Ibiden Co., Ltd.(4062) · PCBs & Substrates

Long-Term Value Research on Ibiden

Other languages
Quick ReadPlain-language overview · read this first

Ibiden is a Japanese company focused on high-end chip substrates. The report rates it as “Watch”: the company is worth studying, but at the current price there is no need to rush into buying it.

Its strongest capability is producing an extremely difficult class of chip package substrate and supplying it reliably to top-tier customers such as Intel, AMD, and Nvidia. Think of this product as the high-precision base beneath a chip that supports it and routes its connections. It is benefiting directly from booming demand for AI servers. Once a customer has qualified a supplier, switching becomes difficult, which gives Ibiden a capability that competitors cannot easily take away. The issue is that these three major customers account for half of its revenue, so any order cuts from them could make earnings swing sharply. The company also has a more traditional ceramics business, so the story is not entirely pure-play.

The report is most concerned with the quality of earnings. On paper, the company earned about 63.7 billion yen this year, which looks strong, but 49.4 billion yen came from a one-off gain on selling investment securities, not from the true earning power of the core business. More importantly, the company is spending heavily to expand capacity, with about 500 billion yen of additional investment planned over the next three years, and most of its profit has not turned into cash that can be distributed to shareholders. In other words, the profit is real, but its ability to convert that profit into cash is currently not strong.

The report is blunt on valuation. Based on current earnings, buying the whole company would take more than 80 years to pay back, which is clearly expensive. The report’s more realistic buying range is about 3500 to 6000 yen, while the current price is 17,570 yen, meaning investors have already paid upfront for several years of good news. The biggest risk is that if AI demand cools or the capacity expansion falls short of expectations, the share price could fall by more than half. The report respects the company, but does not agree with the current price. Its suggestion is to put the stock on the watch list first and wait for a cheaper entry point.

The above is only a plain-language explanation of this report and is not investment advice. The stock market involves risk; invest with caution.

Lead

Ibiden is a global leader in high-end IC package substrates (ABF/FC-BGA), with its electronics business contributing about 59% of FY2025 total sales of JPY 416.2 billion, core customers including Intel, AMD, and NVIDIA, and a mature ceramics/DPF business alongside a heavy AI/HPC-driven expansion cycle of roughly JPY 500 billion over FY2026-28, backed by a 57.3% equity ratio and net cash of about JPY 100.4 billion. The core thesis is that Ibiden owns valuable process know-how, yield capability, customer qualifications, and an early position in AI/HPC substrates, but its capital intensity, customer concentration, and mixed business profile make the current valuation demanding. Research rating Watch: a strong company worth studying, but not a price that offers enough margin of safety today.

Full report

Conclusion First

From the perspective of a long-term business owner, Ibiden is a company with high technical barriers and deep customer relationships, but also very heavy capital expenditure and a business mix that is not especially pure. Its most valuable assets are not the factories on the balance sheet, but its process capability, yield, customer qualification status, and early position in high-end IC package substrates for AI/HPC. At the same time, the same company also contains a more mature ceramics/DPF business that is partly exposed to internal-combustion-engine cycles, so this is not the kind of simple business one can confidently overweight for ten years at a glance. The latest complete annual results available are the FY2025 full-year results for the year ended March 31, 2026. The FY2025 Annual Securities Report will not be filed until June 18, 2026, so the most reliable materials for now are the latest full-year earnings release, briefing materials, and the audited FY2024 Financial Review/Integrated Report 2025.

My view on Ibiden can be summarized as follows: the investment rating is Watch; the core judgment is that the company itself is worth studying, but the current price is not worth rushing into. The current price has no margin of safety. It is suitable for long-term investors who understand the semiconductor packaging cycle and can tolerate valuation drawdowns; it is less suitable for ordinary conservative investors. The biggest uncertainties are concentrated in three areas: the durability of AI substrate demand, the return on the JPY 500 billion capacity expansion, and whether the current valuation has already over-discounted the 2030 target.

As of 11:30 JST on June 9, 2026, Ibiden's share price was about JPY 17,570, implying a reference market value of about JPY 4.95 trillion. Yahoo! Finance Japan showed a company-forecast PER of about 84.6x, PBR of about 8.92x, and dividend yield of about 0.20%. Compared with the company's FY2026 guidance of JPY 58.0 billion in net profit and EPS of JPY 207.70, the market is already paying for a very strong, very long, and very smooth expansion story, rather than paying for cash flows that have already been realized. For an investor with a "balanced but conservative" risk preference and a holding period above 10 years, this looks more like a good company at a bad price.

If I had to compress my preliminary view into one sentence: Ibiden is a quality participant in the high-end AI substrate track, but by Buffett-style value-investing standards, it currently looks more like an excellent asset that requires patience than a buying opportunity with sufficient margin of safety. The key to this judgment is not short-term share-price movement. It is that the company is in a heavy capital expansion phase, true distributable cash flow has not exploded in step with accounting profit, while the market valuation has already run far ahead.

Business Understanding and Industry Structure

Ibiden currently has three main businesses: Electronics, Ceramics, and Other. FY2025 segment sales were JPY 243.3 billion, JPY 82.6 billion, and JPY 90.3 billion, respectively, versus company-wide total sales of JPY 416.2 billion. Electronics is the core business. Ceramics provides some diversification, but it also introduces cyclical and structural variables that differ from semiconductors. In its Integrated Report, the company defines "cutting-edge semiconductor IC package substrates" as the growth center of the group and states that it has a global leading share in high-end IC package substrates.

Put simply, this business makes money by manufacturing extremely difficult high-performance IC package substrates and delivering them reliably to top-tier chip customers, while also selling DPFs, holding materials, and specialty graphite in the ceramics business. The electronics business is essentially a compound business of "difficult manufacturing + joint customer development + repeated qualification + long-term quality control." It is not ordinary PCB. The ceramics business contains more mature product lines, some of which may face long-term substitution. The company's website also clearly states that the electronics business focuses on plastic packaging, printed circuit boards, and related products, while the ceramics business focuses on DPF and special carbon products.

On customers, the audited FY2024 Financial Review disclosed major customers under the FY2024 basis including JPY 76.7 billion from Intel, JPY 40.7 billion from AMD, and JPY 75.1 billion from NVIDIA. Against FY2024 total sales of JPY 369.4 billion, these three customers together accounted for about 52% of revenue. This shows that Ibiden has top-tier customer endorsement, but also clear customer concentration risk. If a product generation changes, a customer insources or shifts orders, or the demand rhythm changes, profit volatility can be very large.

On revenue stability, this is not highly predictable subscription-software revenue, but it is also not a pure spot business. Once high-end substrates pass qualification, customers face high costs when switching suppliers. Qualification cycles are long, yield ramp is slow, and the cost of mistakes is high, so revenue has a degree of "semi-recurring" character. Still, order volume is affected by chip cycles, customer CapEx, server-platform iterations, and product-mix changes. In the FY2025 Q&A, management explicitly said that when a new-generation product is first introduced, the company can often obtain nearly 100% share, but after about 3 to 6 months, share may fall back by 20% to 30%. This shows that its advantages are real, but also that competition in the industry never disappears.

In cost structure, Ibiden is a high fixed-cost, depreciation-heavy, CapEx-heavy manufacturer. FY2025 depreciation and amortization was JPY 66.2 billion, and additions to fixed assets and intangible assets were JPY 64.3 billion. In February 2026, the company formally announced that it would invest about JPY 500 billion in high-performance IC package substrates over FY2026-FY2028, including about JPY 220 billion for Phase 1 projects such as the Gama Plant. This means that even if the business is strong, it is not an asset-light cash machine. It is a manufacturing platform that must keep spending to maintain process leadership.

The industry needs to be viewed in parts. The sub-industry of high-end advanced package substrates remains in an AI/HPC-driven growth phase. A Yole 2026 report summary indicates that the advanced IC substrate market is expected to expand at about a 9% CAGR, mainly driven by AI accelerators. Reuters, citing SEMI, also showed that semiconductor wafer-fabrication equipment sales are expected to rise 9% in 2026 and another 7.3% in 2027. However, Ibiden is not a pure AI substrate company. It also has more mature segments such as ceramics/DPF, so the group as a whole is not a single-line growth stock, but a hybrid of growth track plus traditional manufacturing assets.

On competitive structure, industry materials usually list Ibiden, Unimicron, Shinko Electric, Samsung Electro-Mechanics, and AT&S as major participants in advanced IC substrates. One of the closest Japanese comparables is Shinko Electric. AT&S describes itself on its website as a "leading supplier of high-end PCBs and IC substrates," while Shinko Electric's FY2024 disclosures also clearly state that its plastic molded packages, flip-chip type packages, plastic BGA substrates, and related businesses are being affected by both AI demand and server/PC competition. In other words, this is an industry with high technical barriers but also large cyclical swings. Profit pools are concentrated among a small number of players that can make high-end products and deliver reliably, but price wars in low-end products and mature platforms can erode margins at any time.

