Korea AI PCB Ecosystem: 10 Companies Across FC-BGA, CCL, SoCAMM, Materials and Equipment

Tue, 05 May 2026 23:55:00 +0900

Sector map: This is the company-matrix companion to the AI PCB and Substrate Thesis and the AI PCB and Substrate Hub. The earlier thesis explained why AI substrates are a system-level bottleneck. This note asks which parts of the Korean ecosystem carry the better business quality, and where the factor setup still looks less fully priced.

TL;DR

The strongest accounting signal in Korea’s AI PCB ecosystem is not always inside the board manufacturers. It is upstream. Doosan Electronic BG’s high-end CCL business is running at roughly a 30% operating-margin profile, and Pamicell’s low-dielectric materials profile also sits around the 30% zone. Most board manufacturers sit closer to the mid-teens to low-20s. When margins cluster like that, the market is telling us where pricing power is concentrated: the shortage is partly in substrates, but the tighter bottleneck may be in CCL and low-loss materials.

Business quality and factor attractiveness are not the same ranking. A simple business-quality stack would put Doosan Electronic BG near the top, followed by Samsung Electro-Mechanics, Kolon Industries’ mPPO option, Isu Petasys, Daeduck Electronics and Pamicell. A price-aware factor stack looks different: Daeduck Electronics screens better because it has FC-BGA plus MLB exposure at a less demanding multiple; Kolon Industries is interesting because the mPPO option is diluted inside a low-multiple chemical and industrial-materials company; Pamicell offers upstream leverage but carries customer-concentration risk; Doosan has the best business, but the public-market access route is Doosan the holding company.

Doosan Electronic BG is the best business in the map, but not automatically the cleanest stock setup. Investors do not buy a separately listed Electronic BG. They buy Doosan, which includes other subsidiaries and holding-company complexity. The business may be excellent; the access route may already be expensive.

Kolon Industries is the most interesting dilution case. mPPO sales are estimated to move from roughly ₩20bn in 2024 to ₩80bn in 2025 and about ₩180bn by 2027. Yet this sits inside a company with roughly ₩5tn revenue and a single-digit enterprise multiple profile. A sell-side SOTP view can value the mPPO option near ₩1.5tn, while the whole market cap has been around the ₩2.5tn to ₩2.9tn range. The core issue is not whether mPPO matters. It is whether the market can see it through the rest of the company.

Taesung belongs in the ecosystem but not in the same valuation table. With 2025 revenue around ₩38bn, operating losses and a market cap around the multi-trillion-won zone, its price is not earnings-driven. It is an option on PCB capex and glass-substrate equipment. That can be valuable, but it must be held in a separate option basket, not compared one-for-one with profitable substrate and materials names.

My working portfolio lens is not a buy list. It is a factor map: Daeduck Electronics as the balanced core substrate factor; Kolon Industries and Pamicell as the upstream-materials barbell; Korea Circuit as an option; Doosan as a high-quality CCL anchor that requires entry discipline; Samsung Electro-Mechanics, Isu Petasys and Simmtech as higher-recognition names where valuation and pullback discipline matter; TLB as the SoCAMM-specific factor; Taesung as a separate equipment-option basket.

1. From A Manufacturer Screen To An Ecosystem Map

A narrow AI substrate screen usually starts with the board manufacturers: Samsung Electro-Mechanics, Isu Petasys, Daeduck Electronics, Simmtech and Korea Circuit. That is a useful starting point, but it is incomplete.

AI infrastructure does not stop at the board house. Advanced boards require high-end CCL, low-loss resin, glass fiber, copper foil, ABF-like inputs, module boards and the equipment that enables substrate capacity. Once the upstream layer is added, the map becomes more revealing.

AI infrastructure capex
 -> GPU / ASIC / CPU / switch ASIC demand
 -> package substrates, MLB, memory modules, SoCAMM
 -> CCL, low-dielectric resin, mPPO, hardeners, glass fiber, copper foil
 -> PCB and glass-substrate equipment

The ten-company ecosystem looks like this:

Layer	Companies	What they represent
Package substrate / FC-BGA	Samsung Electro-Mechanics, Daeduck Electronics, Korea Circuit	AI accelerator, CPU and ASIC package substrates
MLB / network boards	Isu Petasys, Daeduck Electronics, Simmtech, Korea Circuit	Server boards, switch boards, high-layer MLB
Memory module / SoCAMM	TLB, Simmtech, Korea Circuit, Daeduck Electronics	DDR5 and LPDDR-based server memory modules
CCL anchor	Doosan Electronic BG inside Doosan	High-end copper-clad laminate for AI accelerators and high-speed networking
Low-dielectric materials	Kolon Industries, Pamicell	mPPO, low-loss resin, hardeners and upstream inputs
Equipment	Taesung	PCB equipment and glass-substrate optionality

This is why a simple “AI PCB beneficiary” label is not enough. The ecosystem contains different margin structures, qualification risks, customer concentration risks and valuation regimes.

