GlobalWafers Co., Ltd. (6488.TWO) Q3 FY2025 Earnings Call Transcript & Summary

Good afternoon, ladies and gentlemen. Welcome to the joint earnings call for the third quarter of 2025 hosted by GlobalWafers. I am Leah Peng, the spokesperson of GlobalWafers. Joining us today is Doris Hsu, the Chairperson of GlobalWafers. Here's how this event will unfold. Doris will begin with executive comments, providing insights into overall performance and strategic direction. We will then address investor questions received in advance, followed by an open Q&A session. The call will last for 60 minutes and conclude at 5:00 p.m. [Operator Instructions] Before we begin, I would like to remind you that today's discussion may contain forward-looking statements. Please be aware that these statements are subject to various risks and uncertainties which could cause actual results to differ materially from our expectations. Please refer to the safe harbor notice in our presentation. Now without further delay, I would like to pass the floor to our Chairperson for the executive comments on GlobalWafers. Doris, please.

Thank you, Leah. Good afternoon, everyone. Thank you very much for joining GlobalWafers' Q3 2025 Earnings Call. Leah Peng, our spokesperson, will guide you through the presentation and respond to some of the frequently asked questions we have received recently. I will also take some questions during today's section. First of all, let me share some comments on our overall financial results and also provide you an update on our operational progress. Let me walk you through our financial performance, discuss the key drivers behind the results and share our outlook. 2025 has been both a challenging and pivotal year for GlobalWafers. The external environment remains full of uncertainties, from geopolitical tension, currency fluctuations, to tariff and policy shifts, all adding complexity to the global economy. Even so, GlobalWafers continue to move forward with resilience, supported by our global footprint and operational strength. Now I'd like to share a few very encouraging updates with everyone. First, in terms of our operational footprint. This year marks a major milestone as our global expansion projects begin to bear fruit, following the completion of several brownfield expansion last year. In Japan, our new factory in [ Utsunomiya ], Japan has been successfully ramping up its 12-inch epi wafer for advanced applications and repeatedly achieving record high production volumes. Nearly all major customers have already completed qualifications for this new line in Utsunomiya, and additional advanced product evaluations are progressing as original plan. The project has steadily increased output quarter-by-quarter, and the local government subsidy was received 6 months ahead of schedule, further strengthening financial flexibility and supporting future expansion momentum. In Europe, GlobalWafers is expanding its 12-inch wafer capacity in -- capacity through the new FAB300 in Italy, Novara, Italy. Qualification sample shipments are underway, with multiple customers providing positive feedback. Small volume production for key customers will begin from this quarter, followed by a mass production ramp-up in the first quarter next year. Revenue contribution is expected to increase gradually from 2026 as the production stabilize -- as production stabilized. On the subsidy front, the Italian government is reviewing the first subsidy application now for the project, with the initial disbursement expected in the very near future. Further reviews and payments will continue through 2027 in line with the project construction and ramp-up schedule. In the U.S., the Missouri facility, which is an SOI facility, is now established as an exclusive domestic site for 12-inch SOI wafer R&D and manufacturing site, forming a very critical pillar in the U.S. semiconductor supply chain. The ramp-up to high-volume manufacturing has begun and will continue throughout 2026. The site continue to provide numerous samples and now, routine orders across multiple SOI applications. Multiple customers -- multiple customer supply agreements have already been signed and some others are in discussion. On the subsidy front, Missouri state level grants have already been received. And the first instrument of U.S. CHIPS Act fund -- CHIPS Act grant is expected in near 2027 into the AMIC, advanced manufacturing investment credit -- investment tax credit, will further boost the financials in 2027. In the U.S., we have another site in Texas. The Texas expansion continued to progress steadily and has entered a critical ramp stage. Customer engagement has been very encouraging, with several leading global semiconductor and system companies having aligned to use GWA's U.S.-made wafers for their local production. In total, 2 customers have already achieved the qualification and are in mass production state. In addition, 4 more customers are in qualification phase and other customers are in discussion, which represents up to 90% of the capacity of GWA's 4 Phase 1 production ramp. On the policy front, we received the first CHIPS Act payment, which is USD 200 million and USD 1 million workforce grant in June. The increase of AMIC, advanced manufacturing investment credit, to 35% -- from 25% to 35% in 2026 will further enhance the project economics. Potentially, we are expecting to receive -- to have some meaningful savings from AMIC. These developments, strong customer alignment and government support, confirm that GWA is now firmly positioned as a key node in the U.S. semiconductor supply chain. Looking ahead, there is another very encouraging development on the subsidy front. Based on the current progress of ongoing applications and government discussions in multiple regions, GlobalWafers is expected to receive additional subsidy disbursements, totaling up to around $100 million in 2026, mainly related to our U.S. and European projects. These potential inflows will further strengthen our financial flexibility and support continued investment in capacity expansion and also for the technology advancement. There are also some positive developments on the cost side as well. First, regarding CapEx -- capital expenditures, which have always been very important factors affecting our overall profitability. Our major construction investments, such as the new plant buildings, were largely completed in 2024. Our global expansion CapEx has now entered the consolidation base and total spending for the first 3 months -- first 3 quarters of 2025 has declined significantly, down by about 1/3 compared with the same period last year. So it's about 1/3 down compared with last year's CapEx. And secondly, the recent industry -- recent interest rate cuts will also help ease GlobalWafers financial burden as well. The U.S. Federal Reserve reduced policy rates twice this year, with 25 basis point cuts in both September and October, this bringing the federal -- the federal fund rate down to 3.7% to 4%. A major portion of GlobalWafers' financial liabilities is denominated in U.S. dollar. So based on the current loan balance, each quarter point rate cut is resulting in approximately USD 2.2 million saving annual interest savings. So