管理層發言
Welcome, everyone, to UMC's 2025 Fourth Quarter Earnings Conference Call. For your information, this conference call is now being broadcast live over the Internet. A webcast replay will be available within 2 hours after the conference has finished. Please visit our website, www.umc.com, under the Investor Relations, Investors, Events section. Now I would like to introduce Mr. Michael Lin, Head of Investor Relations at UMC. Mr. Lin, please begin.
Thank you, and welcome to UMC's conference call for the fourth quarter of 2025. I'm joined by Mr. Jason Wang, President of UMC; and Mr. Chi-Tung Liu, the CFO of UMC. In a moment, we will hear our CFO present the fourth quarter financial results followed by our President's key message to address UMC's focus and the first quarter 2026 guidance. Once our President and CFO complete their remarks, there will be a Q&A session. UMC's quarterly financial reports are available at our website, www.umc.com, under the Investors Financial section. During this conference, we may make forward-looking statements based on management's current expectations and beliefs. These forward-looking statements are subject to a number of risks and uncertainties that could cause actual results to differ materially, including the risks that may be beyond the company's control. For a more detailed description of these risks and uncertainties, please refer to our recent and subsequent filings with the SEC and the ROC security authorities. During this conference, you may view our financial presentation material, which is being broadcast live through the Internet. Now, I would like to introduce UMC's CFO, Mr. Chi-Tung Liu, to discuss UMC's fourth quarter 2025 financial results.
Thank you, Michael. I'd like to go through the 4Q '25 investor conference presentation material, which can be downloaded or viewed in real time from our website. Starting on Page 4, the fourth quarter of 2025. Consolidated revenue was TWD 61.81 billion, with a gross margin around 30.7%. The net income attributable to the stockholder of the parent was TWD 10.06 billion, and the earnings per ordinary shares were TWD 0.81. The utilization rate in the fourth quarter stayed the same as the previous one, around 78%. For the sequential comparison, revenue grew 4.5% quarter-over-quarter to TWD 61.8 billion. Gross margin improved to over 30% to now 30.7% or gross margin of TWD 18.95 billion. The non-operating income remained similar to that of last quarter. The net income overall contributed to shareholders of the parent is around TWD 10.05 billion or EPS of TWD 0.81 in Q4 of 2025. For year-over-year comparison, on Page 6, revenue grew by 2.3% to reach TWD 237.5 billion for the whole year of 2025. The gross margin rate is around 29% or TWD 68.9 billion. The net income attributable to the shareholders of the parent for the year 2025 is around TWD 41.7 billion or 17.6% net income rate. EPS for 2025 was TWD 3.34, which is a decline compared to that of TWD 3.8 in 2024. On Page 7, our balance sheet at the end of 2025 shows that cash amounts are still more than TWD 110 billion, with total equity of the company now at TWD 379.8 billion at the end of 2025. For ASP on Page 8, you can see that for the last three or four quarters, it mostly remained at similar levels for our blended ASP throughout 2025. In revenue breakdown on Page 9, the quarterly comparison shows an increase in Asia and Europe, with North America representing about 21% in Q4 of last year. For the full year breakdown on Page 10, the change is similar; North America dropped from 25% in 2024 to 22% in 2025. For Page 11, IDM for Q4 revenue still represents about 20%, almost no change. But for the full year number on Page 12, IDM accounts for 19%, increased by 3 percentage points to 19% in 2025. For quarterly revenue breakdown by application, it remains almost similar quarter-over-quarter on Page 13. For the annual performance on the application breakdown on Page 14, consumer increased by 3 percentage points to 31% from 28% in the previous year. We continue to see 22-nanometer to be our key driver of growth for recent quarters and also forward-looking as well. So 22 and 28-nanometer revenue in Q4 '25 now represent 36% of the total revenue pool. On Page 16, for the full year, the increase of 22 and 28-nanometer revenue is by 3 percentage points, and we show about a 2 percentage point increase in 14-nanometer on a year-over-year comparison. Capacity remained flat on a quarter-over-quarter comparison base, but it will decline by roughly 1% due to the annual maintenance schedule. On Page 18, our latest forecast for the 2026 CapEx plan is around USD 1.5 billion, which is slightly declined from USD 1.6 billion in 2025. The above is a summary of UMC's results for Q4 2025. More details are available in the report posted on our website. I will now turn the call over to President of UMC, Mr. Jason Wang.