2. The Key Accounting Clue: Margins Are Higher Upstream

The most important observation is the margin gap.

Category	Company / segment	Working 2027E OPM profile
Upstream CCL	Doosan Electronic BG	~31%
Upstream low-dielectric materials	Pamicell	30%+ profile in recent / sell-side frames
MLB pure-play	Isu Petasys	~22%
Balanced substrate	Daeduck Electronics	~19%
Module substrate	TLB	~16%
Memory substrate	Simmtech	~14%
FC-BGA + MLCC large cap	Samsung Electro-Mechanics	~14%
FC-BGA option	Korea Circuit	high-single-digit to low-teens path
Diluted materials option	Kolon Industries	low-single-digit consolidated OPM, because mPPO is diluted

The margin message is simple: where pricing power is strongest, operating margin tends to show it first. If Doosan Electronic BG and Pamicell sit near 30% while many board makers sit in the 14% to 22% zone, the shortage is not just “boards are scarce.” It is also that the materials stack enabling those boards has pricing power.

This does not mean board makers are unattractive. It means the ecosystem must be separated into two questions:

Question 1: Which businesses have the highest quality and pricing power?
Question 2: Which public stocks still offer the best factor setup after price?

Those two answers are not the same.

3. The Ten-Company Working Matrix

The table below is a working matrix using May 4-5, 2026 price references and sell-side estimate sets where available. It is not a real-time quote sheet. Several lines use segment estimates or company-specific model work rather than standardized consensus.

Company	Reference price / market cap	2027E revenue	2027E OP	2027E OPM	2027E PER	OP growth 26 -> 27
Samsung Electro-Mechanics	~₩68.6tn market cap in the working sheet	₩13.83tn	₩1.94tn	14.0%	46.5x	+26.2%
Isu Petasys	~₩10.94tn	₩2.13tn	₩462bn	21.7%	29.7x	+44.4%
Daeduck Electronics	~₩5.65tn	₩1.88tn	₩363bn	19.3%	19.5x	+48.8%
Simmtech	~₩3.25tn	₩2.07tn	₩284bn	13.7%	18.5x	+70.1%
Korea Circuit	~₩2.19tn	₩2.20tn	₩190bn	8.6%	15.4x	+40.7%
Doosan Electronic BG via Doosan	~₩27.6tn Doosan market cap	₩3.44tn segment revenue	₩1.08tn segment OP	31.3%	20.9x Doosan-level frame	+41.8%
TLB	~₩0.89tn	₩415bn	₩65bn	15.6%	21.7x	+38.8%
Kolon Industries	~₩2.5tn to ₩2.9tn market cap range	₩5.18tn consolidated	₩241bn	4.7%	9.7x	+17.6%
Pamicell	~₩1.2tn market cap range	standardized consensus limited	–	30%+ margin profile	–	–
Taesung	~multi-trillion-won option valuation	consensus limited	loss / early-stage option frame	–	not meaningful	–

Several caveats matter. Doosan Electronic BG is a segment inside Doosan, not a separately listed company. Kolon Industries’ consolidated OPM understates the economics of the mPPO line because the option is diluted inside a much larger chemical and industrial-materials company. Taesung’s valuation is not an earnings multiple; it is an equipment-option price.

The arithmetic behind selected rows:

Doosan Electronic BG 2027E OPM = 1.076tn / 3.441tn = 31.3%
Doosan Electronic BG OP growth = 1.076tn / 0.759tn - 1 = 41.8%

TLB 2027E OPM = 64.8bn / 415.4bn = 15.6%
TLB OP growth = 64.8bn / 46.7bn - 1 = 38.8%

Kolon Industries consolidated OPM = 241.0bn / 5.18tn = 4.7%

4. GARP Re-Ranking

I use a simple GARP score as a sorting tool:

GARP score = operating-profit growth x operating margin / PER

This is not a valuation model. It is a way to keep growth, margin and price in the same frame.