On whether this is a business I can understand, my answer is: the economic logic is understandable, but the technical details are not simple. From an owner perspective, one can understand why it makes money and why it must keep investing. But if one cannot continuously track advanced packaging, customer-node transitions, substrate-specification evolution, yield, and capacity utilization, one has not truly understood it. If I score "understandability" by Buffett standards, I give it 3 points out of 5. If the market closed for 5 years, I would be willing to own this company, but only at a much lower valuation level. This judgment is based on the company's segment composition, customer concentration, industry growth, and cyclical characteristics.

On industry attractiveness, I give it 3 points out of 5. The reason is direct: high-end AI substrates are a good track, but the whole company is not a pure-track asset. At the same time, this track has fast technology upgrades, heavy capital spending, strong customer bargaining power, and large cyclical swings. It does not have the "more mature, more comfortable" character of a consumer-goods monopoly or software platform.

Moat and Management

Ibiden's moat lies mainly not in brand or network effects, but in process capability, qualification, customer relationships, mass-production yield, and capacity-expansion execution. In the Integrated Report, the company writes that it can use the technology, experience, and advanced production facilities accumulated over "more than 40 years" to meet substrate specification requirements for larger sizes, more layers, and finer lines. Management further stated in the FY2025 Q&A that, for silicon bridge-type substrates, its technological advantage can be maintained "at least until FY2030." This indicates that the company's strongest barriers are manufacturing know-how and joint development capability, not a visible consumer brand.

More specifically, Ibiden's moat profile is roughly as follows:

Moat Type Judgment Explanation
Brand advantage Weak End consumers have little awareness; brand is not the core of pricing power
Cost advantage Medium High-end products depend more on yield/process than lowest cost
Scale advantage Medium to strong Only a few companies can expand advanced substrate capacity at very large scale
Network effects None This is not a platform business
Switching costs Medium to strong Customer qualification, quality certification, and platform switching costs are high
Channel advantage Weak Mainly B2B direct supply, not dependent on distribution channels
Patent/regulatory barriers Medium More reflected in process experience and manufacturing barriers than in a single patent
Data advantage Weak This is not a data-platform company
Corporate culture/operating capability Medium to strong Long manufacturing experience, quality control, and global multi-site supply capability matter
Capital allocation ability Medium Improving, but not enough for a high score

The company's recent signal is that the high-end moat is strengthening, while low-end/mature moats are weakening. Management itself acknowledged that the competitive environment for PC-related products will intensify further, and FY2025 included a JPY 16.4 billion impairment for the Philippine electronics business and idle assets. On the other hand, the company believes demand is very strong for high-performance substrates related to AI servers, ASIC/GPU, and switches, and that total industry demand will exceed supply capability. In other words, investors should not loosely say "Ibiden has a strong moat." A more accurate statement is: it has a relatively strong moat in the most advanced AI substrate niche, but the moat is not wide in mature platforms and some traditional businesses.

The company's ability to raise prices in an inflationary environment is limited but real. In the FY2025 Q&A, management said that increases in material costs had been passed through with customer consent, but price negotiations were conducted within a "reasonable range" because the company values long-term customer relationships. This answer resembles that of a strong B2B industrial company: it has some bargaining power, but cannot raise prices freely like a luxury-goods company. It indicates that the company is not a price taker, but also does not have absolute pricing power.

In an economic downturn, the company is likely to survive very well, but may not necessarily earn very comfortably. FY2025 electronics segment margin rose to about 18.6%, but FY2024 was only about 13.6%. The ceramics business had an FY2025 margin of about 9.2%, significantly below FY2024's about 14.5%. This difference shows that group-level earnings resilience is decent, but profitability is still clearly affected by product mix and cycles. This is not a business that can sustain high ROIC and high FCF through every cycle.

On management credibility and capital allocation, I give it 3 points out of 5. On the positive side: the board has 12 directors, of whom 6 are outside directors, and all 6 are independent directors; the company is publicly advancing "management conscious of capital cost and stock price" and has set an FY2027 ROE target of above 10%; as of FY2025, the company had also completed ahead of schedule its target for reducing policy shareholdings, with a reduction ratio of 81.2% based on market value at the end of FY2023. These items show that management is moving from a traditional Japanese manufacturer toward greater emphasis on capital efficiency.

More importantly, at the FY2025 briefing, the company explained its capital allocation logic relatively clearly: cumulative operating cash flow over FY2026-FY2028 of about JPY 800 billion before R&D will be used mainly for growth investment, while also balancing debt reduction and shareholder returns; the target dividend payout ratio is 20%; growth investment will in principle be completed within the framework of operating cash flow and customer advances. This shows that management is at least verbally aware that it cannot chase growth through unlimited external financing, and is trying to ground expansion in customer demand and cash flow. The company even stated clearly that it is not considering additional financing at present. From a shareholder perspective, this is a positive signal.

However, conservative investors should not ignore the other side. First, management's own shareholding is not high. The 2025 shareholders' meeting materials show Representative Director and President Koji Kawashima holding about 39,200 shares, and Chairman Takeshi Aoki about 89,500 shares. Relative to 281.7 million shares outstanding, these stakes are very small. Alignment depends more on position, reputation, and equity incentives than on large personal wealth tied to common shareholders. Second, FY2025 net profit included JPY 49.4 billion in gains on sales of investment securities, while repurchases were only at the JPY 10 million level, indicating that current capital allocation is clearly focused on expansion and financial adjustment, not large buybacks at undervaluation. Finally, the FY2024 auditor identified the "accuracy of accounting treatment for large-scale capital expenditures in the electronics business" and assumptions in "impairment testing of the Philippine subsidiary" as key audit matters. This reminds us that management is not opaque, but the areas most worth watching in this type of company are returns on capital expenditure and impairment assumptions.

Overall, I would score Ibiden's moat strength at 3 points out of 5. In the high-end AI substrate niche, it is close to 4 points. But at the whole-company level, because of customer concentration, mature-business drag, and cyclical and CapEx volatility, it has not reached the level of a business that can be compounded long term without much thought.

Financial Quality and Owner Earnings

First, consider the core financial profile over the past six fiscal years. The table uses the company's FY basis, with amounts in JPY billion. FY2020-FY2024 mainly come from the historical data pages of Integrated Report 2025, while FY2025 comes from the latest full-year results release and cash-flow statement. The "free cash flow" column follows the common disclosure basis used on the company's historical pages, which is closer to operating cash flow plus investing cash flow. This differs from strict Buffett-style FCF, so I make a more conservative Owner Earnings adjustment after the table. The company's financial trajectory over the past five years has not been a smooth upward line. It has clearly been affected by cycles, product mix, and CapEx phases. The historical financial chart in the company's Integrated Report 2025 also visually shows the volatility in ROE and CapEx.

Fiscal Year Revenue Operating Profit Operating Margin Net Profit Attributable to Owners Net Margin ROE Equity Ratio Interest-Bearing Debt CapEx Depreciation and Amortization Company-Basis Free Cash Flow
FY2020 323.5 38.6 11.9% 25.7 7.9% 8.8% 54.6% 150.2 78.2 35.4 -43.4
FY2021 401.1 70.8 17.7% 41.2 10.3% 12.1% 54.8% 170.0 60.6 52.7 40.7
FY2022 417.5 72.4 17.3% 52.2 12.5% 13.3% 48.9% 270.0 131.3 54.9 21.7
FY2023 370.5 47.6 12.8% 31.5 8.5% 6.9% 43.8% 343.5 146.6 46.0 68.0
FY2024 369.4 47.6 12.9% 33.7 9.1% 6.8% 45.3% 343.0 157.3 54.2 -45.3
FY2025 416.2 62.0 14.9% 63.7 15.3% 12.2% 57.3% 192.5 64.3 66.2 54.0

The first thing worth emphasizing in this table is not growth, but growth quality. On an FY2020 to FY2025 basis, revenue compounded at about 5.2% and operating profit at about 9.9%, which looks good. But FY2025 net profit attributable to owners of JPY 63.7 billion included JPY 49.4 billion in gains on sales of investment securities, which is clearly not sustainable operating profit. If this portion is stripped out, the "sustainable component" of FY2025 net profit is much less impressive than the headline suggests. In other words, high growth in headline net profit does not mean the company has already entered a true cash-flow harvest phase.

On margins, the recent two-year gross margin improved from about 30.7% in FY2024 to about 31.6% in FY2025, and operating margin rose from 12.9% to 14.9%. This is real improvement. But it mostly reflects a higher mix of AI-related high-value-added products and better electronics-business conditions, rather than the kind of long-term stable expansion seen in a mature monopoly. FY2025 electronics segment profit reached JPY 45.2 billion, a clear improvement from JPY 26.8 billion in FY2024. By contrast, ceramics segment profit fell from JPY 12.2 billion to JPY 7.6 billion. This again reminds investors that group profit improvement is highly dependent on the product mix of the electronics business, not on all businesses becoming stronger at once.

