[TECH Weekly] Simec Recast as an AI Infrastructure Beneficiary Beyond a Traditional Materials, Parts and Equipment Company
TECHWORLD ·
✦ AI Summary
Simec makes PCB for semiconductor packaging and PCB for memory modules, and is expanding its business into SOCAMM substrates for AI servers and CXL expansion substrates.
As AI servers and HPC spread and demand for high-speed, high-capacity memory increases, shipments of next-generation high-value-added substrates are expanding, and the market is recasting Simec as a key beneficiary of AI infrastructure.
However, Simec's earnings are heavily influenced by the memory semiconductor cycle and upstream IT demand, while large-scale capital investment and fluctuations in raw materials and exchange rates are headwinds.
As digital transformation accelerates and interest in related companies broadens, reporter Joo Ga-young covered a piece that reexamines Simec from the perspective of an AI infrastructure beneficiary as part of a series sharing promising KOSPI and KOSDAQ names with individual investors in mind, alongside global macro issues. The topic of the article is Simec's reevaluation, and Simec is a global materials, parts and equipment company.
Simec operates businesses in PCB for semiconductor packaging and PCB for memory modules. The company has world-class technology and a high market share, with high-density and fine-circuit technology as its core strength. It also specializes in manufacturing substrates that maximize memory semiconductor performance.
As the AI server market and the HPC market expand, demand for high-speed, high-capacity memory is surging, and this is leading to a full-scale ramp-up in shipments of next-generation high-value-added substrates. Against the backdrop of rising memory demand and expanding shipments of next-generation substrates, Simec's technological competitiveness is drawing renewed attention and forming the basis for a shift in valuation.
Simec has a new growth engine in an 'SOCAMM substrate for AI servers.' In the past, Simec was viewed as a maker of general-purpose memory module PCB and package substrates centered on PCs and general servers. But the AI server market has surged recently.
In response, Simec has upgraded its existing module PCB manufacturing capabilities. It also succeeded in expanding its portfolio into the 'SOCAMM substrate for AI servers' segment. While Simec used to be seen largely as a general-purpose memory substrate company, it broadened its business focus in line with AI server market growth.
In the process, Simec expanded dedicated plants in places such as Ochang. It also carried out large-scale investments. As a result, Simec has moved beyond the category of a traditional materials, parts and equipment company.
As a result, the market is recasting Simec as a key beneficiary of AI infrastructure. The stock's upward momentum continues. The company's earnings momentum also continues.
Simec has two main products. Main product 1 is memory module PCB (Module PCB). Memory module PCB (Module PCB) is a substrate mounted on motherboards in PCs and servers, among others. It is a core substrate that efficiently connects multiple DRAM chips and supports capacity expansion. Simec has long maintained a global No. 1-level position in memory module PCB.
Simec operates a Package Substrate business. Package Substrate is a high-value-added substrate that electrically connects a Silicon Die and the main board and protects it from the external environment.
Simec supplies chip scale packages (CSP) for smartphones and a wide range of ball grid array (BGA) substrates for PCs, servers and network equipment. On that basis, Simec has unrivaled competitiveness in the memory semiconductor substrate market and has secured major memory semiconductor makers around the world as stable customers.
When IT set demand recovers or the high-capacity trend centered on high-performance servers accelerates, the selling prices of high-value-added substrates rise. At the same time, the company has a strong operating leverage profile, with operating margin improving sharply.
In addition, the structural growth momentum is strengthening as rising demand for high-performance package substrates from the spread of AI servers, rising demand for high-performance package substrates from the spread of HPC environments, and rising demand for CXL(Compute Express Link)-linked expansion substrates from the spread of AI server and HPC environments are all added together.
Simec is a company with a strong technological edge, but its weakness is that it is highly dependent on the memory semiconductor cycle. Since its upstream industry is memory semiconductors, Simec's results are heavily affected by the memory semiconductor cycle. If the memory semiconductor cycle recovers, Simec is often cited as a representative name showing rapid earnings elasticity.
On the other hand, if the global memory industry enters a downcycle, Simec's earnings volatility could widen. Its earnings volatility could also increase when demand for IT devices such as smartphones and PCs weakens. This means results are highly dependent on the memory cycle and upstream IT demand.