Thank you, Chi-Tung. Good evening, everyone. I would like to share UMC's fourth quarter results. In the fourth quarter, our results were in line with the guidance, with flat wafer shipments amid mild demand across most of the markets. The 4.5% revenue increase during the quarter was supported by favorable foreign exchange movements as well as sequential growth in our 22- and 28-nanometer business, which continues to improve our product mix. With the 22- and 28-nanometer segment, 22-nanometer's revenue increased 31% quarter-on-quarter to a record high, accounting for more than 13% of total fourth quarter revenue. Looking at the full year, UMC delivered solid performance in 2025 with shipment increasing 12.3% and revenue in U.S. dollars up 5.3% year-on-year. Going into the first quarter of 2026, we expect wafer demand to remain firm. UMC is confident that 2026 will be another growth year as tape-outs on our 22-nanometer platform accelerate, and other new solutions continue to gain business traction. We have been working hard to lay the foundation for our next phase of growth, investing for the future in both capacity and technology. In 2025, we completed the new Phase III facility at our Singapore Fab 12i, which is already playing a central role in supporting customers to diversify supply chains. At the same time, we are striving to expand our footprint in the U.S. through innovative yet cost-effective modes of partnership, such as our 12-nanometer collaboration with Intel and the recently announced MoU with Polar Semiconductor. The leadership UMC has built over the past few years across specialty technologies, including embedded High Voltage, Non-Volatile Memory, and BCD, has and will continue to sustain stable business growth. Looking ahead to 2026 and beyond, we expect advanced packaging and silicon photonics to serve as new growth catalysts, positioning UMC to address the evolving needs of high-performance applications across AI, networking, consumer, automotive, and more. Now let's move on to first quarter 2026 guidance. Our wafer shipment will remain flat. ASP in U.S. dollars will remain firm. Gross margin will be approximately in the high 20% range. Capacity utilization rate will be in the mid-70% range. Our 2026 cash-based CapEx budget will be USD 1.5 billion. That concludes my comments. Thank you all for your attention. Now we are ready for questions.
分析師問答
Yes. Thank you, President Wang. The first question will be coming from Sunny Lin, UBS.
So, I have a few questions. Number one, Jason, may we have your thoughts on the overall market outlook for 2026? And then for semi versus foundry? And if UMC can continue to outgrow your adjustable market for this year?
Sure. Well, for 2026, we expect the AI-related segment to remain as the primary growth driver in the semiconductor industry. Furthermore, with the continuous commercial deployment of Edge AI applications, demand for chips using general-purpose servers is also expected to rise. In contrast, the adverse effect of the memory supply imbalance could put some pressure on specific consumer electronics. But overall, the semiconductor industry is projected to grow by mid-teens in 2026. The question for the foundry market is that we believe AI demand will remain strong and is the main contributor behind the low 20% growth projection in the foundry market this year. On the other hand, although memory pricing may impact demand in the foundry market, at UMC, we estimate that our addressable market will grow by a low single-digit percentage. UMC's growth was expected to outperform the average growth of our addressable market.
Got it. So, then my second question is on pricing. Lots of discussions and obviously, Chinese peers are raising pricing. So how should we think about the pricing outlook for mature foundry and for UMC through 2026? Would UMC be able to start to reflect better value? And if yes, which product categories should we expect more upside from here?
Okay. Well, we do anticipate a more favorable ASP environment in 2026 versus 2025. This outlook reflects our disciplined pricing strategy and the positive impact from multiple reasons: product mix optimization, improved loading, and reduced exposure to more commoditized market segments. As you're referring to Chinese players, we expect strong growth momentum in our 22-nanometer demand to support our product mix in 2026 as well. Overall, our pricing strategy remains consistent and is anchored to the value we deliver through technology differentiation and manufacturing excellence. So, we do think the 2026 pricing environment is more favorable now. Now the question about which product specifically, we generally do not comment on pricing for any specific product or node. But in general, we do see the environment as more favorable now.
No problem. That's very helpful. And then a follow-up would be on the overall industry supply versus demand for the coming few years. TSMC on the recent earnings conference talked about the plan to optimize capacity for mature nodes to better support cloud AI demand in the coming few years. So from your perspective, how should we think about the opportunity here? Are you starting to see more client engagement for new products in the coming few years?