Company / segment	OP growth	OPM	PER	GARP score
Doosan Electronic BG via Doosan	+41.8%	31.3%	20.9x	62.6
Simmtech	+70.1%	13.7%	18.5x	51.9
Daeduck Electronics	+48.8%	19.3%	19.5x	48.3
Isu Petasys	+44.4%	21.7%	29.7x	32.4
TLB	+38.8%	15.6%	21.7x	27.9
Korea Circuit	+40.7%	8.6%	15.4x	22.7
Kolon Industries consolidated	+17.6%	4.7%	9.7x	8.5
Samsung Electro-Mechanics	+26.2%	14.0%	46.5x	7.9

The GARP table says three useful things.

First, Doosan Electronic BG screens extremely well as a business. High growth plus a 31% operating-margin profile is rare.

Second, Daeduck Electronics remains the most balanced substrate-factor setup among the board manufacturers because it combines FC-BGA, MLB and a still-reasonable valuation frame.

Third, Kolon Industries looks weak on consolidated GARP precisely because mPPO is diluted. That is not a reason to ignore it. It is the reason to analyze it separately.

5. Doosan Electronic BG: Best Business, Imperfect Access Route

Doosan Electronic BG is the cleanest business-quality leader in the map. Public reports around Doosan’s 1Q26 results show Doosan’s own business revenue rising sharply, with Electronics BG strength driven by high-end CCL for AI accelerators, memory semiconductors and high-speed systems.

This is the business-quality case:

Point	Why it matters
OPM around 30%	Pricing power is concentrated in high-end CCL
AI accelerator and networking exposure	CCL content rises with GPU, CPU, switch ASIC and high-speed board complexity
Capacity expansion	Suggests management sees a multi-year cycle rather than one strong quarter
Customer qualification	High-end CCL is not a commodity slot once qualified into the AI supply chain

But the public-market access route is not pure. Investors buy Doosan, not separately listed Doosan Electronic BG. Doosan includes Electronics BG, Doosan Bobcat, Doosan Enerbility exposure, other self-business pieces and holding-company discount or premium dynamics.

That is the whole reason business quality and stock setup diverge. The business itself can be first-rate while the stock entry point is less clean.

What a Doosan share includes:
 - Electronics BG economics inside the holding company
 - Other affiliate values
 - Holding-company structure
 - Non-CCL volatility

What it does not give:
 - A pure listed CCL instrument
 - Direct 100% exposure to Electronics BG margin
 - Isolation from other subsidiary volatility

So Doosan Electronic BG is the benchmark for business quality. Doosan the stock needs price and holding-company discipline.

6. Kolon Industries: The mPPO Dilution Case

Kolon Industries is a different kind of opportunity. The interesting part is not consolidated operating margin. The interesting part is that mPPO can be material while still hidden.

The mPPO revenue path in the working sheet:

Item	2024	2025	2026F	2027F
mPPO revenue	~₩20bn	~₩80bn	~₩130bn	~₩180bn

A DS-style SOTP view can assign about ₩1.5tn of value to the mPPO option by applying a high multiple to estimated 2027 EBITDA. That is meaningful versus a total market-cap range around ₩2.5tn to ₩2.9tn.

The dilution mechanism is straightforward:

Kolon Industries consolidated:
 - revenue around ₩5tn
 - operating margin around mid-single digits
 - market sees a chemical / industrial-materials company

mPPO line:
 - much smaller revenue base
 - faster growth
 - higher implied margin
 - tied to high-end CCL and low-dielectric material demand

Result:
 - the consolidated multiple stays low
 - the option is visible to analysts but not fully separated by the market

If mPPO reaches about ₩180bn revenue in 2027 and around ₩59bn EBITDA, its standalone economics look nothing like a low-single-digit-margin commodity chemical business. The analytical question is whether the market begins to value Kolon as two businesses: a mature industrial-materials core plus a fast-growing mPPO option.

There are two ways for that to happen. The first is accounting visibility: more frequent IR disclosure, clearer segment commentary and sell-side SOTP models. The second is structural separation, which would require company action and should not be assumed. The more realistic path is accounting visibility.

7. Pamicell: Upstream Of The Upstream

Pamicell deserves a separate company-level treatment; this ecosystem note keeps the lens narrower: where does it sit inside the ten-company map, and why does that position matter for the broader AI CCL chain?

It is upstream of Doosan Electronic BG’s CCL.

Pamicell low-dielectric resin / hardeners
 -> Doosan Electronic BG high-end CCL
 -> FC-BGA / MLB / AI server boards
 -> AI accelerator, CPU, NIC and switch systems

That position creates both the upside and the discount.

The upside is operating leverage. If a small upstream supplier becomes qualified into a high-margin CCL cycle, incremental orders can move the income statement quickly. Pamicell’s 2025 operating margin was around 30%, and sell-side 2026 frames have discussed margins in the mid-30s.

