Guolian Minsheng Securities: Remains optimistic about the current storage super cycle

According to the report released by Guolian Minsheng Securities, Zhitong Finance APP has learned that it remains optimistic about the current storage super cycle. The high prosperity of AI demand, strengthened supply constraints, long-term contracts locking in profits, and improved shareholder returns together drive the upward shift of the industry profit center. On the demand side, AI training is rapidly evolving towards inference and Agentic AI, which not only continues to boost the demand for HBM, but also significantly increases the configuration requirements for server DRAM and enterprise-level SSDs, with storage demand expanding from single HBM to broader data center storage. Meanwhile, constraints from advanced process nodes, packaging and testing, and capacity switching continue to limit supply expansion, meaning DRAM and NAND prices are expected to continue rising from a high base, with the persistence of price increases noticeably stronger than traditional cycles.

On the supply side, leading storage manufacturers are using LTA/SCA contracts to lock in procurement volumes and price ranges for several years in advance. Micron has already secured more than 35% of its revenue before 2030 through long-term agreements, significantly improving industry earnings visibility. Supported by robust profitability and free cash flow, shareholder returns for storage manufacturers are simultaneously improving. Micron clearly stated its intention to return 100% of its excess cash to shareholders over the long term.

The brokerage believes that AI is driving a reevaluation of the storage industry from a strong cyclical sector to one characterized by "high prosperity + high certainty + high shareholder return." At present, the industry is still in a phase of upward revisions to profit expectations during the super cycle, and continues to be optimistic about the storage industry chain. Suggested areas of attention: 1) AI chips: NVDA, AVGO, MRVL, CBRS; 2) Storage: MU, SNDK, Samsung, SK Hynix, Kioxia; 3) MLCC: Samsung Electro-Mechanics, Murata Manufacturing, Taiyo Yuden; 4) Substrates: IBIDEN; 5) CPU: INTC, AMD, ARM, Qualcomm.

Main viewpoints of Guolian Minsheng Securities are as follows:

TrendForce forecasts that demand for AI servers will support the continued rise of storage contract prices in Q4 2026. TrendForce notes that in Q4 2026, DRAM suppliers will continue to prioritize allocating advanced process capacity to high-performance server products, and the overall DRAM market will remain in short supply. However, the pace of contract price increases is expected to moderate, with traditional DRAM prices anticipated to rise 10–15% quarter-on-quarter. The NAND Flash market is showing divergent trends with accelerated AI demand and weak consumer demand; prices across segments are still generally rising, and overall NAND Flash contract prices are expected to increase 15–20% quarter-on-quarter.

Updates on Overseas Technology Companies

Micron released its FY26 Q4 financial report, with revenue, gross margin, and EPS all exceeding prior guidance. FY26 Q4 revenue was $54.2 billion, up 379% year-on-year and 31% quarter-on-quarter; Non-GAAP EPS was $33.42, up 33% quarter-on-quarter; Non-GAAP gross margin was 87%, an increase of 2.1 percentage points quarter-on-quarter. Both DRAM and NAND saw simultaneous growth in volume and price. DRAM revenue was $39.8 billion, up 27% quarter-on-quarter, with mid-single-digit growth in bit shipments and an approximate 20% quarter-on-quarter increase in price. NAND revenue reached $14.1 billion, up 42% quarter-on-quarter, about 10% growth in bit shipments, and pricing up about 30% quarter-on-quarter. Data center SSD revenue was approximately $10 billion, accounting for more than two-thirds of NAND revenue. HBM quarterly revenue growth outpaced the company overall, and the vast majority of HBM bit supply for 2027 is already under contract, with significant year-on-year price increases. For FY27 Q1, revenues and EPS are expected to continue growing, but gross margin is expected to be the lowest of the year. The company guides for Q1 revenue of $61.5 billion ± $1.5 billion, a midpoint increase of about 13.4% quarter-on-quarter; Non-GAAP EPS of $38.15 ± $1, midpoint up about 14.2%; Non-GAAP gross margin about 86.25%, a decrease of 0.75 points quarter-on-quarter.

Contract scale continues to expand, making future volume and price floors clearer. The company has already signed 26 SCA long-term agreements, covering over 35% of revenue through 2030; about three-quarters of the expected agreement revenue has clear price frameworks, most with defined upper and lower price limits. Customers have made financial commitments totaling $32 billion, the vast majority in cash deposits. Therefore, even executing at contract floor prices, profit margins will still be significantly higher than previous cycle peaks. Meanwhile, new long-term pricing negotiations are advancing at much higher price levels. Management plans to further increase capital returns from December 9, expecting the cash balance to approach its target level by the end of FY27 Q1. Afterward, most excess cash will be returned through buybacks, and the company plans to seek additional buyback authorization. The long-term goal is to return 100% of excess cash to shareholders.

AMD to Acquire World Labs to Advance Next-Generation AI Computing.

On September 28, 2026, AMD announced it has signed a definitive agreement to acquire World Labs, an AI models and research laboratory led by AI pioneer Li Fei-Fei. This acquisition will bring a world-class team of AI researchers and model experts to AMD, further enhancing its capabilities in developing AI hardware, software, and systems to meet emerging AI model and application demands. The transaction will be conducted entirely in stock, with a deal valuation of approximately $8.2 billion. Pending regulatory approval and satisfaction of other customary closing conditions, the transaction is expected to be completed by the end of 2026.

Risk Warnings: AI development falling short of expectations; AI commercialization not meeting expectations; weaker-than-expected macroeconomic growth, etc.