We're always excited to see more customer engagement. More importantly, we need to prepare ourselves to cope with market dynamics, and we welcome any opportunity to support our customers. We view this landscape shift as an opportunity to further optimize our product mix and gradually improve ASP and margin as well.
No, got it. Got it. And then maybe lastly, just on your Singapore expansion. How quickly are you planning to ramp capacity in 2026 and in 2027? And how should we think about the differentiation of products that you have for Singapore versus the Taiwan capacities for 22- and 28-nanometer? And then with that, how should we forecast the depreciation in 2026 and 2027?
Well, first of all, for the year of 2026, the capacity increase will be around 1.2% year-over-year for us. The expansion of our Singapore facility will start in the second half of 2026 and will continue into 2027. In terms of the nodes available in our Singapore facility, our strategy is to have a geographically diverse manufacturing footprint between Taiwan, Singapore, Japan, and the U.S. and from a technology coverage standpoint, we would like to cover most of the nodes so customers can benefit from different sets of forecasts.
As for the depreciation forecast, we are looking at a low teens annual increase in the full year depreciation expenses. For next year, we don't have the exact number here, but it's very likely to be similar to the amount for 2026. So in a way, we will see the depreciation curve peak either this year or next year with very similar numbers.
Next one, Haas Liu, Bank of America.
Congrats on the results. I would actually like to follow up on the pricing. If we look at the like-for-like pricing environment, based on your current mid- to high 70 percentage of the utilization, if we strip out any of the considerations of product mix improvement, are you able to improve or just to pass on your higher manufacturing costs or material costs to your customers at this stage, or do you still receive a meaningful pushback from your customers?
Well, I mean, the pricing discussion is always ongoing. The overall pricing strategy remains consistent, as I mentioned earlier. In 2026, we see some market dynamic changes. Therefore, for certain customers, we do have adjusted pricing upward. For certain customers, we still have some of the one-time pricing adjustments at the beginning of the year to support their market share expansion, as well as competitiveness. Overall, we believe the environment is more favorable now in 2026.
Okay. Yes. So, when you talk about supporting your customers to gain market share by strengthening their cost structure, do you mean you are actually adjusting down your pricing for those customers? Or is it actually up for this year?
We have a mix of that. For certain customers, we have adjusted pricing upward. For certain customers, we will apply the one-time price adjustment downward.
Okay. Got it. And then just on the near term, a couple of your Fabless customers recently talked about earlier and also stronger inventory restocking because of the memory price hike. I was just wondering what impacts your first quarter outlook here, if your customers are seeing stronger inventory pulling in the traditional low season. Why is your shipment for the first quarter still relatively flat? And then, what's your puts and takes for the first quarter overall business outlook? Just wondering which part of the business is relatively stronger and which is weak?
For Q1, by segment, we are actually in line with our addressable market seasonality. We did not see significant changes due to the inventory restocking. But if you're looking into applications, we expect the revenue contribution from the consumer segment to increase driven by WiFi, DTV, and set-top boxes, while revenue from communication and automotive will decline due to softer demand for ISP and DDI products.
Okay. That's pretty clear. And then since you mentioned seasonality, are you expecting this year's seasonality to look pretty similar to the previous few years where the first quarter could be relatively light and the second and third quarters will see relative strength?
I can provide you with this. If we look at the whole year, with new projects on multiple specialty technologies across embedded high-voltage, non-volatile memory, power management IC, RF SOI, which supports the end markets in communication, consumer, automotive, and AI servers, we're looking at the second half of 2026 outperforming the first half. So that may deviate from the traditional seasonality. We believe the overall shipment for the year will be a growth year, and the second half will be better than the first half.
Okay. Yes. And last question before I jump back in the queue is that, just based on your comment, what is the underlying market unit demand assumption you have right now? Is it the smartphone market will actually grow or decline based on your current base case scenario that the second half will be better? Or is it already factoring in a relatively more conservative expectation that smartphone TV, PC markets will see a unit decline?
Based on current forecasts from our customers, we do see gains on product segments across all applications. We have observed share gains on those applications. The forecast shows that there's more of a share gain compared to the end-market demand.
Next one, Felix Pan, KGI.