The discount is customer concentration. Pamicell is not a broad board manufacturer. It is not Doosan Electronic BG. It depends on the order cadence, qualification status and sourcing decisions of a narrow upstream materials chain.

The key follow-up variables are the same ones from the Pamicell work:

Variable	Why it matters
1Q26 operating profit and margin	Confirms whether high-margin materials are flowing into reported earnings
Follow-on Doosan contracts in May-June	Tests whether the February contract was a new run-rate or timing pull-forward
Third Ulsan plant progress	Determines whether 2027 capacity can support the thesis
KRX reclassification after May 4	Tests whether market perception is moving from biotech to electronics materials

Pamicell is one of the sharper expressions of the upstream thesis, but it must be sized and analyzed as a customer-concentration story.

8. TLB: The SoCAMM-Specific Factor

TLB deserves a separate slot because its factor exposure is different. It is closest to a SoCAMM and server memory-module board angle.

TLB working 2027E:
 revenue: ₩415bn
 operating profit: ₩65bn
 OPM: 15.6%
 PER: 21.7x
 GARP: 27.9

SoCAMM matters because LPDDR-based server memory modules can become part of the next-generation CPU and AI inference memory stack. The more CPU-heavy agentic AI becomes, the more interesting the memory-module board layer becomes.

But SoCAMM is still a factor, not a fully proven revenue base. Vera CPU adoption, OEM design wins and the final module-standard path all matter. That makes TLB cleaner than many diversified PCB names for the SoCAMM angle, but also more dependent on one thesis becoming real.

The practical map:

SoCAMM purity:
 1. TLB
 2. Simmtech
 3. Korea Circuit
 4. Daeduck Electronics

TLB is not the highest-quality company in the ecosystem, but it is a useful single-factor monitor for SoCAMM.

9. Taesung: Do Not Put It In The Same Table

Taesung belongs in the AI substrate ecosystem because equipment matters. If substrate makers expand capacity, equipment suppliers can benefit. If glass substrates become a real production path, specialized equipment can gain option value.

But Taesung should not be ranked like the other nine names.

The reason is valuation structure. The working numbers point to a company with 2025 revenue around ₩38bn, operating losses and a multi-trillion-won market cap. That implies a price-to-sales ratio in the dozens, closer to an option valuation than a normal earnings stock.

Taesung 2025 working frame:
 revenue: \~₩38bn
 operating profit: negative
 market cap: multi-trillion-won zone
 PSR: roughly 60x+

That is not automatically wrong. Options can be valuable. But the analysis method changes.

Taesung driver	What to monitor
PCB equipment cycle	Large orders from substrate capacity expansion
Glass substrate option	Evidence of pilot-to-mass-production transition
Consensus formation	Whether 2026-2027 earnings estimates become real enough to underwrite

Until those variables mature, Taesung belongs in a separate option basket, not a GARP table beside Doosan Electronic BG or Daeduck Electronics.

10. Business Quality Versus Factor Attractiveness

This is the most important conclusion.

Business quality ranking:

1. Doosan Electronic BG
2. Samsung Electro-Mechanics
3. Kolon Industries mPPO, if separated
4. Isu Petasys
5. Daeduck Electronics
6. Pamicell

Price-aware factor attractiveness:

1. Daeduck Electronics
2. Kolon Industries
3. Korea Circuit
4. Pamicell
5. Doosan
6. TLB
7. Isu Petasys
8. Simmtech
9. Samsung Electro-Mechanics
10. Taesung, separate option basket

The point is not that the second list is a trading recommendation. It is a reminder that good businesses and good factor setups are different things.

Good business -> market recognizes quality -> price rises -> incremental alpha shrinks
Hidden or complex business -> market discounts the story -> price lags -> alpha can remain

Doosan Electronic BG is excellent, but the stock route is a holding company. Samsung Electro-Mechanics is excellent, but the market has already recognized a lot of it. Daeduck is less pure, but has balanced FC-BGA and MLB exposure at a less demanding multiple. Kolon Industries looks cheap because mPPO is hidden inside a larger company. Pamicell still carries the old biotech label in some investors’ minds, even though its materials economics have changed.

Alpha rarely comes from the highest-quality business after everyone agrees it is high quality. It more often comes from the place where the market has not yet cleanly separated the valuable part from the rest.

11. What Could Be Wrong

Several parts of this framework can fail.

Doosan Electronic BG’s 30% margin could be a peak-cycle outcome rather than a structural margin. If CCL capacity catches up quickly, the margin signal weakens.

Kolon Industries’ mPPO revenue path could lag. If the line moves slower than the ₩130bn to ₩180bn path, the dilution discount may be justified.