Cash-flow quality is more revealing than the income statement. FY2025 operating cash flow was JPY 106.4 billion, versus JPY 118.9 billion in FY2024. Cash-flow details show that FY2025 receivables increased by about JPY 10.1 billion, work in process increased by about JPY 6.0 billion, and customer advances decreased by about JPY 11.1 billion, indicating that while the business was expanding, working capital was consuming cash. In other words, strong orders did not automatically translate into "cash that shareholders can take away." Part of the growth has to be supported by changes in receivables, work in process, and advances received.

If one uses a stricter FCF definition closer to what value investors usually prefer - operating cash flow minus expenditures for purchases of PPE and intangible assets - FY2024 strict FCF was about negative JPY 79.6 billion, and FY2025 was about negative JPY 0.8 billion, far weaker than the "free cash flow" shown on the company's historical pages. This is not an accounting problem, but a definition problem. The company's historically disclosed FCF includes buffers from sales of investment assets, such as FY2025 cash inflow of JPY 57.4 billion from sales of investment securities. For business owners, this portion cannot be relied upon over the long term, so cash-generation ability must be evaluated more strictly.

This is why I view Ibiden as a manufacturing company whose growth requires capital, rather than a compounding machine that becomes lighter as it grows. FY2025 additions to fixed assets and intangible assets were JPY 64.3 billion. Management has also made clear that it will continue to invest JPY 500 billion over FY2026-FY2028 to expand high-performance substrate capacity, and the FY2026 capital spending plan alone is JPY 210.0 billion. As long as this expansion round continues, true distributable cash flow will struggle to look as good as accounting profit.

The good news is that the balance sheet is very solid. At the end of FY2025, cash and cash equivalents were about JPY 292.9 billion. Based on balance-sheet details, estimated total interest-bearing debt was about JPY 192.5 billion, implying net cash of about JPY 100.4 billion. The equity ratio was 57.3%, and operating profit/interest expense was about 45.7x. This means the company has ample survivability in economic downturns and industry volatility. Permanent capital loss is more likely to come from valuation mean reversion after buying at a high price than from a balance-sheet blowup.

On accounting risk, I do not see clear signs of financial fraud, but I would rate the risk of "aggressive accounting or earnings manipulation" as low to medium and requiring ongoing monitoring. The reason is that, on one hand, disclosure is relatively adequate and the audit opinion is unqualified. On the other hand, the key audit matters in the prior year's audited financial statements focused heavily on when large capital expenditures begin to be depreciated after transfer to fixed assets, and on sales-growth, margin, and discount-rate assumptions in the impairment testing of the Philippine subsidiary. This is not a negative conclusion, but it points exactly to where financial judgment is most likely to go wrong for this company.

From an Owner Earnings perspective, I use a more conservative basis. Taking FY2025 as an example, operating cash flow of JPY 106.4 billion minus one-off subsidy-related cash inflow of about JPY 8.7 billion gives normalized operating cash generation of about JPY 97.7 billion. The company does not separately disclose maintenance CapEx, so I estimate it in a range of JPY 50.0 billion to JPY 70.0 billion. This assumption is inherently uncertain, but given that company depreciation has reached JPY 66.2 billion and high-end substrate technology iterates quickly, I am unwilling to set maintenance CapEx too low. On this basis, FY2025 conservative Owner Earnings were roughly JPY 27.7 billion to JPY 47.7 billion, with a midpoint of about JPY 37.7 billion. Against the current market value of about JPY 4.95 trillion, the market is paying roughly 104x to 179x this portion of true distributable cash flow, with a midpoint of about 131x. This is not cheap in a value-investing sense. The above is my estimate, not company-disclosed data; the key uncertainty lies precisely in maintenance CapEx.

Put more plainly: Ibiden's profits are real, but its ability to immediately turn profit into distributable cash is not strong, at least during the current expansion cycle. This does not negate the company's long-term value, but it has a major impact on whether buying today can produce a good return.

Intrinsic Value and Margin of Safety

I use three valuation methods, and in each case I try to lay out the assumptions. The most important point first: if valuation assumptions are unclear, the conclusion is meaningless. For Ibiden, the hardest and most critical variable is not next year's profit, but how far this AI substrate expansion can lift Owner Earnings, how long that level can last, and how much capital it requires.

First, the Owner Earnings discounted-cash-flow method. I use the conservative FY2025 Owner Earnings midpoint of about JPY 37.7 billion as a reference, then adjust starting point, growth rate, and discount rate by scenario. The settings are as follows: the conservative scenario starts from JPY 30.0-40.0 billion, grows at about 10% over the next five years, then slows to 3%-4%, with a discount rate of 9.5%-10.5% and terminal growth of 2%; the neutral scenario starts from JPY 40.0-50.0 billion, grows 18% for the first five years and 6% for the following five years, with a discount rate of 8.5%-9.5% and terminal growth of 2.5%; the optimistic scenario starts from JPY 50.0-60.0 billion, grows 25% for the first five years and 8% for the following five years, with a discount rate of 8%-9% and terminal growth of 3%. Based on these assumptions, I estimate per-share intrinsic value at roughly: conservative range JPY 2200-3200, neutral range JPY 4750-6870, optimistic range JPY 9200-13370. This is my estimate based on conservative cash-flow logic, not a price forecast.

Compared with the current share price of about JPY 17,570, the above results imply that the current price is at a premium of about 4.5x to 7.0x to conservative intrinsic value, about 156% to 270% to neutral intrinsic value, and still about 31% to 91% above optimistic intrinsic value. To make the current price reasonable, the market broadly needs the company to lift Owner Earnings at a very high speed over the next five years and still maintain very high returns after expansion is complete. Simply put, the market is not buying "today." It is buying an "almost perfectly delivered 2030 story."

Next, relative valuation. Ibiden's current reference metrics are roughly: PER 84.6x, PBR 8.92x, and P/S about 11.9x. Based on FY2025 operating profit plus depreciation, EV/EBITDA is around 37-38x. By comparison, one of the closest Japanese comparables, Shinko Electric, had the last visible public-market data of about 2.81x PBR, about 6.52% ROE, FY2024 revenue of JPY 215.0 billion, and corresponding market value of about JPY 797.5 billion, implying P/S of about 3.7x. The Yahoo snippet for AT&S shows Trailing P/E of about 43x and P/S of about 3.24x. The Yahoo snippet for Samsung Electro-Mechanics shows Trailing P/E of about 161x, Forward P/E of about 111x, and P/S of about 11.27x, but it is a more diversified large electronics-components company. The conclusion is not "peers are all expensive, so Ibiden is reasonable." The opposite is true: relative valuations across the AI chain are also high, and a basket of high valuations cannot be used to prove each other cheap. Against a heated peer backdrop, Ibiden still looks very expensive, especially with PBR far above Shinko Electric and far above the Japanese broad market.

The asset method better shows how far away the floor is. At the end of FY2025, company net assets were JPY 557.4 billion, book value per share was JPY 1969.85, cash and cash equivalents were JPY 292.9 billion, investment securities were JPY 32.3 billion, estimated total interest-bearing debt was about JPY 192.5 billion, and net cash was about JPY 100.4 billion. This shows that the balance sheet is sound, but the asset method provides almost no support for the current share price. According to Yahoo Finance Japan, current PBR is already 8.92x. Of course, the asset method undervalues high-end substrate customer qualifications and process assets. But if the investment case for a heavy-asset manufacturer is almost entirely unprotected by book assets, investors must be much stricter about future cash returns.

Combining the three methods, I give the following valuation ranges:

Range Per-Share Value
Conservative intrinsic value range JPY 2200-3200
Fair intrinsic value range JPY 4750-6870
Optimistic intrinsic value range JPY 9200-13370

Based on this framework, I would set the ideal buy-price range at JPY 3500-6000. This is the range after retaining a margin of safety below the "fair intrinsic value" range. The acceptable hold-price range is roughly JPY 6000-9000, assuming you already understand the business deeply and are willing to bear cyclical and valuation volatility. The clearly overvalued range starts at above JPY 13500. The current price of about JPY 17,570 is already in this category.

Therefore, the margin-of-safety conclusion is very clear: the current price does not provide enough margin of safety. The most fragile assumption in the valuation is that the market is assuming the company can execute this JPY 500 billion expansion at high quality and move close to a path of "JPY 1 trillion in revenue and JPY 300 billion in operating profit" around FY2030. If growth falls short, margins pull back, or valuation multiples merely contract to more normal manufacturing levels, investment returns may be poor. For conservative investors, the right action is not to debate whether it is a good company. It is to recognize that today is better spent waiting for a better price.