I just have a couple of questions about the future growth drivers, particularly in your remarks about advanced packaging and silicon photonics. So my first question will be besides the Interposer, what else might we have engagement for advanced packaging? For Interposer, what's the capacity expansion plan for 2026? My second question will be about silicon photonics, particularly in the Singapore fab. There are rumors about potential customers. Is there any color or engagement we can expect or any contribution generated from this segment?
Okay. That's a big question. Let me see if I can cover that. Looking back, we have delivered solid performance in 2025 with a 12.3% shipment growth and 5.3% revenue growth, which outperformed our addressable market. This result is supported by our differentiated 22-nanometer technology and other specialty offerings across both 12-inch and 8-inch amid a broad-based market demand recovery. We view 2026 as a year of both continuity and evolution. We believe UMC will take shares again and outperform its addressable market. We will also see several positive trends. As our guidance suggests, we are seeing a more favorable pricing environment resulting from tighter global supply and our differentiated technology and geographical footprint, driving growth for the next few years. We're on track with our 12-nanometer cooperation with Intel, expecting to start seeing tape-outs in 2027. On advanced packaging, we see two distinct opportunities: enablers and 10 extenders. We're seeing 2.5D and 3D packaging as well as the chiplet technology being adopted beyond just the data center and ultra high-end chips, which will also be used in mature nodes. This allows for better power efficiency and differentiated products. Meanwhile, we are working with over 10 customers in advanced packaging currently and expect more than 20 new tape-outs in 2026.
Yes. Okay. But just let me quickly follow up and rephrase my question. So for silicon photonics, what's the earliest timetable we can see the revenue contribution?
For the 12-inch PIC and the pluggable product, we'll be expecting to ramp this year.
Okay. And about the Interposer, currently it's the bottleneck for our partner to expand their capacity. Is there any color we can give or how much capacity growth for the Interposer, like how much year-on-year growth or anything like that?
The capacity planning for Interposer will be aligned with customer ramp plans and market outlook. We will provide clarity closer to that, but in 2026, the focus will be on tape-outs.
Next, Gokul Hariharan, JPMorgan.
Could you go a little bit deeper into the advanced packaging comment that you made? What is the involvement level of UMC in some of these advanced packaging solutions? Are you doing full stack, or is it largely focused on Interposer? And in terms of the tape-outs you have, what are the nature of these tape-outs? Are these mostly data center ASIC-related products or is this a much wider array of products other than data center ASIC?
Sure. We have reported strong building capabilities in advanced packaging, including wafer-to-wafer stacking, TSC, and Interposer technologies. We provide advanced packaging capabilities to many of the current products. For example, we have wafer-to-wafer hybrid bonding with RF SOI solutions already in production for the mobile sector; we do not provide memory ourselves, so customers have to provide the memory wafer.
Got it. Any plans to further expand your Interposer capacity? I think you expanded it up to 6,000 and then kind of stopped it there.
There are discussions around that. We have common tools in place, which we can leverage from our 40-nanometer and 65-nanometer capacities. Unique tools for future expansion will be planned for the customer ramp profile, probably happening in 2027.
Got it. Understood. That's clear. Another question I had is on the expectations for communication and consumer segments, which represent over 70% of revenue. Given concerns about smartphone and PCs, how are you budgeting for this?
We remain cautious on that topic. Currently, we have not observed demand impacts on our customers' forecasts for the year despite the recent pricing surge. Our technology predominantly supports customers addressing the higher end of the market segment, where demand tends to be more resilient. We remain attentive to potential impacts on the memory market, but our current assessment is that any headwinds are manageable.
Got it. My last question is on the geographic split of revenues. Could you talk a little about the Intel 12-nanometer progress and how revenue or profits from this partnership will be booked?
The 12-nanometer cooperation project with Intel continues to advance smoothly. We're on schedule to deliver the PDK and associated IP in 2026, with product tape-outs commencing in 2027. This partnership represents significant steps toward commercialization and future revenue growth. The application areas for this project include products such as digital TVs, WiFi connectivity, and high-speed interface products.
Okay. And any thoughts for the Xiamen capacity?
For Xiamen, it's a core part of our competitive advantage, and the fab is currently operating at full capacity. We anticipate continued engagement and will optimize our customer engagement and product loading regionally.
Okay. One more on blended ASP. You mentioned the ASP environment is more favorable, but overall utilization is still in the mid-70s as of Q1. Do you expect blended ASPs to move up meaningfully like 5% to 10% this year, or does that require higher levels of utilization?
