[Foreign language] Welcome to TSMC's third quarter 2019 earnings conference and conference call. This is Elizabeth Sun, TSMC's Senior Director of Corporate Communications, and your host for today. Today's event is webcast live through TSMC's website at www.tsmc.com. If you are joining us through the conference call, your dial-in lines are in listen only mode.
As this conference is being viewed by investors around the world, we will conduct this event in English only. The format for today's event will be as follows: first, TSMC's Vice President and Chief Financial Officer, Mr. Wendell Huang, will summarize our operations in the third quarter 2019, followed by the guidance for the fourth quarter. Mr. Huang and TSMC CEO, Dr. C.C. Wei, will jointly provide company's key messages. We will open both the floor and the line for the Q&A. For those participants on the call, if you do not yet have a copy of the press release, you may download it from TSMC's website at www.tsmc.com. Please also download the summary slides in relation to today's earnings conference presentation.
As usual, I would like to remind everybody that today's discussions may contain forward-looking statements that are subject to significant risks and uncertainties, which could cause the actual results to differ materially from those contained in the forward-looking statements. Please refer to the safe harbor notice that appears on our press release. Now, I would like to turn the microphone to TSMC CFO, Mr. Wendell Huang, for the summary of operations and current quarter guidance.
Thank you, Elizabeth. Good afternoon, everyone, thank you for joining us today. My presentation will start with the financial highlights for the third quarter, followed by the guidance for the current quarter. Third quarter revenue increased 21.6% quarter-over-quarter to TWD 293 billion , driven by new product launches, both in premium smartphone and High Performance Computing applications using TSMC's industry-leading 7- nm technology. Gross margin increased by 4.6 percentage points sequentially to 47.6%, mainly due to a solid improvement in capacity utilization. Total operating expenses accounted for 10.7% of net compared to 11.2% in the second quarter, due to better operating leverage. Operating margin increased by 5.1 percentage points sequentially to 36.8%. Overall, our third quarter EPS was TWD 3.9, ROE was 25.7%. Now, let's take a look at revenue by technology.
Seven nanometer technology saw very strong demand and accounted for 27% of wafer revenue in the third quarter. 10 -nm was 2% and 16 -nm was 22%. Advanced technologies, which are defined as 16 nanometer and below, accounted for 51% of wafer revenue, up from 47% in the second quarter. Let's take a look at revenue contribution by platform. All four of our growth platforms saw demand increases in the third quarter. Smartphone increased 33% quarter-over-quarter to account for 49% of our third quarter revenue. HPC increased 10% to account for 29%. IoT increased 35% to account for 9%, and automotive increased 20% to account for 4%. Moving on to the balance sheet.
We ended the third quarter with cash and marketable securities of TWD 585 billion, a decrease of TWD 180 billion from the second quarter, mainly as we distributed TWD 207 billion of cash dividends for 2018. On the liability side, current liabilities decreased by TWD 127 billion quarter-over-quarter as we distributed 2018 cash dividend and accrued another TWD 65 billion, or TWD 2.5 per share, for the second quarter 2019 cash dividend. That will be paid in January of next year. On financial ratios, accounts receivable turnover days decreased one day to 41 days. Days of inventory decreased 11 days to 65 days, reflecting higher wafer shipments during the quarter. Now let me make a few comments on cash flow and CapEx.
During the third quarter, we generated about TWD 142 billion of cash from operations, spent TWD 98 billion in capital expenditures, and distributed TWD 207 billion of cash dividend. As a result, our overall cash balance decreased by TWD 197 billion to TWD 452 billion at the end of the quarter. In US dollar terms, our third quarter capital expenditures was $3.14 billion. I have finished my financial summary. Let's turn to fourth quarter guidance. Based on the current business outlook, we expect our fourth quarter revenue to be between $10.2 billion and $10.3 billion, which is a 9% sequential increase at the midpoint. Based on the exchange rate assumption of one US dollar to TWD 30.6, gross margin is expected to be between 48% and 50%. Operating margin is expected to be between 37% and 39%. This concludes my financial presentation.
Let me follow by making a few comments about 2019 capital expenditures and TSMC's long-term financial objectives. I will first talk about our capital budget for this year. In TSMC, we build capacity according to our customers' demand. To forecast such demand, we take into consideration not only from each individual customer's indication, but also our own forecast based on macro as well as market segment outlook. Given the stronger outlook for 5G deployment next year, the demand for our 7-nm and 5-nm has increased significantly in the last few months. We have therefore decided to raise our full year 2019 CapEx by $4 billion to meet this increased demand. We now expect our 2019 CapEx to be between $14 billion and $15 billion. About $1.5 billion of the $4 billion CapEx increase is for 7-nm capacity, and $2.5 billion is for 5-nm capacity.
