Good morning, everyone. I am Koji Kawashima, Representative Director and President of Ibiden Co., Ltd.

Today, I would like to explain the consolidated financial results of Ibiden Co., Ltd. for the fiscal year 2025.

I will proceed with the explanation based on the presentation materials, please refer to them as we go. First, regarding our consolidated results for fiscal 2025, net sales were ¥416.4 billion and operating profit was ¥62.0 billion.

Compared to the previous fiscal year, we achieved increases in both sales and profits. In particular, in the Electronics segment, operating profit grew by approximately 70% year-on-year.

The main drivers were an overall improvement in the product mix and higher sales prices, which contributed significantly to profit growth in Electronics. On the other hand, in Ceramics, both sales and profits decreased, as you can see.

This was mainly due to sluggishness in the automotive market, particularly the slowdown in electric vehicles, as well as limited growth in internal combustion engines, which led to lower order volumes and consequently reduced sales and profits. Other businesses performed steadily, recording slight year-on-year growth.

Net profit came in at ¥63.7 billion, marking a record high. While business profit growth contributed, another factor was the gain on sales of investment securities, specifically the sale of Toyota Industries Corporation shares.

However, even excluding the gain on sale of investment securities and impairment losses, net profit derived from core operations confirmed solid year-on-year growth. Next, I would like to discuss our earnings forecast for fiscal 2026 and beyond.

In terms of business performance, there are currently uncertainties surrounding geopolitical risks in the Middle East. At present, we believe the direct impact on us will be minor, but as the situation remains unpredictable, this forecast basically does not incorporate potential impacts from Middle East risks.

For fiscal 2026, we forecast full-year net sales of ¥230.0 billion in the first half and ¥270.0 billion in the second half, totaling ¥500.0 billion. For reference, net sales of ¥220.7 billion in the second half of fiscal 2025 was our historical record high; therefore, we project setting new record highs in both the first and second halves of fiscal 2026.

Operating profit is forecast at ¥38.0 billion in the first half and ¥52.0 billion in the second half, totaling ¥90.0 billion. In the second half, the operating profit margin is expected to reach 19.3%, which will also be a record high.

Net profit is forecast at ¥23.0 billion in the first half and ¥35.0 billion in the second half, with the net profit margin expected to be in the range of 10% to 13%. Our foreign exchange assumptions for fiscal 2026 are ¥150 per US dollar and ¥158 per euro.

On the next slide, I will explain our full-year outlook in more detail. For the full year of fiscal 2026, our forecast is net sales of ¥500.0 billion, operating profit of ¥90.0 billion, and an operating profit margin of 18.0%.

Growth will be driven primarily by the Electronics segment. There are two main reasons: first, capacity expansion at new plants, particularly the Ono Plant, which began ramp-up operations from October of fiscal 2025.

Although ramp-up began as planned in fiscal 2025, capacity ramp-up slowed slightly. In fiscal 2026, we plan to accelerate this ramp-up steadily toward full capacity.

Second, the product mix is transitioning to next-generation products, mainly centered on AI, which increases the proportion of high-value-added products and enhances operating profit. On the other hand, Ceramics is expected to face continued tough conditions in fiscal 2026, with net sales and operating profit both forecast to decline.

Other businesses are expected to remain roughly flat year-on-year. Net profit is forecast at ¥58.0 billion.

While this represents a slight decline compared to fiscal 2025, excluding the previously mentioned extraordinary gain, net profit from core operations will continue to show solid growth. Both net sales and operating profit are projected to reach record highs.

first, in Electronics, full-year net sales for fiscal 2026 are projected at ¥330.0 billion, an increase of ¥20.0 billion compared to our previous disclosure on October 31. As shown in the breakdown, general-purpose servers, switching IC packages, and AI servers—mainly GPUs and ASICs—are expected to drive growth significantly.

On the right, the waterfall chart shows operating profit. Compared to our previous estimate of ¥57.0 billion disclosed on October 31, we now project ¥75.0 billion.

The largest positive factor is the improvement in product mix and average selling price (ASP). Both AI and general-purpose server products are shifting to next generations, enhancing added value while keeping sales prices relatively stable.

On the negative side, expenses will increase slightly due to ramp-up costs and rising prices of maintenance materials and components. Next is our medium-term business outlook.

Here we present our sales and operating profit forecast through fiscal 2030. We have upwardly revised the medium-term targets through fiscal 2030 compared to our previous disclosure, driven primarily by the strong expansion in Electronics.