Risks, Counterarguments, and Opportunity Comparison

The most important risks almost all point to permanent capital loss rather than "quarterly volatility." I summarize them as follows:

Risk Why It Matters Current Evidence
Competition risk Price wars in mature PC/general-server substrates can consume profits The company has impaired the Philippine business and acknowledged intensifying PC-related competition
Technology substitution risk Advanced packaging routes evolve too quickly, and a wrong bet can make heavy assets lose depreciation value The company needs continuous expansion and R&D, while process difficulty keeps rising
Expansion execution risk If the JPY 500 billion investment falls short on yield/ramp/demand, returns can collapse The company is in a heavy capital expansion phase
Customer concentration risk Concentration among top customers amplifies platform-switching and order volatility Intel, AMD, and NVIDIA together accounted for about 52% of FY2024 sales
Cyclical risk If AI strength slows, high valuation will pull back before earnings do The current market valuation clearly bets on sustained strength
Overvaluation risk Even a good company can produce no return for years if bought too expensive Current PER/PBR are significantly high
Supply-chain risk Materials, equipment, labor, and delivery delays can affect expansion Management says further above-plan expansion before FY2028 is difficult
FX and interest-rate risk High overseas revenue means both the yen and interest rates affect profit and valuation Overseas sales exposure is high, and the 10-year JGB yield has risen to around 2.73%
Accounting and impairment risk Large CapEx capitalization and impairment judgments affect profit quality Audit KAMs focus on capital-expenditure accounting and Philippine impairment
Business-model disruption risk If high-end substrates migrate toward new materials/new architectures and the company fails to keep up, heavy assets could become a burden The company discusses directions such as glass core and silicon bridge, but technology updates quickly

These risks are not theoretical. The strongest counterargument is very simple and powerful: Ibiden is not a Buffett-style asset-light, high-cash compounding company. It is a heavy-asset advanced manufacturer whose valuation is being lifted by the AI narrative. It may well be a good company in its track, but "good company" does not automatically mean "good investment." In FY2025, the headline surge in net profit benefited heavily from gains on sales of investment securities, while true distributable cash flow was still being swallowed by expansion. At the same time, the group includes ceramics/DPF assets affected by internal-combustion-engine and mature industrial demand, which lowers business purity and the sustainability of a pure-AI valuation.

If I were the short seller, I would say this: the market is now valuing Ibiden as a "pure AI substrate leader," while it is actually a hybrid manufacturer of "high-end substrates + mature ceramics business + heavy capital expansion + customer concentration." When valuation has already risen to nearly 9x PBR and more than 80x PER, any slowdown in AI demand growth, decline in customer advances, uneven expansion ramp, or retreat in market risk appetite could cause the stock to fall much more than fundamentals deteriorate. This counterargument is not extreme. It is exactly the argument that most needs to be taken seriously today.

What facts would make me admit that my current cautious view is wrong? There are four main ones. First, by around FY2028, new capacity ramps according to plan and high-end AI substrates deliver sustained ROIC above 15%. Second, strict-basis free cash flow turns clearly positive after the expansion peak, and Owner Earnings sustainably reach above JPY 120.0 billion to JPY 150.0 billion. Third, in high-end new-generation products, the company can do more than "capture initial launch share" and can still maintain clearly high margins and stable share after expansion. Fourth, management fulfills its commitment to "expand within operating cash flow without additional financing" and does not dress up cash returns by selling financial assets. If these facts materialize one after another, my current "Watch" view would turn more positive.

The biggest permanent-capital-loss scenario is not bankruptcy. It is buying at a high price and then suffering the double hit of "earnings are not bad, but valuation returns to normal." A simple asset-method stress test: if the market reprices the company to 3-4x PB, then based on FY2025 book value per share of JPY 1969.85, the corresponding share price would be only about JPY 5900-7880, a drawdown of about 55%-66% from the current JPY 17,570. This is not unimaginable among high-valuation AI-chain assets.

Compared with other opportunities, I do not think Ibiden is clearly superior to the index today. TOPIX is a broad benchmark for the Japanese equity market. On June 8, 2026, Nikkei 225's market-cap-weighted PBR was about 1.86x, and its price-weighted PBR was about 2.73x, while Ibiden was at 8.92x PB. This means you are paying a valuation far above the Japanese broad market for Ibiden, while receiving a company with a more complex business structure and more concentrated customers. Unless you have extremely high confidence in Ibiden's track position over the next five to ten years, it is not meaningfully superior to buying the index.

Compared with the risk-free rate, Ibiden's current appeal is also weak. On June 9, the Japan 10-year government bond yield had risen to about 2.73%, while Ibiden's current forecast dividend yield was only 0.20%. Even if one uses my more conservative Owner Earnings yield estimate above instead of dividend yield, the cash yield implied by the current price is not attractive. For conservative long-term capital, this combination of "high valuation, low visible cash return, and strong dependence on execution" is not worth chasing.

On comparison with "the strongest peer competitor," I would view it this way: Shinko Electric is the more natural Japanese comparable, but it has entered a privatization process, so public-market price information has less reference value. AT&S and Samsung Electro-Mechanics can serve as international valuation references, but their business mixes are not fully homogeneous. Even so, Ibiden currently does not look "clearly cheaper than peers." It looks more like one of the highly sought-after names under the AI theme. If I could hold only 5 assets, I do not yet think it qualifies for a core portfolio targeting steady returns.

Investment Checklist and Final Judgment

First, compress the core judgment into a checklist:

Checklist Item Conclusion
Can I understand this business? Pass, but only for the economic logic; technical understanding has a high threshold
Does it have long-term stable demand? Pass, but high-end substrate demand is strong while traditional ceramics demand is ordinary
Does it have a durable moat? Partial pass
Does it have pricing power? Partial pass
Can it generate stable free cash flow? Fail
Is its return on capital excellent? Uncertain
Is management trustworthy? Pass
Is capital allocation rational? Partial pass
Is the balance sheet solid? Pass
Is valuation below intrinsic value? Fail
Is the margin of safety sufficient? Fail
Would I feel comfortable holding it long term? Uncertain
What key facts would make me sell? Expansion returns falling short, worsening customer concentration, strict FCF staying negative for a long time
Am I tempted to buy only because of share-price momentum or market sentiment? This motive should be watched very carefully today

The essence reflected in this checklist is clear: operationally and technically, Ibiden deserves respect; valuation-wise, the current price does not. For a value investor, that is enough to make a decision.

【Final Rating】 Watch

【One-Sentence Investment Thesis】 Ibiden is a strong player in the AI high-end substrate track, but the current share price has already priced in too much future success in advance and lacks the margin of safety most valued in Buffett-style investing.

【Core Bull Case】

  • Electronics has become the group's core business. FY2025 electronics segment sales and profit improved significantly, and demand for high-end AI/HPC substrates is strong.

  • The company has a global leading position in high-end IC package substrates, and management believes its technology advantage in silicon bridge-type substrates can last at least until FY2030.

  • The balance sheet is solid. At the end of FY2025, the company was in a net cash position, the equity ratio was 57.3%, and interest coverage was high.

  • Management's posture on capital efficiency is improving. The reduction of policy shareholdings was completed ahead of schedule, and the company has clearly set a payout-ratio target and a principle of "expansion within operating cash flow."

  • If the JPY 500 billion expansion is executed according to plan and at high quality, the long-term earnings ceiling will rise significantly.

【Core Bear Case】

  • Current valuation is extremely high: about 84.6x PER and 8.92x PBR. This is no longer "buying a good company cheaply," but "paying a high price for perfect execution."

  • FY2025 net profit included large gains on sales of investment securities, so headline profit is above the sustainable level of operating cash returns.

  • The company is in a heavy capital expansion phase. Strict-basis free cash flow is not attractive, and Owner Earnings are far below what the market value implies.

  • Customer concentration is high. Intel/AMD/NVIDIA already accounted for more than half of FY2024 revenue, which amplifies industry volatility.

  • The company is not a pure AI substrate asset. It also includes mature ceramics/DPF businesses, so business purity and valuation purity do not match.

【Key Assumptions】

  • AI/HPC substrate demand remains strong for at least the next three to five years.

  • The JPY 500 billion expansion ramps according to plan, and customers support returns through advances/order visibility.

  • The company maintains a sufficiently long technology lead and yield advantage in high-end new products.

  • After expansion is complete, strict-basis free cash flow and Owner Earnings improve significantly.

【Fair Buy Price】 JPY 3500-6000. This is based on my Owner Earnings-derived neutral intrinsic value range of about JPY 4750-6870, with a necessary discount for manufacturing cyclicality, customer concentration, and expansion risk.

【Target Holding Period】 More than 10 years, but only if bought at a reasonable or cheap price. If the purchase price is wrong, a longer holding period does not automatically repair returns.

【Expected Annualized Return】 Buying at the current price of about JPY 17,570, and estimating under conservative, neutral, and optimistic cash-flow realization paths, the next ten years could roughly correspond to: about -14%/year in the conservative scenario, about -6%/year in the neutral scenario, and about +2% to +4%/year in the optimistic scenario. This is not a short-term forecast. It is a calculation based on different Owner Earnings growth paths and terminal valuation multiples. Its point is that upside has been severely compressed at the current price.

【Maximum Loss Risk】 If AI strength fades, expansion returns fall short, and valuation reverts to a more typical range for an excellent manufacturer such as 3-4x PB, the share price could return to about JPY 5900-7880, implying a drawdown of about 55%-66% from current levels. In a more pessimistic case, if earnings and valuation both contract, the decline could be deeper.

【Tracking Indicators】

  • Sales volume, ASP, and product-mix changes for high-end IC package substrates.

