AI demand drives MLCC and ABF into an upward pricing cycle, with Samsung Electro-Mechanics upgraded to top pick

Overview

The report argues that Samsung Electro-Mechanics is shifting from a traditional cyclical electronic components company to a platform supplier highly leveraged to AI infrastructure investment. Hyperscale cloud demand is tightening supply of high-end MLCCs, while AI ASICs, high-speed networking, and advanced packaging are accelerating demand for ABF substrates. Supply expansion requires at least one to two years, and together with long-term supply agreements and raw material cost pass-through, pricing upside is expected to be highly sustainable. The company’s synergy across MLCCs and substrates, along with new products such as embedded passive devices, silicon capacitors, and glass substrates, is expected to extend the earnings growth cycle.

Core views

The core thesis consists of three points: first, the high-end AI MLCC market is highly concentrated, certification barriers are strict, and new capacity will mainly be released after 2027, so shortages will translate into stronger pricing power; second, ABF demand mix is shifting from PCs to servers, AI GPUs, AI ASICs, and high-speed networking, with the supply-demand gap expected to widen to about 25% by 2030; third, the company’s AI-related revenue mix is rising rapidly, and its product portfolio is tilting toward high-margin MLCCs and advanced substrates, which should bring earnings quality and valuation levels above those of a traditional cyclical recovery. Near-term optical business is weighed down by weak consumption, but this is insufficient to change the medium- to long-term AI-driven uptrend.

Analysis framework

The report combines supply chain channel checks, observations of customer lead times and book-to-bill ratios, bottom-up MLCC and ABF supply-demand models, company management guidance, and 2Q 2026 actual results to update earnings forecasts. It uses a residual income model to derive the target price and sets bull, base, and bear scenarios to assess risk-reward.

Methodology notes

• Valuation frameworkResidual income valuation model

  Estimate equity value based on cost of equity, terminal growth rate, and residual income over the forecast period.

  The base case uses a 10% cost of equity, 1.0 beta, 6.5% equity risk premium, 3.5% risk-free rate, and 5% terminal growth rate, with 2026 as the base year and a forecast period of 2026–2036.

• Financial forecasting frameworkMorgan Stanley ModelWare

  Uniformly build forecasts for revenue, profit, cash flow, returns, and valuation multiples.

  The model incorporates 2Q 2026 actual results, the latest management guidance, and the industry supply-demand outlook, and compares them with Refinitiv Estimates consensus expectations.

• Industry supply-demand analysisBottom-up ABF supply-demand model

  Infer the ABF substrate supply-demand gap by end application, capacity addition cycle, and product mix.

  The model estimates an ABF substrate supply shortfall of about 25% in 2030 and judges that new capacity will require at least two years to come online, so pricing in 2026–2028 will be mainly determined by changes in end demand.

• Growth stage analysisS-curve analysis

  After AI penetration crosses an inflection point, deployment volume, value per unit, and profitability will accelerate.

  The company’s AI-related revenue is expected to rise from 20% in 2026 to 31% in 2027, while increased content of multi-layer core ABF, MLCCs, and embedded devices will drive continued margin improvement.

• Scenario analysisBull, base, and bear scenario valuation

  Assess potential upside and downside risk through different growth, competition, and profitability assumptions.

  The bull-case value is ₩3,000,000, the base-case target price is ₩2,620,000, and the bear-case value is ₩1,350,000, corresponding to approximately 15x, 13x, and 7x 2027E P/B, respectively.

Asset mapping & comparison

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

• Samsung Electro-Mechanics (009150.KS)

  The report’s core recommended stock, directly benefiting from demand growth for high-end MLCCs and ABF substrates driven by AI servers, ASICs, and high-speed networking.

  Strengths

  Leading position in the high-end AI MLCC market; capabilities in both MLCCs and package substrates; increasing ASIC and networking customer orders; product mix migrating toward high-margin AI applications; embedded devices, silicon capacitors, and glass substrates provide subsequent product cycles.

  Weaknesses

  The optical business remains affected by smartphone and consumer electronics demand; capacity expansion capex is high; near-term free cash flow may be pressured by capacity investment.

  Comparison

  In the high-end AI MLCC market, the company and Murata form a Japan-Korea duopoly, and the company is clearly the second-largest supplier in the overall MLCC market; compared with commodity-type Chinese manufacturers, its advantages mainly come from process and customer certification barriers in high-capacitance, low-equivalent-series-inductance products.