Although we're not able to give you a formal guidance for our next year's CapEx until next January, we currently plan next year's CapEx to be somewhat similar to our revised 2019 CapEx. Let me state our long-term financial objectives. As the company's new CFO, I'm happy to tell you that TSMC's long-term financial objectives remain the same. Our goal is to achieve revenue and net income together in the next few years to be between 5% and 10% in US dollar terms. Gross margin to be about 50%, operating margin to be about 39%, and ROE to be above 20%. Regarding our cash dividend policy, we reiterate that we will distribute about 70% of free cash flow as cash dividend. More importantly, TSMC is committed to a sustainable cash dividend on both an annual and quarterly basis. I will turn the microphone to C. C.
Thank you, Wendell. Good afternoon, ladies and gentlemen. Let me start with our near-term demand and inventory. We conclude our third quarter, as reported by CFO, with revenue of NT$293 billion, or $9.4 billion. That is slightly above our guidance due to better demand from smartphone related applications in our forecast three months ago. Moving into fourth quarter this year, we expect demand from both smartphone and high performance computing related applications will continue to increase, thanks to our industry-leading 7-nm technology that powers these applications. On the inventory front, our fabless customers' overall inventory is being gradually digested throughout the third quarter. We now expect it reduce to a few days above seasonal level at hitting third quarter, and approach seasonal level by the end of this year.
For the full year of 2019, we forecast both the overall semiconductor market, excluding memory, and the foundry segment to decline by a low single digit from their 2018 level. However, we continue to expect TSMC to do better and achieve a high annual growth. Let me talk about the progress and development of 5G. 5G will drive AI applications and bring many benefits to the market. The performance will be greatly improved with data transmission speed up to 10 times faster as compared to 4G network. In addition, 5G latency will have about a 90% reduction as compared to 4G, allowing for real-time response and control. The benefit from 5G will unlock new usage cases such as AR, VR, real-time translation, and high-quality gaming, to name a few. We believe smartphone OEMs will come out with many more innovative applications to take advantage of the 5G infrastructure.
Since the middle of this year, we have been seeing an acceleration in the worldwide 5G development. This was speeded up the introduction and deployment of 5G network and smartphone in several major markets around the world, which leads to the increase of our CapEx for this year. We expect a faster ramp of 5G smartphones as compared to 4G, with a penetration rate of 5G smartphones to reach 15% of the total smartphone market in 2020. Meanwhile, we expect the silicon content of 5G smartphones will be substantially higher than that of 4G smartphones. That is due to increasing functionality and additional ICs for more camera, RF circuit, modem, Power Management IC, et cetera. Power efficiency, speed, and ability to incorporate additional functionality are critically important to 5G smartphones, which require TSMC's leading-edge technology and will continue to fuel our growth for the next several years.
Now I will talk about our N5 and N3 status. Our N5 technology has already entered risk production with good yield. N5 will adopt EUV extensively and is well on track for volume production in the first half of next year. With 80% A0 logic density gain and about a 20% speed gain compared with the 7-nm, our N5 technology is a true full node stride from our N7. We believe it will be the foundry industry's most advanced solution with the best density, performance, and power until our 3-nm arrives. With N5, we are further extending our customer product portfolio and increasing our addressable market. The initial ramp will be driven by both mobile and HPC applications. We are confident that 5-nm will have a strong ramp and be a large and long-lasting node for TSMC. Now I will talk about N3.
We are working with customers on N3, and the technology development progress is going well. Our N3 will be another full node from our N5, with PPA gain similar to the gain from N7 to N5. We expect our 3-nm technology will be the most advanced foundry technology in both PPA and transistor technology when it is introduced. Now, I will talk about the ramp-up of N7+ and the status of N6. Today, we are completing our second year ramp of N7. We continue to see very strong demand across a wide spectrum of products for mobile, HPC, and IoT applications. Our N7+ is the industry's first commercially available EUV lithography technology. N7+ provides 15%-20% higher density with improved power consumption when compared to N7, and is already in high volume production with yield similar to N7.
We expect the strong demand for N7+ continue into next year. We are increasing CapEx to meet this demand for multiple customers. N6. Our N6 provides a clear migration path for the second wave N7 products as the design rule are 100% compatible with N7, while providing 18% logic density gain with performance to cost advantage. N6 use one more EUV layer than N7+. N6 risk production is scheduled to begin in fourth quarter next year, with volume production starting before the end of 2020. We reaffirm that seven nanometer will contribute more than 25% of our wafer revenue in 2019. We expect even higher percentage in 2020 due to worldwide development of 5G accelerates while demand from HPC, mobile, and other applications continue to grow. I'll talk about TSMC's advanced packaging business.