overall, lower U.S. interest rates help reduce our financing costs and provide greater flexibility to navigate future challenges. So that's a very good support for us as well. In addition, our inventory level has returned to the very healthy range seen in late 2023 -- in early 2024 now, reflecting solid inventory management and a stable foundation for the next upturn. I'll quickly update the financial highlights in Q3 2025 in the following sections. Revenue. GlobalWafers reported consolidated revenue of TWD 14.5 billion in the third quarter of 2025, representing a 9.5% decrease compared to Q2. For the first 3 quarters of the year, revenue reached TWD 46.1 billion, down 0.4% Y-o-Y. This is our revenue performance in NT dollar. But in the first half of 2025, the NT dollar appreciated notably against the U.S. dollar and only slightly weakened after August. So as most of GlobalWafers' revenue is denominated in U.S. dollar, this currency movement made our reported result in NT dollar appears so much softer. In local currency terms, third quarter revenue declined 9.5% Q-on-Q. And cumulative revenue for the first 3 quarters was roughly flat year-over-year. However, if I change this to U.S. dollar terms, our business performance remains stable. So our third quarter revenue reached USD 0.5 billion. Our Q3 revenue down only 0.7% Y-o-Y. If I use U.S. dollar as a currency is Y-o-Y 0.7% down, while cumulative revenue for the 3 quarters of this year -- for the first 3 quarters this year grew 2.5% from the same period last year. That's based on U.S. dollar currency. So gross profit and operating income wise, in the third quarter, our gross margin was 18.4% and operating margin was 8.5%. For the first 3 starting -- for the first 3 quarters of the year, gross margin was 23.6%, while operating margin stood at 13.6%. Our net profit and EPS. In the third quarter, our net profit margin was 13.6%. And for the first 3 quarters of the year stood at 11.1% net profit. EPS for Q3 came in at TWD 4.12 per share and TWD 10.68 per share for the first 3 quarters this year. The higher net profit compared with operating income was mainly due to the valuation gain from our shareholding in Siltronic, reflecting the recovery of the stock price during the quarter. As of the end of Q3, our prepayments amounted to TWD 26 billion, equivalent to approximately around $0.87 million. That's our current prepayment status. Let me explain a little bit about our gross margin. I think you definitely have already noticed that our gross margin in Q3 is lower than previous quarter. Let me explain the main reasons behind the lower gross margin in the third quarter, which can be summarized into 3 key reasons. Number one is that, first, new production lines across our sites, including those in the U.S. of the Texas new site in St. Peter, in Missouri, 2 new sites in the U.S. and 1 in Italy, Novara, and 1 in Utsunomiya Japan. These new fabs remained in the trial production and customer qualification base is not completely qualified yet. So still working on the qualification and trial run -- trial production right now. These fabs are still in the early ramp-up and process optimization stage. We started pilot and related costs continue to be recognized. Each site impacts the overall margin by roughly 2% to 4% percentage point. That means that these new lines, although revenue contribution is very minimum, but the expenses and all the costs are we see -- we have to be recognized. That's the first reason. So the first reason why the gross margin is much lower, it's because of those new lines start ramping -- start making the trial runs and test optimization. So that's the first reason. And the second one is that the Texas fab, our GWA certain costs that were previously capitalized by end of June, a lot of costs were capitalized. But those -- many of the costs have been expenses from this quarter -- from Q3 under accounting standard. So these costs are now recorded under COGS, cost of goods sold, instead of OpEx or capitalized with directly reduced gross margin. Under accounting principles, once a production line enters the prior production phase and is capable of producing wafers that meet customer specification, starting up in the pilot line costs are no longer eligible for capitalization and must be expensed. That's why starting from Q3 this quarter, GWA or the production cost is recognized as COGS. So even the revenue contribution is very minimal. But that has already been recognized as COGS. That's the second reason. And the third reason is that energy cost, higher electricity cost during the summer season also increased the overall cost ratio. The third quarter coincided with the summer peak period when electricity and utility expenses rose across most regions. As production volumes remain low during the trial production phase, fixed cost absorption was limited so further raising the unit cost. Overall, the lower gross margin this quarter mainly reflected the accounting change at GWA from capitalization to COGS, the transition from capitalization to expenses under COGS, together with the expenses recognition from other fabs still in the trial production stage and the impact of the CNO cost increases. These are the 3 key reasons for the gross margin -- for the gross margin down. These are temporary pressures during the early ramp-up phase of new capacity. As customers' qualifications are completed and production volumes increased utilization -- increased yield increase, then we will improve -- the utilization will improve and margins are expected to recover progressively. That's for the gross margin. Global inflation. I think global inflation -- global inflation expectations remain broadly stable, allowing major central banks to begin gradually easing policy rates, providing some support for economic growth. However, tariffs and policy uncertainties continue to weigh on consumer sentiment and business investments. So we're still very cautious for monitoring the global economy and also, we follow very closely with each country's new policies and taxes tariffs. So looking ahead, we believe that global GDP growth is expected to slow slightly between 2024 and 2026. Even so, I think the overall economy remains resilient, supported by strong AI-related investment in the United States. Okay. Yes, the semiconductor market continued to show strong momentum in AI-related advanced nodes, while demand for mature nodes remained relatively conservative, also for small diameter like 6-inch or 8-inch or the demand is much more conservative than the advanced one. So strong demand for AI and advanced logic chips has supported overall industry growth, driven primarily by high-end wafer demand from cloud computing and hyperscale data center expansions. In contrast, markets related to mature nodes and consumer electronics remain relatively soft as what we are seeing on the semiconductor industry current status. Okay. So I think that's pretty much the information I want to share with everyone. Next, Leah will present industrial outlook, highlight our competitive strength and also review our overall operational performance and actually some key investor concerns, followed by some Q&A. Thank you very much. Leah, please.