At the beginning of February, we announced an additional investment of ¥500.0 billion split between the Ono Plant and the Gama Plant. Utilizing this ¥500.0 billion investment, we target an incremental sales increase of approximately ¥500.0 billion from current levels toward 2030, aiming for net sales of ¥1.0 trillion in fiscal 2030, led by Electronics.

Meanwhile, Ceramics and other businesses are projected to grow steadily at a modest pace. In terms of operating profit, we aim for ¥300.0 billion in fiscal 2030, corresponding to an operating profit margin of 30%, driven by Electronics.

By successfully executing the new capital investments and ramping up capacity reliably, we believe achieving ¥1.0 trillion in net sales, ¥300.0 billion in operating profit, and a 30% operating profit margin is well within reach. Moving on to cash allocation and capital expenditure.

On the left, from fiscal 2026 to 2028, operating cash flow—consisting of net profit from business operations, depreciation, and customer prepayments—is expected to total approximately ¥600.0 billion. Regarding allocation, first, we reduced interest-bearing debt by approximately ¥150.0 billion in fiscal 2025 and plan to continue debt reduction by about ¥50.0 billion over the next three years.

Second, for shareholder returns, we will implement progressive dividends targeting a dividend payout ratio of 20%. The remaining funds will be allocated to growth investments, including new plant construction, facility upgrades, R&D, human capital development, and corporate venture capital investments in technology development.

On the right, capital expenditure for fiscal 2026 is projected at ¥210.0 billion. While fiscal 2027 and 2028 are shown with a range, we currently forecast around ¥300.0 billion for fiscal 2027 and around ¥100.0 billion for fiscal 2028.

However, considering potential capacity shortages going forward, additional capital investments may be considered. Regarding strategic shareholdings, we previously set a target to reduce them by 50% by fiscal 2027; with the recent sale of Toyota Industries shares, our reduction rate has already reached 81.2%, putting us well ahead of schedule.

Next, I will discuss our market outlook and business strategy. First, regarding the electronics market, the left chart shows the AI market, where GPU-based AI is indicated in yellow and AI ASIC is shown separately in blue.

In either case, the AI market continues to see very strong demand, and we anticipate a CAGR of around 20%. The right chart shows the CPU market, excluding general-purpose PC CPUs, focusing on servers and data center switching ICs, which require high-end packages.

We project a CAGR of approximately 4%. Going forward, with the shift from training AI to inference AI, the number of CPUs used alongside GPUs is expected to increase, which should support steady growth in the CPU market as well.

on the left, we show SAP total demand. From a manufacturing process perspective, because manufacturing proceeds layer by layer, total manufacturing area directly correlates with capacity.

Taking fiscal 2024 as 100, the index reaches 220 in fiscal 2028. This growth is driven by both an increase in total unit volume and an increase in substrate area and layer count per package.

In particular, dark blue represents general-purpose servers, and yellow represents AI servers, mainly GPUs, with ASICs on top. For these three areas combined, taking fiscal 2024 as 100, demand is projected to grow 4 to 5 times by fiscal 2028, indicating robust demand growth.

On the right, our capacity expansion plan shows that taking fiscal 2024 as 1, fiscal 2028 capacity reaches 2.8 times, which indicates that capacity will remain tight relative to total demand. Considering that our current plan already includes the ¥500.0 billion investment for Ono Phase 2 and Gama Plant construction, adding further capacity beyond this through fiscal 2028 will be very challenging from a human resource standpoint.

However, given strong customer demand, we must consider how to address capacity needs beyond this timeframe.

the Gama Plant on the left and the Ono Plant on the right. The Gama Plant was previously planned as a dedicated plant for a major CPU customer, but following contract adjustments, it was officially decided to produce EMIB-T products here.

We will invest ¥220.0 billion to install equipment and aim for full mass production in fiscal 2028. For the Ono Plant on the right, about half of the facility was equipped up to fiscal 2025; we will invest ¥280.0 billion out of the ¥500.0 billion into the remaining half to bring the entire plant into mass production.

Partial mass production will begin in fiscal 2027, with full mass production starting in fiscal 2028. Furthermore, anticipating future capacity shortages, both Gama and Ono have adjacent land available where new buildings can be constructed, and we will consider utilizing these sites as needed.

this slide shows package substrates, especially high-performance computing package substrates. The orange lines in the center represent interposer-based packages, which are currently mainstream.

