High-end MLCCs for AI servers support Murata, while Ibiden earnings expectations may be too high

Overview

This report is Morgan Stanley's investor visit summary for the Japanese electronic components industry. Based on discussions with institutional investors in Singapore and Hong Kong, it focuses on MLCCs, high-value-added components for AI servers, pricing of standard MLCCs, and Ibiden's technological competitiveness and valuation in ABF package substrates. The overall industry view is In-Line, but the report clearly favors Murata Manufacturing among individual stocks while remaining cautious on Ibiden.

Core views

There are three core views. First, Murata's MLCC growth does not mainly rely on price hikes for the same products, but rather on rising demand from AI servers for miniaturized, high-capacitance, high-value-added MLCCs and on product mix upgrades. Second, over the medium to long term, the gap in corporate value will widen between companies that can continuously enhance product competitiveness and those that simply rely on price increases for standard products. Third, although Ibiden benefits from NVIDIA Rubin-related ABF package substrates and subsequent EMIB-T opportunities, Morgan Stanley believes that both market consensus and the company's targets assume overly optimistic earnings growth.

Analysis framework

The report uses methods such as investor visit feedback, comparison of product technology competitiveness, product mix and ASP breakdowns, market share estimates, and comparisons between company earnings forecasts and market consensus to assess the degree to which Japanese electronic component companies benefit from the AI server supply chain and the associated valuation risks.

Methodology notes

• Investor feedbackInstitutional investor visit summary

  Identify market focus points through 31 one-on-one meetings and institutional lunch meetings.

  Investors mainly focused on the magnitude of price increases for standard MLCCs, but the report argues that the more important earnings variable is the product mix upgrade of high-value-added MLCCs for AI servers.

• Industry comparisonTechnology competitiveness and product mix analysis

  Compare the mass-production capabilities and market shares of manufacturers such as Murata, SEMCO, and Taiyo Yuden in high-end MLCCs.

  The report emphasizes that Murata has advantages in finer material particles, higher layer counts, narrower external electrode widths, and stable mass production, and therefore is better positioned to benefit from AI server demand for high-capacitance products.

• Earnings forecastComparison of company forecasts, FactSet consensus, and Morgan Stanley forecasts

  Compare Ibiden's future OP expectations under different forecasting bases.

  Morgan Stanley's OP forecasts for Ibiden in F3/28 and F3/31 are below the company's targets and market consensus, and it therefore maintains a cautious rating.

• Rating frameworkMorgan Stanley relative rating system

  OW, Equal-weight, NR, and UW represent expected total returns relative to the industry coverage universe.

  The report discloses that Morgan Stanley ratings typically measure risk-adjusted total return relative to the industry coverage universe over the next 12-18 months.

Asset mapping & comparison

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

• Murata Manufacturing

  Core beneficiary of high-value-added MLCCs for AI servers; rating maintained at OW.

  Strengths

  It has a leading MLCC market share and advantages in barium titanate particle refinement, higher layer counts, narrower external electrodes, miniaturization, and stable mass production of high-capacitance products.

  Weaknesses

  Its pricing strategy for standard MLCCs is relatively cautious and it may not pursue short-term profit maximization; earnings growth depends on continued enhancement of the high-value-added product mix.

  Comparison

  The report estimates its 2025 MLCC share at 40.8%, higher than SEMCO's 22.5% and Taiyo Yuden's 11.3%; the report believes Murata is the only company capable of stably mass-producing multiple key MLCCs for AI servers.

  Risks

  If demand for high-end MLCCs for AI servers, volume growth, or ASP mix improvement falls short of expectations, upside to earnings may narrow; the pricing environment for standard products could also alter the competitive landscape.

• Taiyo Yuden

  A beneficiary of AI/data center MLCCs, but earnings contribution is expected to be limited.

  Strengths

  Sales of AI/data center MLCCs are expected to grow rapidly, and improved capacity utilization should support earnings.

  Weaknesses

  The timing for stable mass production of all key MLCC products for AI servers remains unclear, and after freezing capex in 2H 2024, F3/26 capacity growth is estimated at only about 5%.

  Comparison

  Its 2025 MLCC market share is estimated at 11.3%, below Murata and SEMCO; even if it catches up on existing products, Murata may already be in the mass-production stage of smaller and higher-capacitance products.

  Risks

  Capacity and mass-production capability constraints may prevent it from fully capturing incremental orders, and the contribution of high-value-added products to earnings may fall short of market expectations.

• Ibiden

  A growth name in ABF package substrates, but rating maintained at UW.

  Strengths

  The report expects that Ibiden began mass shipments of ABF package substrates for NVIDIA Rubin in the fourth quarter of F3/26, and Rubin-related sales may exceed Blackwell in the first quarter of F3/27.

