三星电机电话会称AI服务器MLCC、FC-BGA需求继续收紧并推进长协

[Four Key Questions from Samsung Electro-Mechanics’ Conference Call]

Q1) (Company-wide) What is the outlook for Q3 earnings?

A1) In Q2, the company delivered year-on-year and quarter-on-quarter growth, driven by increased sales of high-value-added MLCCs and FC-BGAs for AI servers, networking, and automotive applications, as well as an improved product mix resulting from a higher proportion of premium products.

In Q3, robust demand for related components is expected to continue, supported by sustained investment in AI data centers and advances in autonomous driving. Supply-demand conditions for MLCCs and FC-BGAs are also expected to tighten further.

For MLCCs, the company plans to actively expand the supply of high-value-added, premium-priced products for AI servers, networking, and ADAS applications, including ultra-high-capacitance, high-temperature, and high-voltage products. It will continue discussions on strategic long-term supply agreements with major hyperscalers and semiconductor companies, while also responding proactively on pricing in light of supply-demand conditions.

For FC-BGAs, the company plans to actively expand mass production of new products for AI accelerators and server CPUs supplied to global big tech companies, as well as the supply of high-end products for data centers. It will also continue price negotiations reflecting market conditions. Domestic and overseas capacity expansions tied to LTA discussions with top-tier semiconductor customers will proceed as planned.

For camera modules, the company plans to increase the supply of high-value-added automotive products designed for increasingly advanced ADAS applications and ramp up mass production of new products for humanoid robots.

In Q3, the continued tightening of supply-demand conditions for high-value-added products such as MLCCs and FC-BGAs, the expansion of LTAs, and higher average selling prices are expected to drive record-high earnings.

These trends are expected to continue into Q4 and strengthen further in 2027, supporting sustained earnings improvement.

The company will focus its capabilities on high-growth markets such as AI data centers and automotive applications, while responding quickly to changes in market conditions and supply chains through measures including proactive capacity expansion.

In new business areas such as silicon capacitors and glass substrates, the company will further strengthen collaboration with global big tech customers and proceed with its preparations for commercialization over the medium to long term.

Q2) (MLCC) What is the potential for additional supply agreements for AI-related MLCCs, and what is the company’s future capacity strategy?

A2) The AI server MLCC market is expected to maintain rapid growth as the continued expansion of hyperscaler data-center investment, improvements in GPU computing performance, and faster server-network speeds drive increasing adoption of cutting-edge, high-end MLCCs.

Cutting-edge products, including ultra-high-capacitance products of 100㎌ or more and high-temperature products rated at 125°C, require the highest levels of technology in terms of both performance and reliability. Only a small number of suppliers, including the company, are capable of meeting these requirements. As a result, supply-demand conditions are becoming increasingly tight, while demand for long-term supply agreements to secure stable supply volumes is rising among major AI server customers.

The company has completed LTAs with around 10 customers, including top-tier hyperscalers and major semiconductor companies, and is actively responding to additional LTA requests from key customers.

To strengthen its leadership in the AI server MLCC market, the company is improving yields and productivity while expanding Capex compared with previous years to establish a production base capable of maximizing supply.

As technical requirements for AI server MLCCs advance rapidly, the company will leverage its ultra-high-capacitance product portfolio to launch cutting-edge specifications across different sizes, voltages, temperatures, and operating conditions in a timely manner, enabling it to respond quickly to customers’ diverse requirements.

The company will reinforce its technological leadership in the AI server MLCC market while continuing to deliver revenue growth that outpaces the market.

Q3) (Silicon Capacitors) The company disclosed a KRW 1.5 trillion silicon capacitor supply agreement in May. What is the market outlook, and is there potential for additional orders?

A3) Silicon capacitors offer superior electrical characteristics compared with MLCCs, including low equivalent series inductance, or low ESL, and are also thinner. They are therefore key components placed in the closest proximity to semiconductor chips within AI server semiconductor packages, supporting overall system performance and stability.

Their applications are expanding from mobile devices to AI data centers, and the market is expected to continue growing as the adoption of silicon capacitors in AI data centers increases.

Based on the complementary characteristics of MLCCs and silicon capacitors, the company is currently discussing additional supply agreements with major AI-related semiconductor companies.

Leveraging its position as the only company worldwide capable of providing MLCCs, silicon capacitors, and package substrates on a turnkey basis, the company intends to take the lead in securing the emerging market and drive substantial revenue growth.

Q4) (FC-BGA) The company is reportedly considering capacity expansion. What are the detailed expansion and future capacity-management plans?

A4) As the AI paradigm shifts from training toward inference, global big tech companies are increasingly developing proprietary semiconductor chips optimized for inference, with the number of participating companies also rising.

As AI semiconductors become more powerful, larger chip sizes and increased HBM adoption are accelerating the transition toward ultra-large and highly multilayered package substrates, further consuming suppliers’ available capacity.

Recently, major customers have also increased their requirements for high-end substrates incorporating features such as embedded components and multi-core designs. Only a small number of suppliers with the necessary mass-production capabilities and technological expertise can meet these requirements, further exacerbating the shortage of high-value-added substrates.

The company is discussing strategic LTAs, including investment support, with most top-tier semiconductor customers involved in AI and data-center applications. Its capacity expansion plans will be based on these agreements.

The company will strengthen collaboration with customers on next-generation product development and concentrate its capabilities on securing differentiated technologies. At the same time, it will proceed with capacity expansion as planned to meet customer demand in a timely manner and expand the business over the medium to long term.