How Does Samsung Electro-Mechanics Diversify Its Exposure to AI Server Components?

First-quarter revenue of KRW 3.2091 trillion came from three divisions. The May 15, 2026 quarterly report shows first-quarter consolidated revenue of KRW 3.2091 trillion, up 17.2% year over year. Component generated KRW 1.4085 trillion, Optical Solution KRW 1.0756 trillion, and Package Solution KRW 725.0 billion. MLCCs belong to Component, FC-BGA to Package Solution, and camera modules to Optical Solution.

The next quarter should show whether the planned start of AI data-center network-substrate supply to a new big-tech customer appears in Package Solution revenue and utilization. Customer volume and identity should remain within the company's disclosed scope. Strong demand for one AI component does not make the company a pure-play exposure because all three divisions continue to shape consolidated results.

First-quarter Package Solution sales grew across applications. Official first-quarter 2026 investor relations materials reported Package Solution revenue of KRW 725 billion, up 45% year over year and 12% from the previous quarter. The company cited increased supply of advanced substrates for AI and server applications as a major driver. These exact comparisons establish strong growth in the division rather than in the entire company.

The KRW 725 billion figure is total Package Solution revenue, not FC-BGA revenue alone. It also includes growth in BGA products. Preserving that distinction avoids assigning every won of segment growth to one product. Future disclosures should show whether high-end FC-BGA production, broader BGA demand, or another mix effect drove the segment and how that mix affected profit.

MLCC demand follows rising power density. Higher AI-server power consumption can increase requirements for high-capacitance, high-reliability MLCCs in power circuits. Samsung Electro-Mechanics cited greater supply of MLCCs for AI servers and advanced driver-assistance systems, or ADAS, as a contributor to first-quarter improvement. This creates a second AI-related path distinct from the package substrate beneath a processor.

Revenue depends not only on server unit count but also on the number of components per machine and their specifications. Commodity MLCC pricing and the mix of premium industrial products should be examined separately. A richer mix can support margin even without identical unit growth, while utilization, yield, and input costs determine how much of the higher-value demand becomes profit.

A new big-tech supply plan requires production evidence. The company said it planned to begin full-scale supply of AI data-center network substrates for a new big-tech customer in the second quarter. This is a company-stated execution plan. Actual shipment and revenue must be confirmed in subsequent results. A planned start is stronger evidence than a general industry forecast but remains different from reported sales.

The customer's name and contract volume cannot be inferred beyond the disclosed scope. The stages are supply commencement, recurring orders, and improvement in product mix. Readers should look for evidence that qualified products shipped as scheduled, that volume continued beyond an initial batch, and that segment sales and margins reflect the contribution without attaching an undisclosed identity to the customer.

Expansion should be read with demand and yield. Growth in multilayer, large-area substrates can require capacity investment, but large spending also creates depreciation and early yield pressure. Reviewing an investment, approving it, executing it, and starting production are different stages. A full-utilization outlook for the second half remains something to verify through actual utilization, delivery, and reported results rather than a completed fact.

The next disclosure should combine Package Solution and Component revenue and operating performance with capital expenditure and cash flow. Strong demand creates value only when new equipment produces qualified output at stable yields. Following customer design, substrate development and qualification, production supply, higher utilization and revenue, and then stable yields and profit keeps the AI thesis tied to measurable operations.

Profit contribution across the three-business portfolio. Suppose KRW 10 trillion in revenue consists of KRW 4 trillion from MLCCs, KRW 2 trillion from substrates, and KRW 4 trillion from cameras, with respective margins of 20%, 10%, and 5%. Their profit contributions are KRW 0.8 trillion, KRW 0.2 trillion, and KRW 0.2 trillion. Revenue share and profit share are not the same.

If segment disclosure combines FC-BGA with other substrates, do not reverse-engineer FC-BGA's standalone contribution.

A profit-contribution table multiplying each business's revenue by its margin is the concrete output behind the idea of “spreading exposure across the portfolio.”