While high utilization is an important factor, it's not the only one. We must ensure our pricing strategy enables us and our customers to remain competitive. The pricing environment is indeed becoming more favorable, but we need to manage the magnitude of those changes. Any clarity will be shared with you as it develops.
Next, Alex Chang, BNP.
I just have a very quick question. I just saw that the company announced the start of mass production of SuperFlash Generation 4. How much revenue contribution has come from the non-volatile memory business in the past quarter or maybe past year?
We don't provide a breakdown for revenue contributions. Specialty revenue represents about 50% of our overall revenue, with high voltage accounting for about 30%. The remainder includes non-volatile memory and DCB.
Next is Laura Chen from Citi.
I want to follow up on the depreciation rate and also the gross margin outlook. Jason, you mentioned that the pricing environment seems to be improving. Together with firm shipment and better product mix, how should we think about gross margin trends? You guided that it will be in the high 20% for Q1, but with these favorable factors, what can we expect for margins throughout the year?
Gross margin can be highly dependent on utilization rate, ASP, product mix, depreciation, and foreign exchange rates. There are many variables, making it difficult to give a firm outlook beyond this quarter. For the first quarter guidance, we expect the gross margin to be in the high 20s, mainly due to higher costs, especially depreciation expenses, which are expected to rise by a low teens annual increase in 2026. These factors may pressure gross margins but we will work on cost reduction and productivity improvements to deliver a stable EBITDA margin.
I think advanced packaging and silicon photonics are key opportunities for UMC. We know you've worked on advanced packaging, especially Interposer. Could you share the current revenue contribution from advanced packaging and how that might look in 2 or 3 years?
Currently, Interposer is exposed to a limited customer base and narrow applications. We've engaged with over 10 customers and expect more than 20 new tape-outs in 2026. We foresee significant revenue growth from advanced packaging by 2027.
Next, we'll have Bruce Lu, Goldman Sachs for questions.
I want to go deeper into the silicon photonics topic. How large do you believe the addressable market for silicon photonics is for UMC in 2 years, and what are your competitive advantages in this space?
The Singapore facility will serve various applications beyond just silicon photonics and will play a critical role in our geographical manufacturing strategy. Our collaboration with INEX allows us to deliver industry-standard PDKs to our customers by 2027. Many competitors focus on 8-inch whereas we are focused on 12-inch, providing us a competitive advantage. We are also integrating silicon photonics with our advanced packaging know-how for improved applications.
What is the next step for collaboration with Intel for 10, 7 nanometers, and beyond?
Our current focus is to deliver the 12-nanometer platform. If it makes sense for both UMC and Intel as well as our customers, we will discuss expanding collaboration to other derivatives and technologies in the future.
Now we'll have our last question from Sappho Neuberger Berman.
It's been a while. Congrats on the progress you've made over the past couple of years. I have a few questions. First, on market dynamics, TSMC is shrinking or defocusing mature foundry processes, and other foundries are doing the same. It seems there's a lot of supply being taken away due to AI demand affecting older nodes while analog markets improve. How do you view current dynamics relative to 2021's severe shortage?
That's a great question. We observe significant market movement and have closely analyzed demand and supply outlooks. Whether short term or long term, AI remains a strong driver. This momentum will drive overall demand for the foreseeable future. The economic aspects also indicate that building mature facilities is not justifiable, suggesting we could sustain this situation longer than in prior cycles. We will continue monitoring the progress, ensuring we are positioned to capture opportunities.
Your earlier comments on pricing indicated you provide annual discounts to certain strategic clients, but pricing seems to be going up for most clients. Do you believe you could reprice those clients later if conditions tighten?
Pricing discussions are ongoing. We collaborate with our customers to reflect market dynamics and cost increases, so repricing will happen if conditions justify it.
A lot of that pricing was communicated before the supply-demand dynamics evolved. Am I correct?
You're right. Discussions have occurred considering conditions, including volume dynamics.
Thank you for attending this conference today. We appreciate your questions. As always, if you have any additional follow-up questions, please feel free to contact ir@umc.com. Have a good day.
Thank you. That concludes our conference for the fourth quarter of 2025. Thank you for your participation in UMC's conference. There will be a webcast replay within 2 hours. Please visit www.umc.com under the Investors, Events section. You may now disconnect. Thank you, again. Goodbye.