Samsung Electro-Mechanics shares set a record high for the second straight day. Samsung Electro-Mechanics closed at 2.2 million won on the 18th, up 8.27% from the prior transaction day, and the rally continued on the 19th, with the intraday price climbing to 2,417,000 won.
Compared with the start of the year, the share price is up about 760%. Demand for artificial intelligence (AI) servers is turning not only high-bandwidth memory (HBM) but also flip-chip ball grid array (FC-BGA), a semiconductor substrate, and multilayer ceramic capacitors (MLCC) into bottleneck parts, which is cited as the driver pushing the stock higher.
However, analysts say that for Samsung Electro-Mechanics' share-price gains to continue, the market needs to watch external variables as well as the company's own competitiveness. That is because the pricing and capacity expansion strategy of Murata Manufacturing (Murata), the global No. 1 MLCC corporations, could affect Samsung Electro-Mechanics' selling prices, long-term agreements (LTA), and market-share defense strategy.
◇ Target price of 3 million won appears… the premise is "price hikes"
As the AI market expands, demand is rising together for MLCCs and semiconductor substrates, the main items of Samsung Electro-Mechanics. KB Securities on the 19th raised its target price for Samsung Electro-Mechanics by 36% to 3 million won from 2.2 million won. The revision reflects the view that MLCCs and packaging substrates have entered a super-boom and that shortages could persist for at least two years.
According to FnGuide, the consensus for Samsung Electro-Mechanics' consolidated third-quarter results this year is revenue of 3.457 trillion won and operating profit of 470 billion won. Year over year, revenue would rise 19.7% and operating profit 80.6%. The full-year operating profit consensus is 1.5895 trillion won, and securities firms commonly analyze that this could be a 74.0% increase from a year earlier.
However, these figures can only materialize if several assumptions hold. ▲ A shortage of MLCCs for AI servers must actually lead to selling price hikes, ▲ customers must lock in volumes in advance through LTAs, and ▲ Samsung Electro-Mechanics must avoid losing share to Murata while increasing the share of high-value products.
Conversely, if Murata quickly adjusts its pricing and expansion strategy, if customers resist price hikes, or if Samsung Electro-Mechanics mis-sequences price hikes and capacity additions, the earnings outlook could change.
◇ A crowded MLCC market with many rivals… big variables also for AI server use
The demand surge and shortages driven by AI expansion are appearing in FC-BGA and MLCCs following memory semiconductors.
Still, it is necessary to look more closely at how long the factors driving Samsung Electro-Mechanics' share gains will last and how they will actually translate into earnings. While HBM and other DRAM are in a structure effectively dominated by Samsung Electronics, SK hynix, and Micron, there are more players in MLCCs.
According to market researcher Dataintelo, the 2025 global MLCC market shares by revenue are estimated at Murata 28%–30%, Samsung Electro-Mechanics 22%–24%, TDK 11%–13%, Taiyo Yuden 8%–10%, and Yageo 7%–9%.
The securities industry estimates that in high-spec MLCCs mounted in AI servers, Murata (45%) and Samsung Electro-Mechanics (40%) effectively split the market. The two hold the lead in AI server MLCCs. However, the more the two concentrate production on high-spec products, the more room there is for other companies to seize opportunities in commodity products. Samsung Electro-Mechanics is not in a structure where it can focus production only on AI server products with higher profitability.
◇ Samsung Electro-Mechanics can raise selling prices only if Murata moves first
Samsung Electro-Mechanics raised selling prices preemptively during the 2017–2018 MLCC boom. As a result, Samsung Electro-Mechanics ceded some customers to other corporations and suffered a drop in market share, which took three to five years to recover. Analysts say that in this boom, Samsung Electro-Mechanics is more likely to respond after Murata moves first and customer reactions are confirmed.
In effect, Murata's strategy will determine the magnitude of Samsung Electro-Mechanics' earnings growth. According to the industry and the securities sector, Murata recently notified customers of MLCC price hikes of 15%–35%. In its late-April results announcement, Murata also guided that server-related capacitor sales in fiscal year 2026 would rise 85%–90% year over year and that the average selling price (ASP) for capacitors overall would increase 5%–10%.
iM Securities said that if MLCC selling prices rise 10%, Samsung Electro-Mechanics' 2027 operating profit could expand to about 2.5 trillion won. If prices rise 20%, operating profit could grow to 3.1 trillion won. Conversely, if price hikes face customer resistance or Samsung Electro-Mechanics loses some volume, earnings estimates could fall.
◇ Shares rising without capacity additions… the moment of an announcement could become a risk
MLCC shortages are already appearing. Market researcher TrendForce analyzed that lead times for some high-capacity X6S MLCCs have lengthened from eight weeks to as long as 20 weeks. That means it could take about five months from order to delivery.
The current rise in MLCC-related stocks is also due to expectations of price hikes because capacity additions have not been sufficient. Demand for high-spec MLCCs for AI servers is growing rapidly, but suppliers are not formalizing large-scale capacity expansions in the short term. When demand is visible but supply growth is limited, prices rise, fueling expectations of margin improvement.