Thank you, Doris. I will quickly update GlobalWafers recent performance and the answer questions we received so far. So please turn to Page 12. Rising AI and advanced process demand continue to fuel growth in the wafer fabrication equipment market projected to expand from USD 111 billion in 2025 to USD 122 billion in 2026. Foundry and logic remain key drivers, AI, HPC, power and sensing applications are now central to new industry investments and advanced low expansions. GlobalWafers global expansion strategy is fully aligned with this trend, positioning ahead of the demand to capture the structural growth. In Page 13, this is a tentative signs of stabilization. Our customers are showing signs of stabilization. Solar recovery remains cautious. Early tariff-driven wins lifted inventories. But by Q2, the levels have fallen to around 135 days, indicating steady digestion. Regionally, performance is mixed, except Japan. Most areas saw modest Q2 growth. Overall, AI demand remains the main driver, while other markets is in a mild stabilization phase, with limited visibility for non-AI segments. In Page 14, we can see that the U.S. is rapidly rebuilding its semiconductor ecosystem through large-scale chip reshoring. Over USD 500 billion in private investments are expected to triple wafer capacity by 2032. In the past decade, U.S. capacity grew only 11%, but now is projected to rise over 200% by 2032, far outpacing the global average. Through our U.S. expansions, GlobalWafers is actively supporting the transformation, strengthening local supply chains and deepening partnerships with key customers. In Page 15, you can see that the U.S. is seeing an unprecedented wave of semiconductor investments, creating strong regional clusters across manufacturing, materials and advanced packaging. GlobalWafers is strategically positioned the way in these hubs. Our Texas fab serves as the flagship 12-inch facility, while Missouri focuses on SOI products, both aligned closely with customer localization, strengthening supply chain resilience and long-term partnerships. Our GlobalWafers America GWA is moving quickly through product qualifications with strong assembly demand and small shipments. [indiscernible] in Missouri started at around in Q2 '25 and targets mass production in the second half 2026. Also under the One Big and Beautiful Bill, the advanced manufacturing investment tax credit rate will be increased from the current 25% to 35% starting in 2026. This will further enhance the project's long-term financial fast agility. Next, I would like to highlight an important milestone in Europe, the launch of FAB300 in Novara, Italy. While demand in Europe remains soft, customers are increasingly focused on local supply. But Italy's first and only 12-inch advanced silicon wafer fab, we have received strong feedback from major customers, some already signaling recovery and requesting rush order readiness. FAB300 inaugurated in October this year represents a total investment of approximately EUR 450 million, with 25% subsidized under the IPCI framework. It is equipped with full process capability, producing polished and epi wafers for logic, memory, power and sensor applications, fully powered by renewable energy and the aims to achieve 50% water recycling efficiency during its ramp-up phase. This will fully align with the stringent ESG standards of our European customers. Next, I would like to share our latest progress in next-generation material. We have completed 2 breakthrough prototypes, the square silicon wafer and the 12-inch as [indiscernible] wafer. Both now in customer sampling and pilot pression, marking key advances in wafer size and material innovation. These developments are technically demanding, but with our strong manufacturing experience, we are steadily overcoming challenges and strengthening long-term competitiveness. The key highlights include successful prototype delivery and scaling. In application, the square silicon wafers enable large area design and higher material utilization in 2.5D or 3D packaging, improving design [ feasibility ]. While a 12-inch SIC wafer offers superior heat dissipation for AI and high-performance computing. Please refer to Page 19 for our Q3 financial overview. Following Dori's remarks, I would like to briefly break down the cost structure behind these changes. Compared with the previous quarter on a revenue ratio basis, our depreciation increased by about 1 percentage point of our new fabs enter trial production. Power costs rose about roughly 2% points due to the summer peak season. Labor costs increased by around 2 percentage points, mainly reflecting the annual salary adjustment one-off items. The remaining changes were mainly related to the reproduction costs associated with GWA. Please note that the ratio increases were mainly driven by a lower revenue pace in Q3, which magnified the cost percentages rather than a notable rise in absolute expenses. Our operating margin was 8.5%, down 6.7 percentage points. Net profit margin rose by 3.5 percentage points to 13.6%, and EPS amounted to TWD 4.12 per share. Please note that the income tax rate in Q3 decreased to 10% from the previous 27% Q2. This is mainly due to the tax refunds recognized by our certain subsidiary following the local regulatory changes and the reversal of previously accrued undistributed earnings tax. Also some sites record lower profits in Q3. So the reduced taxable income led to lower tax expenses, contributing to a reduction in the consolidated rate. Our EBITDA margin increased to 2.3 percentage points to 30.2%. This mainly reflects the valuation gains from Siltronic shares we hold. CapEx totaled TWD 7 billion