GLMS Securities has released a research report stating that the current AI-driven memory upturn is not only improving profitability but, more importantly, is driving changes in the industry's business models and capital market valuation logic. For memory manufacturers, the market's focus will gradually shift from "how much further can profit margins rise" to "how long can high profitability be sustained." If leading manufacturers can persistently maintain supply discipline, strengthen their Value over Bit strategy, and extend the upcycle through methods like Long-Term Agreements (LTA), the industry is expected to see reduced profit volatility and enhanced cash flow stability. Consequently, the capital market may grant a valuation system characterized by longer duration and higher quality.

The report suggests focusing on: (1) Memory manufacturers; (2) Module makers and distributors; (3) Companies related to capacity expansion. The key views of GLMS Securities are as follows:

Recent Developments

Recently, SK Group Chairman Chey Tae-won publicly stated that global AI semiconductor demand in 2027 is expected to grow 60% to 100% year-on-year, while new supply is almost negligible. This suggests the supply-demand gap in the memory industry may further widen. Simultaneously, he expressed the view that maintaining high chip prices over the long term is not a healthy state, and that leading manufacturers should focus on expanding supply and stabilizing the industry chain for long-term development. The report notes that this statement reflects a shift in the operational philosophy of the memory industry and implies that the capital market's valuation logic for memory manufacturers may gradually transition from "profit peak" to "profit duration."

Traditional Cycle Framework Focused on Peak Profits

Over the past two decades, the memory industry has consistently followed the classic cycle of "price increases → profit release → capital expenditure expansion → oversupply → price declines." Due to the severe volatility in supply-demand dynamics, profits exhibit strong cyclical characteristics. Consequently, the capital market has tended to view memory manufacturers as typical cyclical stocks rather than growth stocks. Therefore, whether it's Samsung Electronics, SK Hynix, or Micron, market pricing has consistently revolved around peak profits. When Average Selling Prices (ASP) approach the end of an upward trend, valuations often decline prematurely. The fundamental reason is not insufficient profitability, but rather a persistent lack of market confidence in the sustainability of those profits.

AI Era Shifts Focus from Profit Maximization to Profit Sustainability

AI computing infrastructure construction is driving sustained growth in demand for HBM, DDR5, and enterprise-grade SSDs, ushering the memory industry into a new demand-driven phase. Concurrently, the business philosophy of leading manufacturers like Samsung, SK Hynix, and Micron has also undergone significant changes. The industry's competitive focus is gradually shifting from Bit growth to Value growth, and from market share competition to profit quality competition. Recently, the three major manufacturers have continued to strengthen their Value over Bit strategy. Through Long-Term Agreements (LTA), more prudent capital expenditures, and rational supply management, they are actively working to reduce profit volatility rather than pursuing short-term profit maximization.

The recent remarks from SK Hynix management, stating that "excessively rapid price increases are not a healthy long-term state," do not essentially signal a price peak. Instead, they reflect that leading manufacturers are placing greater emphasis on the long-term stability of the industry chain and the sustainability of profits. They aim to achieve maximum enterprise value by appropriately releasing supply and extending the upcycle, rather than solely pursuing peak profit margins.

Valuation Logic May Shift from Cycle-Based to Duration-Based

The report argues that the most significant change for the memory industry in the AI era is not how high profit margins can rise, but how long a state of high profitability can be maintained. If AI demand continues to be released, and leading manufacturers effectively smooth profit volatility through supply discipline, long-term agreements, and rational capacity expansion, then even if profit margins decline somewhat from the cycle peak, the sustainability of profits and the stability of cash flow are expected to be significantly enhanced.

From a valuation framework perspective, what determines enterprise value is not just the level of profit, but also the duration of that profit. As market confidence in the industry's long-term profitability gradually increases, the valuation system for memory manufacturers is expected to evolve from a traditional cyclical stock framework towards a growth manufacturing framework. There is potential for a systematic upward shift in the industry's valuation center.

Risk Factors

Potential risks include: AI capital expenditure falling short of expectations; manufacturers resuming aggressive capacity expansion and weakening supply discipline; and the market reverting to traditional cyclical stock pricing.