  • Ramp speed, yield, and customer qualification progress for new capacity.

  • Whether electronics segment operating margin can stay above the mid-teens.

  • Whether strict-basis free cash flow and Owner Earnings continue to improve.

  • Whether customer advances and long-term supply arrangements persist.

  • Whether the revenue share of major customers such as Intel, AMD, and NVIDIA rises further.

  • Whether FY2026-FY2028 capital expenditure exceeds budget or schedule.

  • Whether the Philippine and other overseas bases incur further impairments.

【Valuation Range】

  • Conservative intrinsic value range: JPY 2200-3200
  • Fair intrinsic value range: JPY 4750-6870
  • Optimistic intrinsic value range: JPY 9200-13370

【Ideal/Fair Buy Price】 JPY 3500-6000

This report is based on public information and does not constitute investment advice. Markets carry risk; invest with caution.

INTCAMDNVDA6967ATX009150

IC SubstratesABF PackagingAI HardwareJapanese ManufacturingSemiconductor Supply Chain
Reader Q&A10

Baillie Framework · Ten Questions for Growth Investing

10

Hunting ten-year five-baggers among great growth stocks — pressing the upside question: "Can it get much bigger?"

Baillie Framework · Ten Questions for Growth Investing — score profile: 44/100 total Ceiling 6/10 · Revenue 2x 5/10 · Next engine 4/10 · Moat 6/10 · Reinvention 5/10 · Management 4/10 · Customer need 5/10 · Unit economics 5/10 · 5x path 2/10 · Blind spot 2/10 0510 How large is its market ceiling? Is it expanding an existing pie, or creating an entirely new market? — 6/10 Ceiling 6 Can its revenue at least double over the next five years? Will growth be driven mainly by volume, price, or new businesses? — 5/10 Revenue 2x 5 Five years from now, what will take over as the next growth engine? Does this “second curve” exist today? — 4/10 Next engine 4 What is its core competitive advantage? Will this moat widen or narrow over the next three to five years? — 6/10 Moat 6 If the core business is disrupted, does it have the DNA to reinvent itself? How does it handle mistakes and bad news? — 5/10 Reinvention 5 Does management, especially the founder, have a long-term perspective and deep alignment with the company? Is it willing to sacrifice current profit for five to ten years out? — 4/10 Management 4 If it disappeared tomorrow, how much would customers miss it? Is its growth sustainable and not dependent on harming society or regulation? — 5/10 Customer need 5 What are the unit economics of this business, including gross margin and incremental returns? Do they improve or deteriorate with scale? Where does the money it earns go? — 5/10 Unit economics 5 What conditions must hold simultaneously for it to rise fivefold in ten years? Are those conditions realistic? What expectations are embedded in today's share price? — 2/10 5x path 2 Why has the market not realized all this yet? Is it too hard to understand, too easy to dismiss, or too far out? What could become the narrative inflection point? — 2/10 Blind spot 2
  • How large is its market ceiling? Is it expanding an existing pie, or creating an entirely new market?6/10

    Conclusion: Ibiden's market ceiling is high enough, but it is not creating an entirely new market. It is expanding and moving upmarket within the existing pie of high-end IC package substrates, while trying to capture the highest-value layer in AI/HPC. This is a positive answer to Baillie Q1: the track has both slope and runway. But it is not a boundless new platform market; demand still ultimately comes from chip shipments and advanced packaging upgrades for AI GPUs, ASICs, server CPUs, data center switches, and similar products.

    The market opportunity is not small. External estimates show that the global advanced IC substrates market is expected to grow from about USD 18.1 billion in 2026 to about USD 33.5 billion in 2033, implying a CAGR of about 9.2% from 2026-2033, driven by AI accelerators, HBM, chiplets, and 2.5D/3D advanced packaging. Another estimate cited from Yole points in the same direction: the combined market for advanced substrate technologies is expected to reach about USD 31 billion by 2030, mainly driven by AI/HPC. This is not a trillion-dollar software market, but for Ibiden, with FY2025 group sales of JPY 416.2 billion and electronics sales of JPY 243.3 billion, it is already large enough to support several years of expansion.

    The key is how the pie gets larger. AI chips do not simply mean selling a few more ordinary boards. They make substrates larger, more complex, higher-layer, and more demanding on yield. In its FY2025 presentation materials, Ibiden says demand will grow for AI GPUs, AI ASICs, server CPUs, and data center switching ICs, and that the production load for AI server substrates, converted on an SAP basis, will rise from the CY2024 baseline of 1.0 to 2.5 times in CY2028. That means the ceiling comes not only from volume, but also from higher unit value: larger size, finer lines, more layers, embedded components, silicon bridge / glass core routes, and similar technologies all increase the importance of high-end substrates.

    Ibiden's view of this ceiling is also reflected in its capital spending. The board has approved about JPY 500.0 billion of investment in the electronics business from FY2026-FY2028 to expand production of high-performance IC package substrates for AI servers and high-performance servers, and the FY2025 presentation materials set an FY2030 target centered on the electronics business of JPY 1 trillion in net sales and JPY 300.0 billion in operating profit. If that target is achieved, Ibiden's revenue ceiling is clearly much higher than the old image of a PCB/ceramics manufacturer.

    But the boundary also needs to be clear: this is not a platform business creating a new market. It is a difficult manufacturing link inside the semiconductor advanced packaging value chain. Competitors still include Unimicron, Shinko, Samsung Electro-Mechanics, AT&S, and others. AT&S also publicly positions itself as a supplier of high-end PCBs and IC substrates, which shows that this market is not Ibiden's alone. Customers will also pursue multi-sourcing, and if AI capital spending slows, substrate demand will fluctuate with it.

    So the answer to Q1 is: the market ceiling is high, and the high-end portion is being lifted materially by AI/HPC; but Ibiden is expanding and upgrading an existing pie, not creating a new one. That is enough to support the imagination of an excellent growth manufacturer, but not enough to treat it as a boundless, asset-light, winner-take-all super-growth platform.

    Jun 9, 2026
  • Can its revenue at least double over the next five years? Will growth be driven mainly by volume, price, or new businesses?5/10

    Conclusion: There is a path, but a five-year revenue doubling is not low-risk or already locked in. Starting from FY2025 net sales, IBIDEN's latest financial results disclosed FY2025 net sales of JPY 416.201 billion and FY2026 guidance of JPY 500.0 billion; the company's FY2025 briefing also raised the target to FY2030 net sales of JPY 1 trillion and operating profit of JPY 300.0 billion. So from FY2025 to FY2030, the target implies about 2.4 times revenue; from FY2026 guidance, it also requires another doubling by FY2030. The answer is: the upside path exists, but delivery depends on capacity expansion, demand, and yield all going well at the same time.

    The first driver is volume, not simple price hikes. The company has already approved about JPY 500.0 billion of capital investment in high-performance IC package substrates in the electronics business from FY2026-FY2028, precisely to expand substrate capacity for AI servers and high-performance servers; the first phase is an investment of about JPY 220.0 billion in Gama Plant Cell6, scheduled to begin operation and mass production sequentially from FY2027. In other words, the core premise of revenue doubling is that high-end ABF/FC-BGA substrate capacity actually comes online and is absorbed by demand from AI GPUs, AI ASICs, server CPUs, data center switches, and similar products.

    The second driver is product mix and ASP improvement, but this is closer to high-end mix improvement than luxury-style arbitrary pricing power. Management said in the FY2025 Q&A that the margin upgrade came from AI-related demand being stronger than expected and a higher share of high-value-added products. The same Q&A also explained that price improvement mainly comes from product mix improvement, maintaining current selling prices, and passing through material costs with customer agreement, while price negotiations will still be kept within a reasonable range. Put differently, future revenue growth will include an ASP contribution, but the essence is selling more, larger, more complex, higher-layer high-end substrates, not pushing revenue up through broad price increases.

    The third driver is new applications or adjacent businesses. The company said it has started mass-production-level support for IC package substrates used in switches, and that it has received switch substrate orders from a major GPU customer. This matters because it extends IBIDEN from GPU/ASIC substrates further into data center networking chip substrates. But I would not treat it as a completely independent new business; it is more like an application expansion on the same high-end IC substrate platform. Ceramics, EV battery safety materials, graphite, and similar areas can provide supplements, but in the short term they are not enough to become the main engine for group revenue doubling.

    The biggest constraint is also clear: the company itself says the FY2030 target is achievable on the condition that it executes the JPY 500.0 billion capital investment as planned. It also acknowledges that, because of labor and lead-time constraints, further expansion beyond the current plan before FY2028 would be difficult. Therefore, the judgment for Q2 should be: a five-year revenue doubling is supported by an official target and real demand. The main driver is volume growth in AI/HPC high-end substrates, followed by ASP improvement from mix upgrade, while new businesses are only incremental upside. But this is a capital-heavy, customer-dependent, execution-dependent path and should not be treated as already done.

