers to a period when global economic growth and technological development converge, leading to an explosive increase in demand for memory and logic semiconductors. Over the past three years, this super cycle has been bolstered by the rising demand for AI semiconductors and data centers. However, the recent stock price correction reveals market concerns about this growth momentum. Even SMIC (Semiconductor Manufacturing International Corporation), a Chinese foundry, has warned that memory supply shortages could constrain automotive and consumer electronics production by 2026. Nevertheless, securities firms and industry experts project that the global memory semiconductor market will maintain a relatively stable trend through 2026, even amidst a moderate economic slowdown. This outlook contrasts with some predictions of a market collapse, suggesting that structural changes in the semiconductor market are providing new growth drivers. Notably, global big tech companies like Google and AWS, in addition to NVIDIA, are expanding their development of ASIC (Application-Specific Integrated Circuit)-based AI chips. This trend is broadening AI infrastructure investment from general-purpose GPUs to specialized areas tailored to specific customers and applications. Such changes further highlight the strategic importance of South Korean companies in the memory semiconductor market, particularly in the High Bandwidth Memory (HBM) sector. HBM, or High Bandwidth Memory, market is considered a key to the domestic semiconductor industry maintaining its competitive edge in the global environment. Specifically, the High Bandwidth Memory (HBM) market is cited as a core driver for this stability, with its size estimated to reach $54.6 billion by 2026, a 58% increase year-over-year. This indicates a rapid increase in HBM's share of the overall memory semiconductor market. Even more noteworthy is the explosive growth of the ASIC-based AI chip market. Goldman Sachs projected that HBM demand for ASIC-based AI chips would surge by 82%, accounting for one-third of the total HBM market by 2026. This clearly demonstrates a shifting paradigm in AI infrastructure investment. The transition from general-purpose GPUs to customized chips optimized for specific applications presents both new opportunities and challenges for memory semiconductor companies. Future of the HBM Market and Global Competitive Landscape Following Blackwell, NVIDIA is setting the standard for next-generation AI infrastructure by introducing its new 'Blackwell Ultra' series of AI accelerators. These new products all adopt HBM3E as a core component, and HBM3E is expected to remain the 'golden standard' in 2026. South Korea's Samsung Electronics and SK Hynix have already secured leading global positions in the HBM3E sector and are focusing on meeting the demand from NVIDIA and other global big tech companies based on this advantage. The reasons why HBM plays a crucial role in the advancement of AI and cloud technologies are clear. Memory semiconductors continue to play a pivotal role in terms of demand and profitability due to the expansion of server and data center investments driven by the proliferation of AI infrastructure. Particularly, as the number of parameters in generative AI models increases exponentially, the demand for high-capacity, high-speed memory is growing even further. These technical requirements are fundamental drivers supporting the sustained growth of the HBM market. Considering counterarguments, caution regarding semiconductor investment is also being raised. Some argue that the super cycle merely repeats existing patterns, and new technology development might progress slower than anticipated. Especially with increasing global economic uncertainty, there are concerns about the risks associated with large-scale capital expenditures. Furthermore, it is pointed out that as competition intensifies, companies lacking a local investment base might be eliminated from the market. However, data and market analysis offer a posi