MAIN ARTICLE
Record sales come with a stronger operating margin
Samsung Electro-Mechanics reported approximately 3.76 trillion won in revenue and 691 billion won in operating profit for July to September 2026, in results announced on October 28. The Samsung affiliate makes electronic components; it should be distinguished from Samsung Electronics when discussing chips and AI. Revenue rose 30% from a year earlier, while operating profit increased 165%. Both measures reached quarterly records, placing the company among the beneficiaries of demand for AI infrastructure components.
Against the previous quarter, revenue increased 9% and operating profit rose 57%. The operating margin consequently climbed to 18.4%, from 12.7%, an improvement of 5.7 percentage points. This is a measure of earnings relative to revenue at the operating level. It does not mean that the same proportion of sales has already arrived as cash, or that it represents net profit after every other financial item. Those distinctions matter when assessing the strength of a result.
A 165% increase means operating profit was roughly 2.65 times its year-earlier level, rather than 165 times as large. Faster profit growth than sales is encouraging, but aggregate figures alone cannot establish how much came from product mix, prices or costs. A sensible interpretation recognises the improvement while leaving that detailed explanation open. Nor does one quarter establish a permanent margin: it creates a reference point against which subsequent delivery and profitability can be assessed.
The fastest-growing division is not the largest
Package solutions revenue reached 939 billion won, up 58% from a year earlier. Sequential growth accelerated to 22%, compared with 6% in the second quarter. According to The Korea Herald, the division accounted for more than half of the company’s revenue increase from the preceding quarter. Higher-value substrates for server CPUs, AI accelerators and networking equipment supported the advance, giving package solutions a prominent role in explaining the overall expansion.
A package substrate connects a processor to the circuit board and carries electrical signals and power. It is not the processor doing the calculations. That distinction reveals a broader industrial relationship: designing a powerful chip and making it useful within a complete system require complementary capabilities. Samsung Electro-Mechanics’ substrate business participates in that supporting layer. Its growth illustrates why the commercial effects of AI investment extend beyond the companies whose processors receive the most public attention.
At roughly a quarter of company revenue, package solutions is a major growth engine without being the biggest sales contributor. Growth rate and business size answer different questions. It would also be incorrect to treat the division’s 58% revenue increase as its profit growth rate, because the figures described do not provide that matching profit breakdown. Keeping these measures separate makes the company’s change in business balance easier to understand without assigning earnings that have not been disclosed here.
Small capacitors carry a substantial business
The components division remained the largest business, with approximately 1.88 trillion won in quarterly revenue, up 36% from a year earlier. Demand for multilayer ceramic capacitors, or MLCCs, used in AI servers and power systems was an important driver.
The division represented about half of total sales. That scale shows why a reading focused only on the faster-growing substrate operation would leave out a large part of the company’s performance.
MLCCs store electrical charge and help stabilise power supply. Samsung Electro-Mechanics expects more powerful AI processors to require higher capacitance, increasing demand for advanced products. The basic relationship is accessible even without specialist knowledge: fast computation needs suitable electrical support. Capacitors serve a different function from data processing, and using one in an AI server does not turn it into an AI processor. Explaining that difference makes the supply chain more understandable without overstating any single component’s role.
The company plans to expand advanced MLCC supplies through long-term contracts with major technology customers. Such arrangements can support production planning, but product quality and delivery still have to be achieved in manufacturing. The reported company-wide totals do not reveal precisely how much revenue is attributable to AI alone or how much profit comes from an individual customer. The direction of demand is clear in the company’s account; its exact contribution requires a more detailed breakdown.
Expansion in Korea and Vietnam is a multiyear commitment
In September, Samsung Electro-Mechanics announced approximately 6.8 trillion won of substrate manufacturing expansion in South Korea and Vietnam. Its investment announcement places the Sejong programme through May 2028 and targets completion and the start of operations in Vietnam by June 2030. These are planned milestones, not evidence that all additional capacity is already operating. The investment should therefore be read on a different timetable from the earnings generated during the three months just reported.
Last week, the company also announced a long-term supply partnership with Nvidia for advanced substrates used in AI processors and data-centre networking equipment. That relationship clarifies the direction of customer demand, but it does not make all third-quarter revenue a result of a newly announced agreement. Investment spending likewise cannot be counted as sales already booked or profit guaranteed to follow. The connection between a customer relationship and future earnings must pass through actual production and deliveries.
For Vietnam, the relevant question is what manufacturing capability emerges, alongside the skills and quality requirements needed to sustain it. A large project creates an opportunity to participate in a demanding supply chain, but its announced value does not establish a number of new jobs or the share of business reaching local suppliers. Those outcomes need to be assessed through implementation and actual connections. Their importance is precisely why they should not be inferred automatically from the capital commitment.
AI expectations coexist with other business rhythms
Not every division moved in the same direction over the shorter comparison period. Optical solutions revenue was 941.3 billion won, up 3% from a year earlier but down 9% from the previous quarter. Both statements can be true because they use different starting points. The result is a reminder that a company benefiting from AI demand can still contain products with different demand patterns, and that a strong overall quarter does not erase variation within the business.
Samsung Electro-Mechanics expects demand for AI data-centre components to remain strong in the fourth quarter and plans to increase advanced substrate shipments for server CPUs and AI accelerators. That is management’s outlook, to be tested against subsequent results. Useful measures include divisional sales, operating margins and expansion milestones. Strong demand must still be converted through manufacturing, delivery and cost control. Following those stages provides a firmer basis for judging durable progress than simply extending the latest growth rate into the future.