    Jun 9, 2026
  • Five years from now, what will take over as the next growth engine? Does this “second curve” exist today?4/10

    Conclusion: Ibiden's second curve already exists today, but it is not an entirely new business line. It is the extension of the high-end IC substrate platform from AI GPUs into AI ASICs, switches, silicon bridge-type substrates, and advanced package substrates with larger size, higher layer counts, and more complex structures. In other words, the growth engine that is most likely to take over five years from now is still inside the electronics business, rather than ceramics, DPF, or some new business that has not yet been commercialized.

    The evidence is fairly solid. In February 2026, the company approved about JPY 500.0 billion of investment in the electronics business from FY2026-FY2028 to expand capacity for high-performance IC package substrates, with the aim of supporting the FY2030 plan; the first phase includes Gama Plant Cell6 for high-performance server substrates, with operation and mass production planned to begin sequentially from FY2027. This is not conceptual R&D. It is an expansion plan that has already entered board approval and fixed-asset investment. The FY2025 briefing also breaks down the next round of demand more clearly: beyond AI GPUs, ASICs for AI inference, server CPUs, and data center switching ICs are all becoming higher performance; among the company's expansion projects, Gama Cell6 is mainly for AI ASICs, while Ono Cell8 is mainly for AI GPUs. This shows that the second curve at least already has product definition, customer demand, and capacity landing points.

    The more important point is the technology roadmap. Silicon bridge, glass core, embedded components, photoelectric conversion, and similar technologies mentioned in the report are not all near-term revenue-generating second curves, but they point in the same direction: high-end substrates are moving from ordinary ABF/FC-BGA toward larger size, more layers, low-power high-speed interconnect, and connection structures for advanced packaging. In the FY2025 Q4 Q&A, management explicitly said that the back-side interconnect technology for silicon bridge-type substrates is extremely difficult and that the company believes its technological advantage can be maintained at least until FY2030. The same Q&A also said the company had begun mass-production-level support for IC package substrates used in switches in FY2025 and had received switch substrate orders from a major GPU customer. This carries more weight than simply talking about future technology trends, because it is already connected to customer orders and mass-production support.

    But the distinction needs to be honest: this second curve is an adjacent growth curve, not a business model leap. It still depends on a small number of leading semiconductor customers, capital-heavy expansion, yield ramp-up, and judgments about the advanced packaging roadmap. From today's vantage point, it exists, but it has not been fully derisked. The real items to track are three: whether Gama/Ono new capacity can ramp on the post-FY2027 timetable, whether ASIC/switch/silicon bridge revenue can take over from GPU substrates as a new growth layer, and whether strict free cash flow can turn positive after expansion. If these happen, Ibiden's second curve will move from visible to verifiable; if not, it will just be a capital-heavy amplification of the current AI substrate boom.

    Jun 9, 2026
  • What is its core competitive advantage? Will this moat widen or narrow over the next three to five years?6/10

    Conclusion: Ibiden's core competitive advantage is the combination of process yield, customer co-development, qualification, and stable expansion in high-end IC package substrates, not brand, network effects, or a single patent. Over the next three to five years, the moat in high-end AI/HPC substrates is likely to widen; but at the group level, it will show a split between a widening frontier business and narrowing mature substrate and ceramics businesses, so it should not be understood as a quasi-monopoly.

    Its hardest barrier is on the manufacturing side. The package substrates needed for advanced servers, GPUs, and ASICs are becoming larger, higher-layer, and finer-line, and what customers are really buying is stable mass-production capability. In its Integrated Report, Ibiden lists cutting-edge semiconductor IC package substrates as a core growth area and says it has a global leading share in that field, while emphasizing that it relies on more than 40 years of accumulated technology, experience, and advanced production facilities to meet requirements for large size, multiple layers, and high-density fine lines. This type of capability cannot be replicated by simply buying a few pieces of equipment. Yield ramp-up, process detail, failure analysis, and customer qualification keep latecomers outside.

    Customer relationships are the second layer of the moat. The report says Intel, AMD, and NVIDIA together accounted for about 52% of revenue in FY2024. That points to customer concentration risk, but it also shows that Ibiden has entered the supply chains of the world's most demanding chip customers. Once a high-end substrate passes platform qualification, customer switching is not a simple price comparison. It involves requalification, yield volatility, delivery delays, and risks to system launch schedules. Management also said in the FY2025 Q&A that at the initial introduction of new-generation products, the company can sometimes obtain close to 100% share, though it may decline by 20%-30% after about 3-6 months. This sentence is important: it shows that first-launch generational advantage is real, but not permanent exclusivity.

    Over the next three to five years, the high-end moat is likely to widen because the technology roadmap is moving in a harder direction. In the same Q&A, the company said the back-side interconnect technology for silicon bridge-type substrates is extremely difficult and judged that its technological advantage can be maintained at least until FY2030. At the same time, the company has announced about JPY 500.0 billion of three-year capital investment in high-performance IC package substrates in the electronics business from FY2026-FY2028, with the first phase centered on expanding capacity for high-performance server substrates at Gama Plant Cell6. The logic is that when products become more complex, customer delivery windows tighter, and capacity scarcer, suppliers that can ramp on time and deliver stably become more valuable.

    But this moat has boundaries. First, customers will multi-source and will not allow a single substrate supplier to hold absolute bargaining power for long. Second, the competitors are not small players. AT&S calls itself a leading supplier of high-end PCBs and IC substrates; Samsung Electro-Mechanics has also shown mass-produced high-end AI/server FCBGA technology; and Shinko Electric also provides flip-chip package substrates for high-performance chips used in PCs, servers, and other products. Third, the report has already pointed out intensifying competition in mature PC/general server substrates and impairment of the Philippine electronics business and idle assets, which shows that the moat in low-end or older platforms is narrowing.

    So my judgment is: Ibiden's moat in high-end substrates for AI servers, ASICs, GPUs, switches, and similar products will widen over the next three to five years, but the group's overall moat is only moderately widening and becoming more bifurcated. The real items to track are three: whether new capacity ramps on time, whether high margins in the electronics business can be maintained, and whether key customers continue to support expansion with orders and prepayments. If these happen, it will move from a strong manufacturer closer to a scarce infrastructure supplier; if yield, share, or customer prepayments fall, the moat will be eroded in reverse by heavy capital requirements and peer capacity expansion.

    Jun 9, 2026
  • If the core business is disrupted, does it have the DNA to reinvent itself? How does it handle mistakes and bad news?5/10

    Conclusion: It has reinvention DNA, but not the kind that can switch tracks at any time. IBIDEN's strength is repeatedly migrating the same capabilities in materials, sintering, precision processing, lamination, and yield control into new demand: from a power company founded in 1912, to using electricity to enter electric furnace/carbide manufacturing, to betting on PWB, specialty carbon, and ceramic fibers after the oil crisis, and later developing IC package substrates and DPF. On its R&D page, the company summarizes this logic as making products needed by future society with core technologies, and explicitly says that after the oil crisis hit the building materials main business, management shifted resources to three priority products. That shows it is not a manufacturer that only defends old production lines, but one with a history of reallocating technology assets under crisis.

    Today, the direction of reinvention is also clear: if PC/general server substrates are eroded by price competition, the company is not continuing to fight in the old battlefield. It is concentrating capital and customer collaboration on high-end AI/HPC IC package substrates. The official February 2026 announcement says that about JPY 500.0 billion will be invested in the electronics business from FY2026-FY2028 to achieve the FY2030 target and expand capacity for high-performance server substrates. The FY2025 briefing also maps the future technology route toward larger size, more layers, embedded components, glass core, silicon bridge, and photoelectric conversion. This is positive evidence: it knows the core business will keep being rewritten by next-generation packaging architectures, so it is proactively upgrading old substrate capabilities into advanced packaging platform capabilities.

    But this reinvention DNA deserves a discount. IBIDEN's migrations mostly remain within the main channel of materials/electronics/ceramics manufacturing. They are not jumps from hardware manufacturing into a software platform. Each reinvention requires large CapEx, customer qualification, and years of ramp-up. If the value of advanced packaging in the future shifts sharply from organic substrates to silicon interposers, glass substrates, or other architectures, whether it can keep up cannot be inferred from history alone. Management said in the FY2025 Q&A that the technological advantage of silicon bridge-type substrates can be maintained at least until FY2030. That sounds more like a window of leadership than permanent immunity from disruption.

    I would rate its attitude toward mistakes and bad news above average: it does not simply gloss over problems, but it is not maximally transparent either. The FY2025 results announcement disclosed directly that competition would intensify in IBIDEN Philippines' main PC-related products and that significant demand growth was hard to see, so the company conservatively revised the business plan and recognized an impairment of JPY 10.630 billion. The same page also disclosed impairment of idle assets across electronics, ceramics, and other segments. This shows that management is willing to write down low-expectation assets rather than force the balance sheet value to stand. An earlier Financial Review also listed the Philippine impairment as a key audit matter, citing slower-than-expected recovery in PC demand and intensified price competition caused by product commoditization. Such disclosures are valuable bad-news front-loading for investors.