If the cycle remains sluggish, shipments may decline and selling prices may also fall. There is also the possibility that lower shipments and lower selling prices occur at the same time. In that case, the burden on earnings could become even heavier.
In addition, compared with global leading substrate makers, Simec has a low share of system semiconductors, or non-memory chips. Accordingly, in a phase where the AI semiconductor ecosystem centered on non-memory chips expands, the scope of benefits for Simec could be limited. In other words, a memory-centered structure could act as a relative constraint in a non-memory expansion phase.
When the recovery in the cycle is delayed, fixed-cost burdens could increase. Short-term earnings weakness could also act as downward pressure on the share price. Therefore, when approaching Simec, monitoring of the cyclical upturn and downturn is necessary, and monitoring of upstream IT demand volatility is also necessary.
Global top-tier substrate makers are rapidly increasing the share of substrates for non-memory semiconductors, which are key in the AI era. By contrast, Simec is maintaining a portfolio centered on memory substrates and module PCB, showing a different portfolio direction from global top-tier substrate makers. As the semiconductor ecosystem may be reorganized from a memory-centered structure to one centered on non-memory chips and AI accelerators, if that ecosystem restructuring accelerates, Simec's structural growth pace could be constrained, and valuation could also be constrained.
Simec is carrying out large-scale facility investments to preemptively respond to demand for new products such as SOCAMM for AI servers. However, large-scale facility investment is a risk factor, and after expansion there is the possibility that the recovery in upstream demand will be delayed, as well as the possibility that yield stabilization during initial mass production will be delayed. If the recovery in upstream demand or yield stabilization is delayed, substantial depreciation and fixed costs could arise, and the burden of depreciation and fixed costs could lead to weaker short-term earnings and greater downward pressure on the share price.
On the cost side, the key raw materials for substrate manufacturing are gold, copper and copper-clad laminate. Price volatility in these raw materials and exchange-rate trends directly affect cost margins. If increases in raw material prices cannot be reflected in final product selling prices in time, profitability could be hurt.
On the demand side, the upstream customers are the world's three major memory semiconductor makers. Examples of the world's three major memory semiconductor makers include Samsung Electronics, SK hynix and Micron. Their CapEx cycles and inventory policies are directly linked to results.
The sensitivity of earnings to demand fluctuations from major customers also needs to be considered. If external macroeconomic uncertainty widens, investor sentiment could deteriorate sharply because the business has a high export share.
Even so, the investment stance on Simec among South Korea's brokerage community remains broadly positive, or buy-rated. This reflects momentum in new products for AI servers. There are also cases of upward revisions to mid- to long-term target prices. Overall, the view on Simec in the domestic brokerage community is optimistic.
Simec is being evaluated as having passed its earnings bottom and entering a clear phase of profit improvement. As it reduces the share of low-margin general-purpose products and reorganizes its portfolio around high-value-added mSAP substrates, fixed-cost absorption is increasing in the process. Lower cost ratios from higher fixed-cost absorption, along with price increases and ASP gains, are combining to drive a sharp rise in operating margin.
Simec has a strong operating leverage profile. The combination of lower cost ratios and higher ASP is expanding the degree of profitability improvement, and the company is being seen as entering a new stage of profitability.
Major brokerage reports are focusing on the expansion of shipments of SOCAMM for AI servers and high-value-added package substrates. SOCAMM is a low-power memory module substrate for next-generation AI servers. The expansion in shipments of these products is driving product mix improvement, and product mix improvement is accompanying higher ASP, raising expectations for a structural leap in profitability.
A securities industry official explained that Simec has traditional strengths in memory substrates and has secured growth engines through SOCAMM for AI servers and CXL expansion substrates. However, the official said cyclical fluctuations in upstream conditions and risks tied to large-scale investment remain. The official added that, in the medium to long term, it is necessary to pay attention to technological leadership and improving fundamentals.
Source: TECHWORLD · Joo Ga-young
Original: https://www.epnc.co.kr/news/articleView.html?idxno=407481
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.