As shown at the top, die sizes are expanding and layer counts are increasing, leading to larger SAP requirements across all packages. Looking ahead, through the investment at Gama Plant, we will ramp up EMIB-T products, which use silicon bridges shown at the bottom center.

Instead of placing semiconductor dies on a large interposer, semiconductor dies are connected via silicon bridges embedded in the substrate. These silicon bridge-type products will advance alongside interposer-type products in the market.

Furthermore, package technologies will incorporate embedded passive components for stable power supply, as well as glass cores and co-packaged optics (CPO) architectures. We will pursue multi-directional development to address these technological advancements.

the left chart shows the powertrain market, which is almost identical to our previous forecast. The blue line in the middle separates internal combustion engine (ICE) vehicles below and fully electric vehicles (BEVs) above.

ICE vehicles will continue to decline gradually, while EVs will expand. In ceramics, we offer products for both ICE and EV applications.

For ICE, we will capture demand in emerging markets such as China and India while accommodating product mix changes to sustain sales and profits. For EVs, we will expand sales centered on NEB, which started ramping up last year, to contribute to profit growth this fiscal year.

On the right, we have newly added a chart for nuclear power. Anticipating rising power demand for data centers, we expect nuclear power will become increasingly necessary and plan to expand our graphite products for nuclear applications.

We have already concluded a long-term supply contract with a major US customer and view this as an area of potential growth. Regarding R&D activities, we plan to allocate approximately 7% of total net sales to R&D expenses.

Key R&D themes include NEB, where we are expanding safety components beyond current vent-related safety parts, actively collaborating with major Chinese customers. In specialized substrate areas, leveraging package technologies, we are developing robotics-related motor applications and ceramic substrate-based power modules for high-voltage power devices.

In new fields, we are conducting exploratory research in biotechnology and CO2 reduction in collaboration with universities and government bodies.

for fiscal 2025, we plan a year-end dividend of ¥15 per share, bringing the annual dividend to ¥30 per share. Based on our progressive dividend policy, our consolidated dividend payout ratio for fiscal 2025 was 53% on an adjusted basis.

For fiscal 2026, we plan to increase the dividend by ¥5 to an annual dividend of ¥35 per share, representing a payout ratio of 17%. We remain committed to stable shareholder returns.

This concludes my presentation.

Thank you very much. I have two questions.

First, regarding the medium-term sales and profit targets on page 9, could you share the customer mix image for fiscal 2027 and fiscal 2030? Specifically, how will the mix evolve among the leading GPU player, the top two CPU players, and others including ASICs?

Also, regarding the substantial increase in operating profit margin compared to previous expectations, is this driven by product mix improvement or higher average selling prices (ASP)? What factors have changed over the past few months to drive this margin increase?

Thank you for your question. Regarding product mix, for fiscal 2025 and 2026, we expect AI-related products, mainly GPUs, to account for roughly 50%, CPU-related customers for 35% to 40%, and the remainder from ASIC-related products.

Looking to fiscal 2030, the split will be roughly similar, with GPU-centric customers at just over 50%, CPU-centric customers around 30%, and the remaining 15% to 20% from other customers, including ASICs. Regarding operating profit margin, both product mix and ASP contribute.

Although I did not detail the Philippines earlier, Ibiden Philippines recorded an impairment, making the organization leaner and more cost-competitive. We will allocate high-layer CPU products to the Philippines and concentrate higher-value-added products in Japan, improving overall product mix.

In addition, generational transitions in AI GPUs bring higher value-add. As for ASPs, material cost increases are passed through, but beyond that, we have maintained stable pricing instead of price declines, leading to higher profitability.

Thank you. Following up on the medium-term outlook, how much visibility do you currently have?

For fiscal 2027, you previously mentioned gaining visibility through ASIC contracts and EMIB-T decisions. How much clarity do you have now toward 2030?

For fiscal 2027, as you noted, visibility is quite high, including contracted portions. Beyond that, targets are based on customer demand forecasts.

Based on current customer demand and inquiries for fiscal 2030, as long as we ramp up capacity steadily as planned, we believe these targets are achievable.

Thank you. My second question is regarding capacity.

To confirm, will the ¥500.0 billion investment be sufficient to achieve the fiscal 2030 sales and profit targets, or will additional capacity expansion beyond the ¥500.0 billion be needed? If additional capacity is needed, around when will decisions need to be made?

For the fiscal 2030 targets, executing the ¥500.0 billion investment will be sufficient to achieve these figures. However, as shown on the demand slide, customer demand exceeds our capacity.