Murata is increasing investment in capacitors to meet data center demand. To capture expanding data center demand, it has also identified small, high-capacity capacitors and power modules as key growth tasks. However, it will take time for capacity additions to translate into actual increases in high-spec MLCC supply. That is because MLCCs for AI servers have longer manufacturing lead times and are harder to secure production yield than general MLCCs.
Samsung Electro-Mechanics also needs to add capacity. In its first-quarter results this year, Samsung Electro-Mechanics said AI-related demand is growing faster than previously expected, so investment could more than double from a year earlier and investment over the next three years could be much larger than in the past. iM Securities estimates Samsung Electro-Mechanics' capital expenditures this year at more than 1.8 trillion won. However, there has yet to be a large-scale official announcement on the timing for starting up new AI server MLCC lines, which the market watches most closely, or on the specific scale of production capacity expansion.
An industry official said, "As long as shortages persist, Samsung Electro-Mechanics and Murata can hold the upper hand as suppliers," but added, "The moment large-scale expansion plans are formalized, the market will bring forward the timing for the end of shortages."
If Samsung Electro-Mechanics formalizes expansion quickly, it could undercut its own bottleneck premium. Conversely, if expansion is late, it could lose customer volumes to Murata. The industry sees a strong likelihood that Samsung Electro-Mechanics will adjust its MLCC supply strategy while watching Murata's investment pace, the intensity of its price hikes, and customers' LTA demands.
◇ AI servers and in-house ASICs spread… MLCC demand "surging"
Despite various variables, one thing is clear: MLCC demand is exploding. MLCCs store electricity and supply it stably to various electronic devices while reducing voltage fluctuations and noise. They are used in most electronic devices such as smartphones, PCs, and cars, and are called "the rice of the electronics industry."
MLCCs are essential for AI semiconductors such as central processing units (CPU), graphics processing units (GPU), and application-specific integrated circuits (ASIC) that use electricity. AI servers consume far more power than general servers, and sudden power fluctuations around GPUs or ASICs are much larger. If power delivery wavers, AI chip performance can drop or errors can occur. That is why MLCCs that are smaller, store more electricity, and withstand higher temperatures in tight spaces are needed.
MLCCs for AI servers are more technically challenging than mobile and IT products but are also more expensive, making them key products for securing profitability. While about 800–1,200 MLCCs are mounted on a smartphone circuit board, AI servers reportedly use around 20,000–30,000. As AI chip performance rises, even more MLCCs may be required. Nvidia's next-generation AI server platform, the Vera Rubin-series NVL72 server, is estimated to mount about 600,000 MLCCs, more than 30% higher than the previous Blackwell-series GB300.
Another pillar spurring recent MLCC demand is the expansion of cloud service providers' (CSP) in-house ASICs. Big Tech companies such as Google, Amazon Web Services (AWS), Meta, and Microsoft (MS) are focusing on developing their own AI accelerators to reduce dependence on Nvidia GPUs and boost profitability.
As CSPs increase their in-house ASICs, MLCC demand naturally rises as well. ASICs, like GPUs, consume a lot of power and require precise control of power fluctuations. TrendForce analyzed that the spread of in-house ASICs by global CSPs is concentrating high-spec MLCC demand on certain premium specifications. Demand is converging on high-value products such as 47µF 2.5V X6S 0402 and the 100µF X6S series.
Rising MLCC usage is also being confirmed in next-generation AI accelerator platforms. On the AMD MI450 platform, usage of 47µF 2.5V X6S 0402 MLCCs increased from 1,440 per board to 10,544 during validation, a 632% jump. On Nvidia's Vera Rubin platform, usage of 100µF 4V X6S 0805 MLCCs reportedly rose from 320 to 500 per board.
In the second half of this year, major in-house ASIC platforms such as Google's TPU V8t/i, AWS's Trainium 4, and Meta's MTIA 400 and 450 are expected to ramp up mass production. As these products enter full-scale production expansion, MLCC demand is positioned to reach new highs.
◇ Samsung Electro-Mechanics' differentiation cards are MLCCs, FC-BGA, and silicon capacitors
Murata has strengths in ceramic materials, powders, thin-film stacking, and miniaturization technologies. It covers passive components including MLCCs, communications modules, high-frequency components, and power modules. Its strength is that it can propose not only MLCCs but an entire suite of power stabilization components to AI data center customers.
By contrast, Samsung Electro-Mechanics has expanded its position in AI server MLCCs through high-capacity products and responses to server and automotive customers. In addition, it has FC-BGA, a package substrate for AI Semiconductor, and silicon capacitors, which is cited as a differentiator enabling it to propose comprehensive AI semiconductor power stabilization solutions.
NH Investment & Securities analyzed that, unlike other MLCC makers, Samsung Electro-Mechanics has a packages division, highlighting differentiated competitiveness through synergies. The point is that it can propose a power stabilization solution that considers both substrates and MLCCs, rather than supplying only MLCCs. Samsung Electro-Mechanics recently signed a 1.5 trillion won supply contract for silicon capacitors, components mounted inside high-performance semiconductor packages such as AI server GPUs and HBM to improve power delivery stability.
A researcher at a market research firm said, "Murata has formalized data center capacitor demand and ASP increases, and how much Samsung Electro-Mechanics can capture in price and volume within that trend will be the factor that determines the future share price."