and depreciation was TWD 2.5 billion, slightly lower than Q2, but this reflects the foreign exchange translation effects and the completion of depreciation for certain legacy equipment. Please refer to Page 20 for 3 quarters financial performance. Revenue for the first 3 quarters reached TWD 46 billion, representing 0.5% Y-o-Y down, and gross margin declined by 8.6% to 23.6%. Operating margin was 13.6%. Net profit margin, 11%. And EPS, TWD 10.68 per share. Our EBITDA margin amounted to almost 28%. CapEx totaled TWD 26 billion, significantly lower than last year, but most major investments were completed in 2024, and projects have not entered into the final phase. Depreciation was TWD 7 billion, and will continue to accumulate as the expansions progresses. Okay. So let's -- please refer to Page 23 for the income statement. To reflect the true operating performance, we presented the simulated financials, excluding the influence of major expansions and the Siltronic mark-to-market valuations. Excluding these factors, our first 3 quarters gross margin would have reached to 31%, EBITDA margin to 36%, and net profit margin to almost 18%, and EPS would have amounted to TWD 16.7 per share. This will significantly higher than the reported figures now. Please refer to Page 24 for our balance sheet. Cash and cash equivalents decreased to TWD 16 billion in Q3, but this mainly reflects the internal fund reallocations to support the CDR refinancing needs with part of the balance reclassified under other current financial assets. However, if we include all the cash-related items classified under the other assets in accordance with accounting standards, our total cash will amounted to TWD 51.3 billion. Changes in borrowings during the third quarter mainly reflected the funding needs for capacity expansion and debt structure optimization. We issued a corporate bond and commercial paper to support the global expansion and enhanced financial flexibility, while some long-term loans were reclassified as the current portion due to maturity. Now I would like to address first the questions we have received from investors recently, and those we anticipate will be raised. The first is the guideline of gross margin and depreciation. In 2025, the gross margin will continue to be affected by higher depreciation expenses and new fab cost, keeping the overall pressure in the short term. The slight decline in Q3 depreciation, this is the result of NT dollar appreciation and the expiration of order access. If we exclude this temporary effect, we forecast that in the full year in 2025, the depreciation will surpass the figure in 2024, and they remain at a high level. If government subsidies, such as the U.S. AMIC and Italy's IPCI programs are recognized, this will partially offset the impact of higher depreciation and support our margin stability. The second question is that has the growing U.S. localization demand translated into actual orders or qualification progress? Yes, the answer is yes. With rising concerns over tariff policies and the supply chain resilience, our U.S. customers' demand for local sourcing has become increasingly evident. GlobalWafers has received a request from multiple customers to accelerate simple delivery and the qualification process and to reserve additional capacity for them. Please indicate that the localization trend is translating into concrete actions. So looking ahead, the U.S. market is expected to become our key [indiscernible] to long-term growth driver for GlobalWafers. And we will continue to strengthen its local supply presence. And the third question is that please share with us the progress of your partnership with Apple. In August, we have entered into a new partnership with Apple to strengthen the U.S. semiconductor supply chain as part of Apple's American manufacturing program, GWA and Apple are working together to build demand for GWA's 300-millimeter wafers throughout the supply chain. We expect this to be a multiyear process. Another one is the status of the Section 232 tariff in investigation and our counter major. The U.S. Station 232 investigation of semiconductors is still ongoing and has not yet reached any decision phase. By law, the Department of Commerce must submit its report to the White house by January next year, and the President will have 90 days to decide on any potential actions. In the meantime, semiconductors and related equipment remain tariff-free, and policy discussion continue to emphasize the U.S. supply chain security and the domestic manufacturing. In view of potential policy changes, we continue to closely monitor the government announcement and industry development, and is well prepared for the possible tariff from semiconductors and the raw materials. Overall, the U.S. tariff policy is expected to further accelerate the trend toward the local supply and manufacturing replacement. Our localized capacity, I think we will have quite a good opportunities because of this. Okay, the next question is regarding memory. What is the outlook for memory and power wafer market? Do you agree that the worst is behind us? Our answer is that the memory market has been gradually improving since the second half, driven by rising demand for HBN DDR5 and the enterprise storage applications. These are clear signs that the overall industry has emerged from the bottom and entered a recovery phase. We have also observed a modest increase in wafer shipment volumes, indicating that the end market demand is gradually flowing through to the upstream supply chain. Above is our response to the questions. Now I would like to turn it over to Doris for the Q&A session. [Operator Instructions]

