TechFlow Logo
Login/ Sign up
ETH Gas
Gwei
Fear
gas
Morgan Stanley Research Report Analysis: Minor Fluctuations in the Memory Chip Cycle, AI Demand and LTA Reshape Long-Term Value

Morgan Stanley Research Report Analysis: Minor Fluctuations in the Memory Chip Cycle, AI Demand and LTA Reshape Long-Term Value

2026.08.07
Share
Long-term agreements are fundamentally changing the earnings visibility and capital return logic of the memory industry.

By: Rita

DRAM contract prices QoQ increase in Q3 fell from expected 20% to around 15%, memory module inventory rose to 12.5 weeks, signs of weakness in China consumer electronics demand. Morgan Stanley admitted in its latest memory industry research report released in August that the second derivative of pricing is peaking, and small cycle fluctuations have emerged. But Morgan Stanley's core judgment is that this is just a normal cycle slowdown, not a turning point. AI capex is still accelerating, capex growth rate of the four major hyperscalers was raised from previously predicted 14% to 29%. Long-term agreements are fundamentally changing the earnings visibility and capital return logic of the memory industry. Morgan Stanley maintains Overweight ratings on Samsung and SK Hynix, with target prices of 381,000 KRW and 2,600,000 KRW respectively, implying 74% to 84% upside potential.

Second Derivative of Pricing Peaks, AI Demand Changes Cycle

Q3 DRAM contract price early trading QoQ increase was about 15%, lower than Morgan Stanley's previously expected 20%. NAND contract prices QoQ increase was about 20%. Reasons for narrowing price increases come from both ends. On the demand side, China consumer electronics demand weakened, PC and smartphone customers have hit cost tolerance limits, procurement urgency declined. On the supply side, module manufacturers and distributors inventory rebounded from low levels to 12.5 weeks, procurement strategy tending towards normalization. DDR3 spot prices have fallen below contract prices, an early signal of weak consumer demand.

But Morgan Stanley believes these signals do not constitute a trend reversal. AI-related demand remains strong, enterprise SSDs are in short supply, cloud service providers are still actively signing long-term agreements for DRAM, aiming to cover at least 90% of demand. HBM pricing is still under negotiation, suppliers aim to flatten HBM profit margins to levels comparable to commodity DRAM, implying 50% to 100% annual price increases. DRAM has become one of the key bottlenecks in AI construction, but the market seems unconvinced of the cycle's sustainability, current memory stocks trade at about 3x forward P/E, with almost no growth premium priced in. Morgan Stanley's judgment is that if investors begin to reprice earnings durability and capital return prospects, multiple expansion will become a larger return driver.

LTA Changes Industry Structure, Capital Return Accelerates

Morgan Stanley views the progress of long-term agreements as the most underestimated structural change in this memory cycle. Samsung explicitly disclosed for the first time in Q2 financial reports, targeting to place 60% to 70% of capacity under long-term agreements, has completed signing with the top five data center customers globally, with another five customers in final negotiation stages. Contracts adopt a rolling five-year structure, customers need to pay in advance, about 25% of expected payments have been received. Samsung also set a minimum price floor for mass products to limit volatility risk.

SK Hynix has completed LTA negotiations with about 10 customers, agreement terms are usually about five years, pricing structure customized to absorb volatility. Micron has completed 16 customer agreements, covering about 20% of DRAM capacity and one-third of NAND capacity, terms spanning 2026 to 2030, adopting take-or-pay and non-cancellable structure, minimum price revenue about $100 billion. SanDisk has signed 10 new business model agreements, minimum contract revenue about $93.9 billion.

Morgan Stanley believes that if the market fully values LTAs, memory stocks' trading multiples should be higher. Investors are currently still unwilling to assign significant value until these agreements prove their resilience during an industry downcycle. But LTA has moved from intention stage to concrete disclosure stage, every major supplier has quantified the output or revenue proportion planned to be included in contracts, multiple companies disclosed deposits and prepayments. As LTA verifies its downside protection role in the next cycle, the memory industry's earnings durability may face repricing.

Samsung and SK Hynix Valuations Attractive

Morgan Stanley maintains Overweight on both Samsung and SK Hynix. Samsung target price 381,000 KRW, based on residual income valuation model, implying 2027 expected P/B ratio about 2x, consistent with peak levels of this industry commodity cycle. Morgan Stanley expects Samsung HBM will improve in second half of 2026 and may gain market share, commodity pricing trend positive, 50% of free cash flow used for capital return. Current stock price corresponds to 2026 expected P/E about 5x, 2027 about 3x.

SK Hynix target price 2,600,000 KRW, implying about 74% upside potential, 2027 expected P/E about 5.8x. AI inference demand will support commodity cycle to remain strong in 2026 to 2027, HBM leadership position solid. 2026 EPS forecast raised 13%, mainly reflecting Q2 asset disposal gains, 2027 to 2028 forecasts basically unchanged. SK Hynix target price corresponds to 2027 expected P/B ratio about 2.6x.

Morgan Stanley's sensitivity analysis shows the market has not yet assigned any meaningful P/E premium to LTA-supported memory earnings and free cash flow. Current valuations are at historical lows, Samsung and SK Hynix's NTM P/B ratios are about 1.4x and 1.3x respectively, far below the 2x or more during commodity cycle peaks. If AI capex ROIC story continues to verify, and LTAs prove their value in the next downcycle, memory stocks may face a reshaping of valuation system.

Disclaimer

This article is a compilation and interpretation by TechFlow Research of a third-party broker research report (Morgan Stanley, August 6, 2026), combined with public market information. Ratings, target prices, earnings forecasts and related judgments cited in the text are the views of the broker's analysts, represent only their affiliated institution's position, do not represent TechFlow Research's views, nor constitute any investment advice.

Market involves risks, decisions must be independent. This article should not be used as a basis for buying or selling any securities.