Doosan’s holding-company structure could remain a permanent discount. A great segment does not always create full value for minority shareholders if the access route is complex.

Taesung’s glass-substrate option could remain an option and never convert into revenue. In that case, the option multiple normalizes.

Pamicell’s customer concentration could become the wrong kind of leverage. If follow-on contracts slow, the small-base upside works in reverse.

The framework is useful only if the quarterly data keeps validating it.

12. Verification Checklist

Layer	Company	Signals to track
CCL anchor	Doosan / Electronic BG	Electronics BG OPM near 30%, CCL capacity expansion, customer diversification
mPPO	Kolon Industries	mPPO revenue path, separate IR commentary, SOTP adoption by sell-side
Upstream materials	Pamicell	1Q26 margin, follow-on Doosan contracts, third-plant progress
Balanced substrate	Daeduck Electronics	FC-BGA customer wins, MLB capacity, server-grade product mix
MLB	Isu Petasys	800G / 1.6T demand, pricing pass-through, OPM durability
Premium anchor	Samsung Electro-Mechanics	AI data-center package substrate wins, MLCC price / mix, estimate revisions
Memory substrate	Simmtech	SoCAMM revenue, 2026 OPM recovery
Option substrate	Korea Circuit	Broadcom-linked FC-BGA revenue, SoCAMM ramp, consolidated margin
SoCAMM	TLB	SoCAMM module-board ramp, DDR5 high-value mix
Equipment option	Taesung	Major orders, glass substrate equipment production, credible consensus formation

FAQ

What is the Korea AI PCB ecosystem?

It is the listed Korean supply chain around AI substrates: FC-BGA, MLB, memory module boards, SoCAMM, high-end CCL, low-dielectric materials and PCB / glass-substrate equipment.

Why do margins matter so much in this analysis?

Margins show where pricing power is accumulating. If upstream CCL and low-dielectric material companies show higher margins than board manufacturers, the bottleneck may be closer to the material stack than to assembly alone.

Why is Doosan Electronic BG not simply the top stock?

Because it is not separately listed. The public-market route is Doosan, a holding-company structure with other affiliate values and risks. Business quality and stock access are not the same thing.

Why is Kolon Industries interesting despite low consolidated margins?

Because the mPPO option may be hidden inside a much larger chemical and industrial-materials company. The consolidated margin and PER can obscure the economics of the mPPO line.

Is Taesung comparable to the other nine companies?

No. Taesung is better treated as an equipment and glass-substrate option. Its valuation is not driven by current earnings in the same way as the profitable substrate or material names.

Selected Public Sources

Doosan 1Q26 high-end CCL strength, Seoul Economic Daily: https://en.sedaily.com/finance/2026/04/29/doosan-q1-operating-profit-jumps-717-percent-on-high-end
Doosan earnings document summary, MarketScreener: https://www.marketscreener.com/news/doosan-earnings-document-ce7f59d3dd81f224
Pamicell 2026 Doosan Electronic BG supply contract report: https://www.g-enews.com/article/Securities/2026/02/20260219100622493844093b5d4e_1
Lotte Energy Materials and Doosan Electronic BG high-performance PCB copper foil alliance: https://en.sedaily.com/finance/2026/03/22/lotte-energy-materials-doosan-electronics-unit-partner-on
Taesung financial overview, StockAnalysis: https://stockanalysis.com/quote/kosdaq/323280/financials/

Final Note

The statement “AI substrates are scarce” is true, but it is too shallow. The better question is: where does the scarcity convert into margin?

The ten-company map suggests the answer is upstream. Doosan Electronic BG and Pamicell show margin profiles around the 30% zone, while many board manufacturers sit in the 14% to 22% range. That does not make board makers bad. It means the material stack deserves its own analytical weight.

The most important investment lesson is that business quality and factor attractiveness diverge. Doosan Electronic BG may be the best business, but the listed route is Doosan. Samsung Electro-Mechanics may be the premium anchor, but recognition is already high. Daeduck Electronics may be less glamorous, but the combination of FC-BGA, MLB and valuation looks more balanced. Kolon Industries may look like a low-multiple chemical company, but mPPO can create a hidden option. Pamicell still carries old-label risk, but that is exactly why the recognition gap can matter.

The work from here is straightforward: track margins, contract cadence, mPPO disclosure, SoCAMM adoption, glass-substrate equipment orders and whether the market starts separating the valuable hidden parts from the consolidated wrappers.

Disclaimer: For research and information purposes only. Not investment advice. Names cited are for analytical illustration; readers should perform their own due diligence and consult licensed advisors before any investment decision.

Isu Petasys on Korea Invest Insights