Thank you, Leah. Thank you. Okay. Let me start from the first question. What is the current utilization rate for each wafer size? That's a very important question. For 12-inch, for 12-inch, if I exclude our newly established 12 inches, which are still under qualification, if I exclude those capacity, based on our original capacity, our utilization rate's over 95%. It's almost for the existing capacity. But if you add the new capacity just established, just finished construction in Q1 this year, Q2 this year, then, of course, we have a lot open capacity. But those are still -- those are newly built and underqualification right now. So the first sensor is over 95% for existing old 12-inch capacity. 8-inch capacity, 8-inch not expansion, but 8-inch capacity right now, utilization is below 80%, slightly lower than 80%. And for 6-inch or below, the utilization rate is lower than 70%. That's our silicon wafer capacity status. And compound material compound utilization for silicon carbide compound utilization rate is still low and market price is still pretty weak. So our silicon carbide small diameter, 6 and 8 -- 6-inch and [ 8-inch ] silicon carbide now, utilization is below 50% right now. But for GaN is much more positive. Our GaN utilization rate, gallium nitrate, GaN on silicon utilization rate is overbooked, is over 100%. So we are working on expansion. I think by end of December, we will be receiving new tools, which will be accounting for around 30% capacity up. So for GaN, it's fully loaded now. Expansion is underway. By the way, I just want to add a little bit comment about silicon carbide. We are seeing some -- it seems that we are seeing some upside of silicon carbide, some recovery rebound of silicon carbide demand. It's not very clear yet, but we are seeing some signals. Some customers are asking for rush orders and small volume. But here and there, we are receiving some extra -- seeing some new orders now. So I hope that '26, silicon carbide will be a little bit better. So that's my answer to the first question. And the second question is in terms of 12-inch silicon carbide, any information I can share. I think -- what we can share is that 12-inch silicon carbide, of course, main applications is for better thermal conductivity. So there are several different crystallization method, different crystallization methods are under development right now. We are working with -- we are working with several toolmakers to develop special tools for 12-inch silicon carbide wafer. The 12-inch silicon carbide wafer, we have more than 1 million wafers, 300 million wafer -- silicon wafer -- 12-inch silicon wafer, more than 1 million wafers every month, way higher than 1 million wafer every month. But all the tools we are using basically cannot use for silicon carbide because the hardness is different and also the transparency is different. So we need to work very closely with our tool makers to develop special tools for 12-inch silicon carbide. So we're still working on this and for different crystal -- crystalization method, that's what we are doing. And all the others are pretty sensitive. So sorry, we cannot disclose too much. And next is with AI driving strong growth, what is your view on material demand and market outlook? Mature node, right now, is -- the market demand is very, very divided, is totally different. 300 millimeter an advanced products demand are still strong, but for the mature node or smaller diameter, demand is pretty weak. So right now, we think that it's not -- when we say -- when we say mature node, it's not only for 12-inch mature node, but also for 8-inch as well. So right now, it's still pretty weak. Yes, we haven't really seen some big pickup or upside yet for the legacy. Okay. Next, question is about the fund, the grants from U.S. or from Europe, how we are going to count? That, as we explained several times in the past earnings call, our accounting way is that we will recognize the fixed asset value -- the minus item of the fixed assets. So if -- for example, if our fixed assets amount is 100 and we received 10 from government grant, then we will reduce the fixed asset value down to -- from 100 down to 90. And if its depreciation is up by [ 60 ] months, then we -- the 90, we'll divide it by [ 60 ] months. So every month, you will see a little bit lower depreciation cost. That's our standard accounting way for this. Yes. So that's my explanation for the accounting. If you want to know a little bit more detail, please feel free to reach out to our spokesperson or our accounting team. Okay. And next is the new capacity. So what's our overall capacity -- overall market share for the new capacity? I think you're asking about 12-inch only because of our expansion, mainly for 12-inch, I think it's a little bit hard to calculate, for 2 reasons. One is it's very hard to tell how much China capacity is not very clear. You hear a lot of rumors in the market. You don't really know what's their real capacity. We are talking about capacity. It's not nominated capacity, it's real production capacity. So we don't have real solid information. We are monitoring this but it's not very clear for us yet. And also, we heard that some -- we don't know that some other companies maybe -- maybe they have some expansion or just some upgrade from the existing tool and make some upgrade. We don't really know yet. So I cannot give you a [indiscernible] number, but I think it will be somewhere around 20%. That's our view and that's our goal as well. Okay. Next question is better transparency. All the 3 reasons for lower gross margin should be acknowledged by management 3 months ago, we have [indiscernible] near-term outlook in the future to avoid surprises. Yes, we will improve this one. Actually, yes, [ Bruce ], you are very right that I think we should be -- we should made it a little bit more clear earlier in our last earnings call. What we didn't expect is that -- we didn't expect that our customers asking us to accelerate our qualification. So we have to run a lot of optimization process to make sure that we can reach the volume. So that's the main reason. And also, we didn't expect that we emitted, starting from Q3, we have to transfer -- we have to -- we have -- the whole conditions make that were not eligible anymore to capitalize. We have to make it an expense at the same time. So this we missed these key points. So we'll make it more clear. Thanks for your advice. We will monitor this one as well. Okay. Okay. So the last question is -- is this the last one?