    So the judgment for Q5 is: IBIDEN has manufacturing-platform reinvention capability and can recognize impairments, exit, or reallocate capital when bad news appears; but its speed of reinvention is constrained by heavy assets and the roadmaps of a small number of large customers. In the Baillie framework, this is a positive factor but not a maximum-score one. The real question to track is whether, when high-end AI substrates also enter mature competition, the company can again move its core processes onto the next profit curve as it historically moved from electricity, building materials, PCB, and DPF into IC substrates; and whether it continues to identify risks early and write down low-efficiency assets promptly.

    Jun 9, 2026
  • Does management, especially the founder, have a long-term perspective and deep alignment with the company? Is it willing to sacrifice current profit for five to ten years out?4/10

    Conclusion: IBIDEN management's long-term perspective is real, and it is willing to sacrifice current free cash flow for the 2030 high-end IC substrate opportunity; but this is not a founder owner-operator, and there is no controlling-level personal shareholding alignment, so Q6 is a “pass, but not a strength.”

    IBIDEN is an old-line Japanese manufacturer founded in 1912. Current President & CEO Koji Kawashima is an internally developed professional manager, not a founder or family controller. The Integrated Report 2025 discloses that Koji Kawashima joined the company in 1987, became President & CEO in 2024, and holds 39,200 shares; Chairman Takeshi Aoki holds 89,500 shares. Even after adjusting for the 1:2 stock split on January 1, 2026, the two together still hold less than 0.1% of the 281,721,114 shares outstanding. This shows management has industry experience and organizational continuity, but not the strongly bound structure of having most of their wealth tied to the company.

    The positive evidence for long-termism is mainly in capital allocation. The board has approved about JPY 500.0 billion of investment in high-performance IC package substrates from FY2026-FY2028 to support the FY2030 plan. The FY2025 financial results briefing also raised the FY2030 target to JPY 1 trillion in net sales and JPY 300.0 billion in operating profit. This is not short-term EPS optimization. It is a typical bet on heavy capital, yield, and customer qualification: the FY2026 capital expenditure plan reaches JPY 210.0 billion, which will clearly pressure current free cash flow.

    More importantly, management is not pursuing growth entirely through external financing. In the FY2025 Q&A, the company said the JPY 500.0 billion investment would basically be supported through arrangements such as customers bearing investment costs and prepayments, and that it currently does not consider additional financing. The same Q&A also said price negotiations need to consider long-term customer relationships. This shows the team is willing to sacrifice part of short-term profit maximization in exchange for customer share, capacity position, and control of the technology roadmap over the next five to ten years.

    The deductions are also clear. Although the compensation system includes about 15% share-based compensation and emphasizes medium- to long-term corporate value, the compensation structure is still mainly fixed compensation and bonuses. The shareholder return policy also sets a 20% payout ratio and progressive dividends, indicating that management is balancing growth investment, financial discipline, and current shareholder returns rather than reinvesting all-out against market pressure in the style of a founder. The 81.2% reduction in policy shareholdings is a signal of improved capital efficiency, but it proves governance progress, not deep binding of management's personal wealth.

    So my judgment is: IBIDEN management is credible, long-term, and manufacturing-literate, and it is indeed willing to sacrifice short-term cash flow for FY2030 capacity and technological advantage; but evidence is insufficient for what the Baillie framework values most: high founder ownership, sharing the same pain as shareholders, and persisting for ten years against market pressure. Q6 can receive a moderately positive assessment, but not a high score.

    Jun 9, 2026
  • If it disappeared tomorrow, how much would customers miss it? Is its growth sustainable and not dependent on harming society or regulation?5/10

    Conclusion: Customers would miss Ibiden, but not to the point where everything stops without it. In validated high-end AI/HPC IC substrate projects, it is a high-switching-cost supplier, and its growth mainly comes from real demand for compute infrastructure rather than regulatory arbitrage or revenue earned by harming society. But the deductions are clear: customers will multi-source, electronics expansion consumes energy, chemicals, and capital, and the old ceramics/DPF business still carries long-term structural pressure from the internal-combustion-engine chain.

    In terms of how painful its disappearance tomorrow would be, Ibiden's strongest points are not brand, but customer qualification, yield, delivery reliability, and co-development status. High-end FC-BGA/IC package substrates are critical to AI GPUs, ASICs, switches, and similar products, with specifications moving toward larger size, more layers, and finer lines. The company's product page also describes demand for larger size, more layers, and finer patterns as PCs and data center servers become higher-performance. Harder evidence from management's Q&A shows that when new-generation products are introduced, the company can usually obtain close to 100% share, which may then decline by 20-30% after 3-6 months; for silicon bridge-type substrates, the company believes its technological advantage can be maintained at least until FY2030. This means customers would miss it in the short term, especially high-end projects from NVIDIA, AMD, Intel, and others that have already qualified Ibiden's processes and capacity, because replacing the supplier would bring qualification, yield, and ramp-up risks.

    But this kind of being missed is not consumer-brand irreplaceability. Customers will rationally multi-source, and the report also lists strong peers such as Unimicron, Shinko, Samsung Electro-Mechanics, and AT&S. Management itself also acknowledges that first-launch share will fall back after several months. The moat for this kind of industrial product is more like qualified-supplier status plus yield learning plus capacity lock-in, not permanent exclusivity. If customer platforms switch, design routes change, or competitors pass qualification in the same generation, Ibiden's share and pricing will be renegotiated.

    On sustainability, I lean toward the judgment that business demand is sustainable, but growth quality deserves a discount. On the positive side, in February 2026 the company announced an about JPY 500.0 billion capital investment plan for high-performance IC package substrates in the electronics business from FY2026-FY2028, aimed at expanding substrates for AI servers and high-performance servers. This is not subsidy arbitrage or exploiting gray regulatory gaps; it is tied to the upgrade of compute infrastructure. Management also said this expansion would basically be financed through customers bearing investment costs and paying part of the funds in advance, which shows there are customer demand and commercial contracts behind the expansion rather than self-excited capacity building.

    On the social and regulatory dimension, I do not see core evidence that it makes money by harming society. On the contrary, the company's environmental policy commits to lifecycle carbon neutrality, reducing energy/water/resource use, and complying with environmental regulations in Japan and overseas; FY2024 ESG data disclosed 0 environmental violations and fines exceeding USD 10,000, with ISO 14001 certification coverage of 100% for IBIDEN itself and 92% for the group. That said, ESG risk should not be ignored: the company's FY2024 Scope 3 emissions reached 1.269 million tonnes CO2e, and water, chemicals, and hazardous waste all have real footprints. The faster electronics substrate expansion proceeds, the more these pressures need to be continuously verified by data. In the ceramics business, DPF and automotive holding materials are driven partly by strict environmental regulations, and the company says they are used to improve particulate collection and purification performance; at the same time, they remain tied to the internal-combustion-engine chain and may be compressed over the long term by electrification and changes in emissions regulation. Overall, I would give Q7 a positive but not perfect assessment: customer pain would be high, and growth is not a dirty business; but it is not irreplaceable, nor is it a low-footprint, low-regulatory-risk, asset-light compounding model.

    Jun 9, 2026
  • What are the unit economics of this business, including gross margin and incremental returns? Do they improve or deteriorate with scale? Where does the money it earns go?5/10

    Conclusion: Ibiden's unit economics are good in the electronics business, but at the group level this is not asset-light compounding. High-end IC substrates can bring higher value added and decent segment margins. The real issue is that profit has to be reinvested continuously into plants, equipment, depreciation, and yield ramp-up. As scale expands, accounting margins may improve, but free cash flow first deteriorates and can improve only after new capacity enters production at high utilization.

    Start by separating the metrics. The company does not disclose gross margin for the electronics business separately, so group gross margin cannot be directly equated with AI substrate gross margin. According to official FY2025 results, group sales were JPY 416.201 billion, gross profit was JPY 131.668 billion, and operating profit was JPY 62.027 billion, corresponding to a group gross margin of about 31.6% and operating margin of 14.9%. But the electronics segment is stronger: FY2025 electronics sales were JPY 243.316 billion and segment profit was JPY 45.248 billion, implying a segment margin of about 18.6%; the ceramics segment was about 9.3%. So the core quality of this business comes from high-end electronics substrates, not the group average.

    The upside in incremental economics comes from product mix. High-performance IC package substrates for AI GPUs, ASICs, and switches are larger, more complex, and higher-layer, with high thresholds for customer qualification and yield. The company also says profitability improved because of a higher share of high-value-added products. Management further stated that pricing is supported by product mix improvement, maintaining selling prices, and passing through material costs. If this line scales smoothly, scale can indeed improve margins: the company's FY2026 guidance is sales of JPY 500.0 billion and operating profit of JPY 90.0 billion, implying a group operating margin of 18%, above FY2025's 14.9%.

    But this is not a business that becomes lighter as it gets bigger. FY2025 depreciation and amortization already reached JPY 66.236 billion, including JPY 55.202 billion in the electronics segment. On a cash-flow basis, FY2025 operating cash flow was JPY 106.407 billion, but cash outflows for purchases of property, plant and equipment plus intangible assets totaled about JPY 107.244 billion, leaving strict FCF at roughly negative. This shows that accounting profit is already decent, but there is not much freely distributable cash for shareholders.