Customers are already requesting additional capacity for fiscal 2028 and 2029, effectively wanting capacity as soon as possible. However, due to human resource and timeline constraints, expanding beyond our current plan before fiscal 2028 is difficult.

Therefore, any additional capacity expansion would likely occur in fiscal 2029 or later.

In that case, looking at the slide where AI server ASIC demand grows 4 to 5 times while capacity only increases 2.8 times, to bridge this gap toward 2029 and 2030, would you need to make decisions on further capacity expansion around 2027? If you make those investments, is there potential upside to the 2030 targets?

We are already discussing additional capacity with customers, but nothing is finalized. However, we will likely need to make decisions around 2026 or 2027.

Regarding whether 2030 sales could go higher with additional capacity, we must be cautious because securing the necessary engineering and personnel resources is a major challenge.

Are raw materials not a bottleneck for this plan?

Raw materials, such as glass, are currently tight. However, we expect supply capacity from suppliers to increase from fiscal 2027 onward, we do not anticipate materials being a bottleneck.

Understood. Thank you very much.

Thank you. This is Naito from Citigroup.

I also have two questions. First, regarding pricing and ASPs: you mentioned that ASPs have been maintained and material costs passed through.

Given tight industry supply, are you able to secure higher prices across a broader product range, and could this provide upside to sales?

Thank you. Regarding ASPs, price increases occur due to product mix improvements and passing through raw material cost increases through quarterly, semi-annual, or annual price negotiations.

However, our policy is not to opportunistically raise prices beyond reasonable levels just because capacity is tight. We prioritize long-term customer relationships and aim to set reasonable prices through dialogue.

Thank you. My second question concerns the package substrate roadmap on page 15.

As die sizes and layer counts expand, manufacturing becomes more complex. How do you view the competitive landscape with Taiwanese and other Japanese competitors?

Can Ibiden maintain its high market share?

For interposer-type packages, as substrates become larger and layer counts increase, yield management becomes more challenging for all manufacturers. However, with new product launches this year, our technological advantage will allow us to maintain a share close to 100% in the initial ramp phase.

Competitors may enter after 6 to 12 months, taking 20% to 30% share, but as next-generation products launch, we expect to regain near 100% share. For silicon bridge-type products like EMIB-T, embedding silicon bridges into the substrate is technically demanding.

We believe only one or two companies globally can manufacture this type of product, which will help us maintain strong competitiveness through fiscal 2030.

Understood. Thank you very much.

Thank you. I have two questions.

First, regarding the simultaneous ramp-up of the remaining space at Ono and the Gama Plant over the next three years with the ¥500.0 billion investment: how will you secure the required human resources, how much of the ¥500.0 billion capex will be customer-funded, and how will it be procured? Also, can EMIB-T only be produced at Gama, or can existing plants also produce it?

Is there any possibility of accelerating the fiscal 2027 mass production timeline using existing lines?

Regarding simultaneous ramp-up of both plants, we are utilizing our existing personnel and reallocating resources to manage the ramp-up. For the ¥500.0 billion investment, 100% will ultimately be recovered through business operations, and a portion is being received in advance as customer prepayments, which are currently being negotiated.

We are not considering additional external equity financing. Regarding EMIB-T production, it is not limited to Gama; existing plants, including Chuo Plant, also have limited capabilities for certain processes such as laser processing.

Overall, we will support EMIB-T by combining Chuo and Gama. Mass production will start from fiscal 2027 using Chuo and part of Gama.

Thank you. My second question is about short-term trends.

I understand that the utilization rate of flip chip package substrates was just below 80% in the January-March quarter. Given strong demand, this seems somewhat low.

Is there a mismatch between capacity and demand? Also, for fiscal 2026, sales are projected at ¥150.0 billion in the first half and ¥180.0 billion in the second half.

Does the second-half increase assume contributions from the new Ono facility, and how much can existing lines increase sales?

Utilization rates vary by plant. For example, Ogaki Plant, which produces AI products, is operating at nearly 100% utilization.

Meanwhile, Chuo Plant is running at around 60% to 70%. This is not unused spare capacity, but rather due to contract structures with specific customers where capacity cannot be freely shifted between customers.

Furthermore, different product lines require different equipment configurations. Regarding Ono, operations started in October as planned, but capacity ramp-up faced initial challenges in fiscal 2025.