Yes, this is the last of the text questions, and we will switch to the live voice questions. The last question comes from [ Felix Lee ] from Morningstar. You are asking that if GlobalWafers could share the outlook of spot and contract wafer pricing. Since the utilization is over 95% at 12-inch, there are the conditions of favorable for GlobalWafers to raise prices? Are the new LTA being negotiated?

Okay. There are no new LTAs under negotiation for existing fab. Now all the new LTAs are for new capacity because customers wanted to have advanced one in local supply. So what they want is that you have to be in local supply and your tool, you have to be advanced supply -- advanced wafer supply. So as I reported earlier in my executive comment that we have multiple -- we have signed multiple supply agreements and also several more are under discussion. In the spot in contact wafer, because actually, market is still not that strong. It's only AI-related, high performance related and really, demand is a little bit stronger. So actually, current price is pretty tough. If you want to get more volume, you have to make some special support to our customers for -- by -- for example, by supplying some additional [ tightened ] allowance spec or you have to do improve the price or make some special support. So right now, spot price is very low. And even price LTA prices there, but customers -- actually, many of the customers will ask you to delay a little bit of the LTA volume. They want to convert part of the volume to spot purchase. That mean that they push out the LTA volume a little bit part of the -- part of the wafers are LTA wafers, but part of the wafer will become the spot wafers. There are many ways to -- you have to support your customers. So that's the current status.

Okay. Ladies and gentlemen, the floor is now open for last questions. [Operator Instructions] Donnie, yes, we are sending you the unmute request, please accept it and raise your questions.

Can you hear me?

Yes, yes. Please go ahead.

My first question is regarding to gross margin. So firstly, is that I'm not sure whether you have recognized USD 200 million subsidies in the third quarter result because I think you have announced there has been subsidies from U.S. government in August. So I'm not sure whether it has been reflected into your -- quarter result?

No, that's in Q2, we received by end of June. So it was in Q2.

I see. So there's no subsidies recognized in third quarter, right?

That's correct.

We have some small subsidy from Japan government.

Japan, a little bit very small money, but not the U.S., not the big ones.

Okay. I see. Then probably I'll break down the questions in more detail. So when exactly you will be receiving the next big amount of subsidies in the coming quarters? And also, if you can give us some updates on the actual mass production time by different new fabs? And when exactly the depreciation cost will reach the peak level in the coming quarters?

The next fund -- big fund, I think it will be H1 next year. It's very hard to provide the very specific date or month or quarter because AMIC, we know that when is a due date to applying, which we have already applied. I mean in the U.S., we have already applied for the first AMIC -- AMIC credit application. We have already filed the application by end of October. We have already done that. But we don't know that when we will get the payment yet. And also in Europe, -- we have also -- we have already submitted our application for the first time. And we don't know how many months the government will release the payment. So we guess that -- we also -- we checked the related department, and they told us that it should be Q4, it could be Q4 or H1 next year. So that's the answer. So my best guess is that we should be able to receive some funds from both U.S. and Europe in the first half 2026. And it's very likely that maybe by end of 2026, we'll be able to receive another one. So as I said earlier, our forecast next year, we will be able to receive over USD 100 million fund from U.S. and Europe. That's our expectation.

Understood. And in terms of the mass production time line of the new -- different new fabs and when exactly the diversion costs to pick out?

The new mass production actually will start ramping up. We are -- for example, U.S., we have already had 2 customers under HV, high-volume manufacturing right now, but some customers are still test monitoring. So it's not -- which time point that we start mass production, we are already start doing mass production. For example, [indiscernible] Japan [ one ] is already mass production and it's getting several record high already, the volume record high already. So that's our Japan, already mass production. And -- but the production cost is still high. The main reason is because all the tools are new. So depreciation cost is much higher. But at least you see very clear that EU is much better and quality is much better in volume productivity, everything is much better and start ramping up the volume. So that's Japan. And the second one is U.S. Texas. Texas already started the mass production for 2 customers and several more customers will -- we'll have more and more new customers enter the HVM stage in the U.S. And the third one is SOI Missouri fab. Missouri, actually, our LTA coverage for Missouri is highest among all of our sites because key products is, for silicon photonics and RFSOI and FDSOI, many advanced SOI material 300-millimeter. So our LTA already cover over 90% of our capacity. But SOI is so much more complicated, more difficult than ordinary silicon wafer. So the optimization time and learning curve will be a little bit longer. So I believe that the mass ramp-up will be about a quarter will be maybe Q1, maybe Q2 next year, we'll see more ramp-up Q1. The best case is maybe starting from Q1, we will see the SOI start mass production as well. And the Italy fab also will start mass production very soon from one customer with small volume, though, maybe by end of Q4 this year, we will start seeing -- start ramping up gradually for mass production.