    The money it earns mainly goes to capacity expansion, not large buybacks or high dividends. The company has announced about JPY 500.0 billion of capital expenditure for high-performance IC package substrates from FY2026-FY2028, and the FY2026 single-year capital investment plan is listed in the briefing materials as JPY 210.0 billion. The uses are mainly Gama Plant Cell6, Ono Plant Cell8, and expansion of existing plants, serving high-end substrate demand for AI ASIC/GPU and similar products. The company also says FY2026-2028 operating cash flow will be prioritized toward growth investment, while also considering debt reduction and a target dividend of about 20%.

    So the judgment for Q8 is: unit economics are good at the product level and conditionally good at the company level. As long as high-end capacity is fully loaded, yield ramps smoothly, and customer demand continues, larger scale will improve margins; but during the peak expansion period, larger scale first brings heavier depreciation, higher capex, and tighter cash flow. Ibiden is a high-quality capital-intensive manufacturer, not an asset-light cash compounding machine wrapped in an AI narrative.

    Jun 9, 2026
  • What conditions must hold simultaneously for it to rise fivefold in ten years? Are those conditions realistic? What expectations are embedded in today's share price?2/10

    Conclusion: A fivefold rise in ten years is not impossible, but at today's price it is no longer an opportunity that the market has not seen. It requires AI substrates, capacity expansion, margins, cash flow, and valuation multiples to almost all line up on the bull side at the same time. Based on the June 9, 2026 close of about ¥18,595, market cap of about ¥4.76T, TTM PE of about 79x, and forward PE of about 75x, a fivefold rise would mean a share price of about ¥93,000 and a market cap of about ¥23-24T. That target requires Ibiden not only to achieve the FY2030 plan, but also to keep being priced by the market after FY2030 as a scarce AI infrastructure asset.

    At least four conditions have to hold simultaneously for a fivefold outcome. First, demand for high-end AI/HPC IC substrates must keep expanding for many years, not merely reflect restocking or capacity tightness in 2026-2028. The company's current growth logic has a real foundation: the FY2026 guidance in the official FY2025 materials is sales of ¥500.0B, operating profit of ¥90.0B, and capital investment of ¥210.0B, clearly a step up from FY2025 sales of ¥416.2B and operating profit of ¥62.0B. But this is still only the first stage of moving from several hundred billion yen of revenue toward a larger scale.

    Second, the ¥500B expansion must be executed with high quality. The company has announced about ¥500B of three-year capital investment in high-performance IC package substrates in the electronics business from FY2026-FY2028, and in its FY2025 briefing it raised the FY2030 target to net sales of ¥1T, operating profit of ¥300B, and an operating margin of 30%. That is strong, but note that it only pushes the company to the level of trillion-yen revenue and three-hundred-billion-yen operating profit. If terminal valuation returns to a more normal 25-40 times net profit range for an excellent manufacturer, that target alone still struggles to support a ¥23-24T market cap.

    Third, cash flow after expansion must truly be released. The core of Q9 is not whether it will grow, but whether growth can turn into cash that shareholders can take out. The report has already pointed out that the strict FY2025 FCF proxy was close to zero, and FY2025 profit attributable to owners of parent included gains on sales of investment securities, so it cannot be extrapolated mechanically. If the future remains a pattern of profit growth being swallowed by capital expenditure, depreciation, and working capital, the market will eventually reprice it from a scarce AI asset back into a capital-intensive manufacturer.

    Fourth, valuation multiples cannot fall sharply. At the current PE of about 75-80 times, the market is already buying the future at a very high multiple. If PE is still close to 75 times ten years from now, about fivefold net profit growth can explain a fivefold market cap. But if terminal PE returns to around 30 times, net profit would need to grow roughly a dozen-plus times to be enough. That is more aggressive than the earnings step implied by the company's FY2030 target. The more realistic framing is that a fivefold rise requires a large outperformance versus the FY2030 plan, sustained high-end share, positive free cash flow, and no long-term valuation discount from the market.

    Are these conditions realistic? Partly realistic at the business level, but they should not be treated as the base case for investment return. Management said in the Q&A that achieving the FY2030 plan depends on executing the ¥500B investment as planned, but it also acknowledged that further expansion before FY2028 would be constrained by labor and lead times. It also said the technological advantage of silicon bridge-type substrates can be maintained at least until FY2030. This shows the company is not an empty story, and its technology and customer position are real; but achieving the FY2030 target and rising another fivefold from today's price are not the same thing.

    Today's share price embeds the expectation that AI substrate demand remains strong for a long time, Ibiden maintains high-end share, the ¥500B expansion ramps smoothly, operating margin moves toward 30%, strict FCF is released after expansion, and the market is willing to sustain a high valuation for years. That expectation is already full. Against the report's reasonable buy-price anchor of about ¥6,000, the current ¥18,595 looks more like paying upfront for a near-perfect long-term story than the kind of position in the Baillie framework where the market has not yet understood the company and the ten-year fivefold odds are attractive.

    Jun 9, 2026
  • Why has the market not realized all this yet? Is it too hard to understand, too easy to dismiss, or too far out? What could become the narrative inflection point?2/10

    Conclusion: The market has not missed IBIDEN's AI substrate narrative. It has already recognized it and assigned a fairly high price. The real perception gap is whether the FY2030 expansion can pass through yield, customer share, depreciation, and capital spending, and ultimately turn into sustained free cash flow. So this is not a typical opportunity where the market does not understand the story. It is more like an opportunity where the story is visible, but cash validation is not yet visible. If the market has a blind spot, it is mainly that some capital may treat it as an asset-light AI bottleneck and underestimate that it remains a capital-intensive, cyclical, customer-concentrated advanced manufacturer.

    The evidence is direct: as of the 2026-06-09 close, IBIDEN traded at about ¥18,595, with a market cap of about ¥4.76T, TTM PE of about 79x, and forward PE of about 75x. That is not the pricing of a forgotten asset. More subtly, the average target price from 17 analysts aggregated by S&P Global is about ¥12,653, about 32% below the current price, which shows sell-side analysts also understand the AI substrate logic but are balancing high prosperity against high valuation. What the market has not fully priced is not that AI servers need high-end IC substrates, but whether IBIDEN can turn this expansion cycle into high ROIC, distributable cash flow, and longer customer lock-in.

    The company's own narrative is already clear enough: the board has approved about JPY 500.0 billion of investment in high-performance IC package substrates in the electronics business from FY2026-FY2028, aiming to support the FY2030 plan. The FY2025 briefing also says SAP demand for AI server substrates is expected to exceed industry supply, that the company plans to maintain high-end share through Cell6/Cell8 expansion, and that it will lift the CY24 capacity/load index to about 2.5 times by CY28. This shows the AI substrate bottleneck is no longer a secret in the basement; it is the main line being traded by both the company and the market.

    The genuinely difficult part is the second layer: whether this main line can upgrade from scarce capacity into shareholder compounding. In the Q&A, management said the condition for the FY2030 target to be achievable is executing the JPY 500.0 billion capital investment as planned, and also acknowledged that further expansion beyond the plan before FY2028 is constrained by labor and lead times. The same Q&A also said initial share for new-generation interposer-type substrates can approach 100%, but may decline by 20%-30% after about 3-6 months, while the technological advantage of silicon bridge-type substrates is expected to last at least until FY2030. This is where the perception gap lies: the technology lead is real, but the market still needs to confirm that it is not merely a short-term first-launch share grab, and that share, pricing, yield, and cash returns can remain strong after expansion.

    There are three types of positive narrative inflection points. First, Cell6/Cell8 begin mass production on schedule from FY2027, yield and utilization of AI ASIC/GPU/switch substrates ramp smoothly, and customer prepayments or long-term supply arrangements continue to support investment. Second, electronics margins remain high after new depreciation arrives, strict FCF shifts from being swallowed by capex to sustained positive territory, and the market begins to revalue it from an AI substrate expansion stock into a high-end packaging platform cash-flow company. Third, the FY2030 revenue and profit targets become not just slogans, but verifiable year by year through orders, capacity, yield, customer share, and Owner Earnings.

    The negative narrative inflection points are also clear: expansion delays or cost overruns, major customer share being diverted to Unimicron/Shinko/AT&S/Samsung Electro-Mechanics and others, a slowdown in AI capital spending, or profit growth continuing to be consumed by depreciation and working capital. If these happen, the market will recognize that it is not an undervalued scarce leader in AI substrates, but a capital-intensive manufacturing stock given a high multiple under the AI narrative. So the answer to Q10 is not that the market does not understand. It is: the market has already seen the first-layer story but has not confirmed second-layer execution; the real inflection point is not telling the AI story again, but turning the FY2030 expansion into auditable, sustainable, distributable cash flow.

    Jun 9, 2026
Ask about this report

Members can ask about this report; once answered it appears under "Reader Q&A" on this page. You can also highlight a passage in the text to ask about it directly.