We are steadily improving this in fiscal 2026, aiming for utilization to reach 90% in the second half, up from around 80% in the first half. Ogaki will maintain 100%, and Chuo will improve from 60% to 70% toward 80% to 90%, supporting our second-half sales targets.

Thank you. A quick follow-up on Chuo Plant: as demand grows, will you maintain dedicated lines for specific customers, or will you reallocate capacity to other customers?

Chuo Plant is also required to run EMIB-T processes, we must utilize its lines for EMIB-T as well. Therefore, we do not foresee significant changes to the current customer allocation.

Understood. Thank you very much.

This is Hirata from UBS Securities. Thank you.

I have two questions. First, regarding the fiscal 2026 plan: in Electronics sales, switching IC substrates are projected to grow significantly.

Could you explain the background and customer dynamics? Also, regarding the medium-term adoption of AI ASICs, could you discuss the number of customers and demand changes over the past 3 to 6 months?

Regarding switching substrates, we began mass production support in fiscal 2025. These are high-layer, large-area substrates essential for high-speed server networking.

Currently, we have one dedicated customer for switching substrates, but major GPU customers are also adopting switching substrates, we expect the switching substrate proportion to increase from about 10% toward 15%. For AI ASICs, large tech companies are developing proprietary chips.

Currently, we have confirmed business with two companies. Some ASIC customers also utilize EMIB-related packaging architectures, for instance through Intel routes, while the total number of end-customer engagements is broader, direct finalized customer relationships stand at two companies.

Thank you. Is the profitability of switching substrates comparable to AI products?

Yes, you can consider profitability to be roughly equivalent.

Thank you. My second question relates to AI inference.

As inference demand grows and CPU demand increases, if ARM-based CPUs expand significantly, can Ibiden capture that demand? Previously, Taiwanese substrate makers were strong in ARM-based processors; do you see business opportunities in ARM CPUs, for example if US cloud players use foundries?

Looking at overall architectural trends, as package complexity increases with higher performance, the trend is similar regardless of architecture. While there may be shifts in customer mix, we believe there is ample room for Ibiden to participate.

Thank you very much.

This is Akizuki from Nomura Securities. Thank you.

First, I would like to clarify a point regarding the next-generation GPU model, Rubin. You mentioned a 100% share this year, but previously you had 100% at launch followed by competitors entering after about 3 months.

Has this timeline changed?

To clarify, initial share is close to 100% for roughly 3 to 6 months, after which competitors enter. That structure remains unchanged.

Understood. My first question is on page 13: the capacity index jumps from 2.3 in FY28 to 2.8 in FY29, while capex for FY28/FY29 is projected to drop to around ¥100.0 billion.

Why does capacity increase significantly in FY29 despite lower capex in that year?

The ¥500.0 billion investment will progressively come online, with full mass production contributions reaching the 2.8 level by fiscal 2028-2029. Capacity additions ramp up gradually rather than all at once, the 2.8 capacity level reflects the culmination of investments made through fiscal 2028.

Exceeding 2.8 would require the additional ¥80.0 billion to ¥100.0 billion capex in fiscal 2029 and beyond?

Yes, that is correct.

My second question is regarding page 15. The package roadmap does not show a calendar year 2027 package image.

What is your view on calendar year 2027? Also, regarding EMIB-T customer expansion beyond custom ASICs, as interposers grow larger and interposer splitting or bridge connections emerge, how do you see EMIB-T adoption expanding?

For calendar 2027, interposer-based packages shown in the dotted line between 2026 and 2028 will continue. Some customers may introduce bridge-type solutions around 2027, but the largest GPU customer is likely to introduce them around 2028.

Regarding interposers versus silicon bridges, silicon bridge-type architectures do not completely eliminate interposers; hybrid configurations combining interposers and silicon bridges will likely emerge. While we omitted this from the chart to avoid overcomplication, combinations of interposers and silicon bridges are entirely possible.

Would that start around 2027?

We are not certain yet, but likely around 2027 or 2028.

When EMIB-T demand ramps up in 2027-2028 before the Gama Plant is fully operational, EMIB-T capacity could be very tight. How will you respond?

EMIB-T capacity is currently insufficient relative to demand. Adding capacity earlier than 2027-2028 is challenging, we are actively discussing allocation with customers.

However, the current plan for Gama uses only about half the space, leaving the remaining half available. Filling that space over the next two years would allow us to expand EMIB-T capacity further.

You will manage through customer partnerships and supply chain collaboration?

Yes, we will do our best within achievable limits.

Thank you very much.

Thank you very much.