Understood. So -- lastly, just depreciation cost. When do you expect it to pick up?

Depreciation costs, depreciation costs, I think is -- depreciation cost -- we're still working on because some tools, which haven't been qualified yet, it's not -- we haven't started the depreciation yet. So the tool depreciation cost will increase, but the -- when the revenue increase at the same time, actually, the overall financial performance will be much better because the depreciation -- the equipment -- the building depreciation already started, but the tool depreciation depends on our qualification status and utilization status. So that's what we are still working on. We don't have the detailed date yet or schedule. We will have much better plan visibility by end of the year.

And the next question will come from Bruce. Please accept our request to unmute. Bruce, we cannot hear you.

Can you hear me?

Yes.

Can you share with us about the 12-inch silicon carbide in terms of your competitive advantage? Why you can win the business? Why you believe you can win the business? Why we need 12-inch wafers for that? We cannot use for 8 or 6 inches? And can you tell us what kind of addressable market we looking at? I mean is this just a consumable, for process? Is it like you got to do for every wafer. Can you share anything about that?

You mean 12-inch silicon carbide wafer, right?

Yes.

Okay. For now, I think basically 12-inch silicon carbide wafer, there are several key applications. Number one is for thermal conductivity for -- especially for AI service or high-performance applications. So basically, it's not for any device is for the thermal, they need thermal spend, high thermal conductivity material. That's the number one application for that. And the second one, for 12-inch silicon carbide wafer is for interposer. But it's not the key purpose now. For now, is for thermal conductivity. Why [ interpose ]? It's good -- silicon carbide wafer is very hard. So it's good for interpose, but 2 weak points that is not very competitive to be used for interpose. Number one is price is definitely still much higher than silicon wafer. And number two is that because it's -- because of the hardness, it's very hard to make a [indiscernible] hole and some specialty. If you want to make the thinning process, grinding off the bat side is very hard to process. So that's why interpose looks like that it can be used for interpose, but not now yet. So now current focus is for the thermal conductivity. And number three, application for 12-inch silicon carbide is for AI glasses or AR glasses. And why it's for AI or AI glasses because silicon carbide is a good material for a wave guide. So it's good for that. And for AI glasses, you need a big dimension. If you use 8-inch, you can use 8-inch, but 8-inch, you can -- one 8-inch wafer, you can get only maybe one pair of glasses -- one pair of classes. So it's not the most competitive one. So these are the key applications. But right now, it seems that the most promising right now for the processes for thermal conductivity. But in China, we know that AR glasses is very popular now. There -- a lot of new projects are under development for 12-inch silicon carbide for AR glasses, AI glasses, waveguide application.

Just a quick follow-up for the first application. It's going to be like for like every -- who needs those? I mean for -- it's going to like 2 nanometers, one [ A 16 ] or [ 16 A ] or anything further? Every [ A 16 A ] wafer, you need one wafer for -- one silicon carbide wafer for semiconnectivities? Is that the right way to think about it?

Yes, it's very close.

I see. So what's your competitive advantage in this space? Who are your competitors currently?

I think we are -- maybe we are one of the best for 12-inch silicon carbide wafering capability. The reason for that is that if you check all the whole world of silicon carbide suppliers, no one has experience like GWC who process over 1 million wafers a month for 12-inch wafer. The only thing -- the only technology we have to learn is how to handle a wafer much tougher, harder than silicon. But all of our employees, we have thousands of employees are super familiar with how to handle thin wafer, 12-inch [ rung ] thin wafer. We know of this, and we know the process. But the only difference is that it's much harder. So -- but for all the other -- the silicon carbide wafer company, they have experienced for 4-inch and 6-inch, 8-inch, but they don't know 12-inch silicon carbide wafer. So of course, tools are different, but we still -- we still have -- we have a lot of silicon carbide experience and technology and IPs can be utilized in upgrade for silicon carbide. So right now, what we are doing now is that we process silicon carbide wafers. We have 2 different kind of material source, one is our own silicon carbide -- 12-inch silicon carbide [ board ] or crystal, but that's one source. But we have more -- more businesses that some companies, they can grow 12-inch boards, but they cannot make silicon carbide -- 12-inch silicon carbide wafering process as good as we are. So they ask us to do -- they bring -- they consign their -- the crystal to us and ask us to process for them. That's our strength, our competitiveness, our strength is the technology and IPs for 12-inch wafering capability is so complicated. It's much more complicated than crystal.

But in the longer term, you will do your own crystal right?

Yes, we are doing our own crystal. But volume-wise, it's not as big as the -- from third parties.