  Risks

  MLCC and ABF price increases falling short of expectations, a slowdown in AI capex, Chinese localization competition, weak smartphone demand, poor execution in customer qualification, and high-valuation volatility.

Key data

• Stock ratingOverweightRating reiterated and newly added as a top pick.
• Target price₩2,620,000Raised from ₩2,560,000, implying 113% potential upside.
• Reference share price₩1,229,000Closing price on August 6, 2026.
• AI-related revenue mix20% in 2026; 31% in 2027AI-related revenue mix is expected to more than double to 31% in 2027.
• High-end AI MLCC market shareAbout 85%Combined share of Murata and Samsung Electro-Mechanics.
• High-end MLCC lead timeMore than 20 weeksTightening supply, canceled discounts, and rising book-to-bill ratios jointly support price increases.
• ABF supply gapAbout 25% in 2030Higher than the previous forecast of about 22%.
• ABF price forecastUp 20%–25% YoY in 2026; up more than 25% in 2027Prices are expected to rise further by 25%–40% in 2028, depending on the customer.
• 2027E EPS₩56,728The report says its new forecast is about 40% above market consensus.
• Earnings forecast revisions2026 cut by 5%; 2027 and 2028 each raised by 2%Weak optical business weighs on 2026, but AI server substrates and high-speed networking support upward revisions in subsequent years.
• Revenue forecast₩14,875 billion in 2026; ₩23,316 billion in 2027; ₩32,325 billion in 2028Corresponding revenue growth of about 31.5%, 56.7%, and 38.6%.
• Operating margin forecast15.7% in 2026; 24.5% in 2027; 25.9% in 2028AI product mix, price increases, and product mix improvement drive margin expansion.
• Scenario valuesBull ₩3,000,000; base ₩2,620,000; bear ₩1,350,000The bear case incorporates weak consumption, intensified competition, and pressure on the optical business.

Impact & implications

If MLCC price increases spread to direct sales customers and ABF long-term supply agreements continue to increase, market expectations for the company’s 2027–2028 earnings may be revised up further. AI ASICs, high-speed networking, and embedded passive devices will increase value per unit and reduce the company’s dependence on Samsung Electronics and the smartphone cycle. Because new capacity is being released slowly, pricing and margin elasticity over the next two years may be stronger than in a traditional electronic components cycle, supporting a re-rating from a cyclical stock to a structural growth stock; however, the target price implies a high valuation, making the share price more sensitive to orders, price increases, and earnings delivery.

Risks

• Capital expenditure by AI server and hyperscale cloud customers falls below expectations, slowing demand growth for MLCCs and ABF.
• Earlier-than-expected release of new capacity, a rebound in customer inventories, or a reversal of distributor stockpiling could weaken MLCC shortages and the sustainability of price increases.
• Chinese commodity-type suppliers accelerate certification for local servers, which could compress share and pricing of low- to mid-end products over the long term.
• Continued weakness in smartphone and other consumer electronics demand could further drag on the optical and traditional components businesses.
• The company faces certification and mass production execution risks in expanding Chinese smartphone customers, North American cloud service customers, and new ASIC projects.
• If price increases for raw materials such as ABF film and copper-clad laminates cannot be fully passed through, margins may be squeezed.
• Large-scale capacity expansion, the new Vietnam plant, and new products such as glass substrates face capex, yield, and production delay risks.
• The target valuation is relatively high, and if earnings forecasts or price increases fail to materialize, share price volatility and valuation compression risks are significant.
• Morgan Stanley has shareholding or business relationships with some covered companies, and investors should use this report as one factor in decision-making rather than the sole basis.

What to watch

• Actual MLCC price increases for direct sales customers and contract price execution in the second half of 2026.
• Whether high-end MLCC lead times, book-to-bill ratios, channel inventories, and extreme spot prices return to real end demand.
• Progress on long-term supply agreements with major AI and data center semiconductor customers and customer investment support.
• Mass production ramp-up of FC-BGA products for AI data center networking platforms at large North American technology customers.
• ABF substrate price increases, supply-demand gap, and timing of new capacity additions in 2027–2028.
• Whether the AI-related revenue mix can reach 31% in 2027, and the extent of margin improvement in MLCCs and package solutions.
• Whether 2027 and 2028 EPS forecasts continue to be revised upward, and whether the gap versus market consensus narrows.
• The impact of the optical business, flagship smartphone cycle, and Chinese consumption policies on 2026 results.
• Execution progress of the Vietnam plant, embedded MLCCs, silicon capacitors, and the 2028 glass substrate product cycle.