LTAs Reshape the Memory Cycle; Morgan Stanley Raises SK Hynix Target Price to KRW 2.6 Million

Overview

This report provides an in-depth analysis of how long-term agreements (LTAs) are structurally transforming the business model of the memory industry. It notes that, driven primarily by AI infrastructure demand, hyperscalers (large cloud service providers) are signing 3–5 year LTAs with substantial prepayments and strict contractual terms to secure supply. This shift weakens the traditional cyclicality of the memory industry, enhances earnings visibility, and improves free cash flow (FCF). Based on this logic, Morgan Stanley has upgraded its earnings forecasts and target price for SK Hynix, arguing that the market has not yet fully reflected the valuation re-rating potential enabled by LTAs.

Core views

Structural Shift Driven by LTAs: Unlike soft commitments during prior semiconductor upcycles, today’s LTAs are highly binding. They commonly include large prepayments (some suppliers have already secured prepayments covering 50% of 2027 demand and 100% of 2028 demand) and non-cancellable, non-refundable (NCNR) clauses. Effectively, customers are paying an 'insurance premium' for supply security—substantially protecting suppliers’ margins and pricing power during potential downturns. Weakening Cyclicality and Improved Cash Flow: LTAs lock in both volume and pricing for future deliveries, enabling more targeted capital expenditures (Capex) aligned with confirmed demand rather than speculative forecasts—thereby lowering inventory risk and overcapacity likelihood. Even during broad industry oversupply, the LTA-covered portion buffers price and margin pressure. As the industry moves past the high-Capex phase, higher capacity utilization and stable demand will significantly boost free cash flow yield in 2026–2028—supporting share buybacks and shareholder returns, similar to Apple’s capital return path. Upgraded SK Hynix Earnings Forecasts: Given downside protection from LTAs, stronger commodity pricing assumptions, and HBM price resets in 2027, the report upgrades SK Hynix’s 2026/2027/2028 EPS forecasts by 6%/3%/14%, respectively. The prior expectation of HBM price declines has been revised to an average annual increase of 15% through 2026–2028. Using the Residual Income Model (RIM), the target price is substantially raised to KRW 2.6 million, implying 38% upside. Valuation Re-rating Potential: Sensitivity analysis shows that even under conservative assumptions—e.g., only 50% of commodity memory covered by LTAs at a 6x P/E—the implied blended P/E exceeds current levels. If LTA coverage reaches 70%–80% and receives appropriate multiples, the implied blended P/E could reach 8.5x–10.7x. Currently, the market continues to value LTA-supported revenue using traditional cyclical commodity-memory logic (~5x P/E), indicating clear undervaluation.

Analysis framework

The report follows an analytical chain of 'supply-demand structural change → business model restructuring → financial model revision → valuation re-rating.' First, it contrasts historical LTAs with current ones—highlighting differences in enforceability and prepayment mechanisms—to establish the logic behind weakening cyclicality. Second, it quantifies LTA coverage ratios (>50%) using TrendForce data and management guidance from companies including SK Hynix, Samsung, and Micron. Third, it revises key financial model assumptions—such as HBM ASP growth rates and commodity memory price trends—to recalculate EPS and free cash flow. Finally, it uses sensitivity analysis to decompose valuation multiples between LTA-supported and non-LTA revenue, derives the target price via the Residual Income Model (RIM), and draws analogies to post-pandemic re-ratings of Japanese shipping stocks for supporting evidence.

Methodology notes

• Valuation MethodRIM (Residual Income Model)

  Residual Income Model

  A valuation method based on book value and excess earnings. The report applies RIM to calculate SK Hynix’s target price, particularly suitable for asset-heavy firms undergoing a transition from cyclical to stable earnings, as it captures value created when return on equity (ROE) exceeds the cost of capital.

• Industry/Industrial Analysis FrameworkSupply-demand framework

  Supply-Demand Framework

  An analytical approach focusing on core industry tensions. The report highlights that AI semiconductor demand is growing 50–60% annually, while DRAM supply growth is constrained to ~30% annually due to EUV equipment bottlenecks—this persistent supply-demand imbalance forms the fundamental basis for LTAs’ adoption and their strong enforceability.

• Corporate Fundamentals & Financial FrameworkFree cash flow analysis

  Free Cash Flow Analysis

  An assessment of a company’s ability to generate cash. The report emphasizes how LTAs reduce inventory risk and optimize Capex, thereby significantly improving memory companies’ free cash flow (FCF) yield—a critical financial metric underpinning valuation re-rating and shareholder returns (e.g., buybacks).

• Event-Based Game Theory & Behavioral FinanceExpectation Gap / Expectations Management

  Expectation Gap / Expectations Management

  The discrepancy between market perception and underlying fundamentals. The report argues that the market currently values memory companies as traditional cyclical stocks (low P/E), failing to price in the earnings stability premium conferred by LTAs—an expectation gap representing the primary investment opportunity.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• SK hynix (000660.KS)

  Direct beneficiary, HBM leader, primary LTA signatory

  Strengths

  Technology leadership in HBM, robust downside protection and pricing power from LTAs, significantly improved free cash flow outlook

  Comparison

  Relative to Samsung, SK Hynix’s HBM leadership grants it greater leverage in LTA negotiations and higher earnings elasticity

  Risks

  Intensifying HBM competition eroding margins, weaker-than-expected macro demand

• Samsung Electronics

  Major beneficiary, actively transitioning contract structures

  Strengths

  Massive scale, converting quarterly/annual contracts to 3–5 year LTAs, having secured final contracts from select customers

  Comparison

  Also benefits from LTAs, but the report focuses on SK Hynix’s valuation re-rating potential

  Risks

  Labor-related issues (per other reports), excessive supply-side spending driven by DDR5 competition

Key data

• SK Hynix Target PriceKRW 2,600,000Based on the Residual Income Model, implying 38% upside
• EPS Forecast Adjustments+6% (2026E), +3% (2027E), +14% (2028E)Reflecting downside protection from LTAs and upgraded HBM pricing assumptions
• HBM ASP Forecast+15% YoY (2026–2028E)Revised from prior expectations of 5–10% annual declines to sustained growth
• Expected LTA Coverage Ratio>50%More than half of memory supply expected to be locked in via LTAs within the next 2–3 years
• Implied P/E Sensitivity8.5x – 10.7xImplied valuation under 70%–80% LTA coverage and reasonable multiples

Impact & implications

The report concludes that the widespread adoption of LTAs signals a structural shift for the memory industry—from a 'boom-and-bust' highly cyclical model toward a more stable, high-margin, long-term revenue model. For leading players such as SK Hynix and Samsung, this implies markedly improved earnings visibility and a fundamental re-rating of valuation frameworks. Investors should focus on companies able to lock in long-term profits via LTAs and maintain strong technological leadership in HBM. Additionally, sustained free cash flow generation may prompt increased share repurchases, further boosting earnings per share.

Risks

• Weaker-than-expected end-market demand
• Intensified DDR5 competition leading to excessive supply-side spending
• Elevated inventory levels among cloud and Chinese smartphone customers
• Intensifying competition in the HBM segment eroding margins

What to watch

• Supply-demand outlook for DRAM, NAND, and HBM
• Global demand for data center products (server DRAM, enterprise SSDs)
• Re-acceleration of average selling price (ASP) growth
• Significant increases in capital return