I see. And lastly, for the profitability, I mean, for the fourth quarter, I mean the depreciation, I just mentioned that are we seeing is still going to be the case. But the utility cost is supposed to be slightly lower at fourth quarter. So can we see a clear [indiscernible] for the gross margin in the near term?

Yes, I think so. I think Q4 should be -- I mean, Q4, the only uncertainty, some concern is how about 232 investigation, any special tax. Those are totally uncontrollable. If I assume that no impact from tariff or any 232 or any other factors, then we should be -- I think Q4 should be better than Q3 from gross margin viewpoint because no more new -- I don't think that we will have more depreciation in Q4 because all the [ 2 ] is pretty much, we know that what we have now. And I believe that our EO definitely will be getting better month by month, and energy costs will be lower and EO perform higher. I expect that -- we expect that our revenue from new lines should be a little bit better than what we were doing in Q4. That's our expectation. Okay. I think -- yes, can we have the last one.

Okay. Then Haas, please. Can you hear us? Please accept our request to unmute yourself.

Yes. I just have one question on the demand and also the overall supply demand outlook. In the near term, as you just mentioned, you are seeing strong demand on advanced logic and also memory. Given the number is around 50% of the low wafer consumption in advanced logic is another 10% to 15% of the low wafer consumption, how should we think about the supply and demand landscape for the overall industry right now? I think you mentioned that your 12-inch utilization is running above 95%. But just from your perspective, what is it like for the industry? And separately, are you seeing pricing trends stabilizing? Or you could even test on the material cost hike to your customers by raising your like-for-like pricing?

Thank you, Haas, for great questions and very hard to answer the question. Let me try my best to make it very clear. Yes, first of all, price, let me answer the price issue first. I think we see some upside from demand viewpoint. We see some upside from -- I just answer a little bit earlier that gallium nitride demand is super strong. So I think we see that this should be -- demand is increasing and EO is improving. So this is -- price should be stabilized a little bit. And we see the second recovery, it seems that some rebound. Not 100% sure, but it seems that we see the light, and that is silicon carbide. I'm not talking about 12-inch silicon carbide. I'm talking about 8-inch -- 6-inch, 8-inch silicon carbide. It seems that the demand is picking up a little bit. So we think that the price should be stabilized as well. Not increase, but should be stabilized. And number three, which is more encouraging, is memory. We are seeing some memory demand improving. It seems that maybe Q1 or Q2 will see some price improvement from memory. And the demand is because of the strong demand of -- the reason is because of the demand is strong for memory and for advanced one and including the package, actually, memory needs more wafer. SIM capacity needs more wafers. So that's part of the reason that this will be increasing. So we expect the memory price now is stabilizing, but it should be improved starting from next year, either end of H1 or H2. I think price should be stabilized and even improved for memory. And capacity, when I say 95%, I mean that our existing capacity only, not including our newly expanded capacity. If we include the newly expanded capacity, of course, the total utilization rate is much lower than that. So if you are talking about the whole industry utilization rate, I think it's overcapacity right now because when we say capacity, I think we cannot ignore China capacity. China has very -- has quite a lot of capacity, which we don't really know. There was a capacity now in China, but we believe that China capacity is -- China has very good capacity now. Yes. So I think, in general, for now, capacity is a little bit higher than demand. But we -- we think that based on all the semi forecast and a lot of market forecasts, semiconductor demand keeps growing in the next several years in a very significant way. So if that is the case, actually, according -- based on our forecast, I think it's very likely that starting from 2027, it will be another super cycle increase because no one right now, the past 2 years, business is very low. So basically, no further expansion for wafer companies. Basically, everyone just finished what they have already announced or already started expansion. But wafer company, especially memory and advanced AI-related application, will continue, will keep expanding more and more. We are seeing quite a lot of customers planning to expand further capacity. So our forecast is that starting from '27, capacity should be -- will become it very tight again. That's our forecast. But it's not -- there are still a lot of uncertainties, including taxes, tariffs and also currency development. Currency has a lot of impact as well. So that's our view. Haas, sorry, that's all I can answer now. But if you need more detail, you are very welcome to contact with us.

Yes. That is actually super helpful. And just a quick follow-up before you can wrap up the call is that on pricing, could you also share about the epi wafer pricing trend? And on that is probably the power you are doing more for the logic customers? And then just on the overcapacity, would it be fair to say that the current overcapacity situation is probably only like 5% to 10% of the oversupply versus like 10% to 20% a year ago? Or it's still like 10% oversupply at this stage?

Yes, I think it's about 5% to 10% overcapacity right now. It's not that much, especially for advanced ones. And price-wise, epi price is stable, but I think it will take a little bit time to really increase. Yes, of course, it depends on if it's -- if we're talking about advanced -- real advanced ones and very tight spec, I think the price stepping -- the cost will increase a lot as well. So yes, we are still working on the price negotiation with our partners, customers now.

Thank you, ladies and gentlemen. We would like to express our sincere appreciation to all of you for your valuable participation today. The earnings call concludes now. Thank you, and have a wonderful evening.

Thank you.