  Weaknesses

  Morgan Stanley believes market expectations are too high, that the substantive earnings contribution from EMIB-T is more likely to emerge only after F3/29, and that margins may not necessarily reach the levels of current NVIDIA-related products.

  Comparison

  Morgan Stanley's OP forecasts for F3/28 and F3/31 are below the company's targets and FactSet consensus, showing that it is more cautious than the market.

  Risks

  If the move toward higher value-added ABF, larger SAP sizes, or the mass-production pace and margins of Rubin/EMIB-T fall short of expectations, valuation and earnings expectations may face downward revisions.

• SEMCO

  A major MLCC competitor and the second-largest player by market share.

  Strengths

  The report estimates its 2025 MLCC market share at 22.5%, ranking it near the top of the global MLCC market.

  Weaknesses

  The main text does not provide details of its stable mass-production capability for high-value-added MLCCs used in AI servers.

  Comparison

  Its market share is below Murata's but above Taiyo Yuden's; the report still identifies Murata as the leading supplier of high-end MLCCs for AI servers.

  Risks

  If it cannot continue catching up in high-capacitance, miniaturized MLCCs for AI servers, it may struggle to share in the growth of the highest value-added segment.

Key data

• Number of investor meetings31 one-on-one meetings, plus institutional investor lunch meetings in Singapore and Hong KongThe meetings took place from June 22 to 26, 2026, and mainly focused on MLCCs and ABF package substrates.
• Industry viewIn-LineThe view on the Japanese electronic components industry is In-Line.
• 2025 MLCC market share estimateMurata Manufacturing 40.8%, SEMCO 22.5%, Taiyo Yuden 11.3%Based on this, the report judges that Murata holds a leading position in high-value-added MLCCs for AI servers.
• Murata AI/data center MLCC mix10%-15% in F3/26, expected to rise to 20%-25% in F3/27Related sales in F3/27 are expected to grow 85%-90% YoY.
• Murata F3/27 growth breakdownVolume growth of about 40%, ASP growth of about 50%ASP growth mainly comes from a higher mix of high-priced, high-value-added products rather than unit price increases for the same products.
• Taiyo Yuden AI/data center MLCC mix5%-10% in F3/26, expected to be about 15% in F3/27Related sales in F3/27 are expected to grow 82%-83% YoY, but mass-production capabilities and capacity expansion may limit the contribution.
• Ibiden OP forecast¥94.7bn in F3/27, ¥128.5bn in F3/28, ¥242.6bn in F3/31The F3/28 and F3/31 forecasts are below the company's targets and FactSet consensus, reflecting Morgan Stanley's view that market expectations are too high.

Impact & implications

The investment implication is that AI server upgrades are increasing the total MLCC capacitance required per accelerator board and driving demand for miniaturized, high-capacitance products; companies with stable mass-production capability and continuous product iteration are more likely to achieve medium- to long-term corporate value enhancement. By contrast, although price increases for standard products may boost profits in the short term, they may also lower barriers to entry, attract competitors from China, Korea, and Taiwan, and weaken medium- to long-term market share.

Risks

• If the increase in total MLCC capacitance demand driven by upgrades in AI server GPU generations is lower than expected, the growth thesis for high-end MLCCs may weaken.
• Murata's ASP improvement mainly comes from product mix upgrades; if the share of high-priced products rises less than expected, earnings leverage will decline.
• Price increases for standard MLCCs may lift short-term profits, but they may also lower barriers to entry and trigger accelerated entry by competitors from China, Korea, and Taiwan.
• Taiyo Yuden faces uncertainty in stable mass production and capacity expansion for key MLCC products.
• There is uncertainty around the ramp-up pace and margins of Ibiden's Rubin and EMIB-T products, and long-term earnings targets from both the market and the company may be overly optimistic.
• The report's charts mention OP sensitivity to foreign exchange for covered electronic component companies, so exchange rate changes may affect earnings performance.

What to watch

• The extent of the increase in total MLCC capacitance demand per accelerator board for each new GPU generation of AI servers.
• Murata's sales mix, volume growth, and ASP mix changes in MLCCs for AI/data centers.
• Stable mass-production capability for high-end MLCCs such as 1608 size 100µF, 1005 size 47µF, and 0603 size 10µF.
• Whether Taiyo Yuden resumes or expands capex, and whether it can catch up with Murata's mass-production pace for high-end products.
• When Ibiden's Rubin-related ABF sales will exceed Blackwell, and the scale and margins of EMIB-T's earnings contribution after F3/29.
• Channel price changes for standard MLCCs and the competitive strategies of manufacturers from China, Korea, and Taiwan.