Our advanced packaging solution enable system integration with wafer-level process, allowing seamless integration of front-end wafer process and back-end chip packaging. The solution consists of CoWoS, InFO, System-on-Integrated-Chips or SoIC, and wafer-on-wafer or WoW. We are seeing strong momentum for CoWoS and InFO for HPC applications as we continue to enlarge the integrated chip area to above two vertical sides in one module. We are also working with a few leading customer on SoIC, which is an industry-leading 3D IC packaging solution. SoIC enables 3D integration of multiple chips in close proximity to deliver the best possible performance, power, and form factor. We target to start production in 2021 timeframe with early adoption by HPC applications.
As the industry continues to seek innovation to enhance system-level performance, TSMC's differentiating advanced packaging solution will allow us to grow the business at a pace faster than corporate in the next few years. Thank you for your attention.
This concludes our prepared statements. Before we begin the Q&A session, I would like to remind everybody to limit your questions to two at a time to allow all participants an opportunity to ask their questions. Questions will be taken both from the floor and from the call. Should you wish to raise your questions in Chinese, I will translate it to English before our management answers your question. For those of you on the call, if you would like to ask a question, please press the star then one on your telephone keypad now. Questions will be taken in the order in which they were received. If at any time you would like to remove yourself from the questioning queue, please press the pound or the hash key. Now we begin. First, we will have questions coming from Credit Suisse, Randy Abrams.
Thank you. This is Randy. I want to ask the first question, given the high CapEx for this year and next year, if you could talk about the capital requirement for 5-nm and EUV. Maybe how the equation with EUV coming in, how does that affect the capital intensity or change the CapEx per 1,000 wafers? How do you see this node in terms of pricing and maintaining the same level of profitability in prior nodes?
You answer that?
Yeah. Okay. Randy, you're asking about how much EUV accounts for the capital expenditures?
I wonder how EUV changes your capital intensity or CapEx per K. If you're getting some benefit that is actually lowering that. Given the CapEx, if it's still the same, you're seeing the higher CapEx per K, then how is that translating to profitability? How do you expect five to compare to, say, some of the prior nodes?
Well, I cannot talk too much about details. Definitely, CapEx per K for five is certainly higher than previous nodes. If we combined everything together, as we stated at the beginning, we're still seeing our profit, structural profitability remains the same.
Okay. I'll ask one follow-up to that and then a second question. This year, second half strong demand, leading edge tight, the gross margin guidance for fourth quarter is still, it's a good margin, but it's 48%-50%, and I think the target's going to hit 50%. Maybe if there's still a little bit of drag on the margin, and with next year bringing in a new node five, if you think you can get to that 50% for next year. That's just a follow-up for me.
Okay. For the fourth quarter, it is just like we stated all along for our structural profitability. When you reach 90% of utilization, we target to reach 50% of gross margins. For next year, it is still early to say. The structural profitability or margins may be affected by the ramp in a five nanometer. As at the beginning of ramping of every nanotechnology nodes, it will somehow be affected. However, if the utilization is good, then we are still expecting to see the structural profitability continues.
Okay. The second question I wanted to ask. You had a strong third quarter recovery, guiding another good fourth quarter. As you look to maybe an early look at next year, how should we think about seasonality? There's some concerns before 5G ramps. We have a correction on 4G, and then we have, if the tariffs hit, we might have had some pre-built. How do you see early year risk of, say, a correction? Can you pull in some demand, say, a capacity sighting to maybe mitigate the magnitude of correction early year?
You are talking about the seasonality of the smartphone. That we observe almost every year. I don't expect next year to be dramatically different. However, we could see some of the 5G smartphones, the growth momentum is higher than we expected. I would expect next year's seasonality is not so strong as we observe for this year. It's too early to say, because of the market is very dynamic.
Thank you.
Next question will be coming from JP Morgan's Gokul Hariharan.
Thanks for taking my question. First of all, if you look at the history, whenever TSMC has had a step up in CapEx, that is typically accompanied by a step up in growth as well. Just wanted to kind of narrow down a little bit on the 5% to 10% growth, which is still the kind of growth that we were expecting when we were expecting TWD 10 billion-TWD 11 billion. Could you give a little bit more detail or maybe narrow down the focus a little bit more for us? If we say a TWD 14 billion-TWD 15 billion range of CapEx, that's closer to the high teens in terms of capital intensity, higher than our previous range.
Gokul, let me answer the question carefully. Let's say that TSMC always build capacity, working closely with the customer and to meet their demand. That's number one, okay. We discuss with the customer on their demand, we make our judgment also. Now, we are increasing the CapEx quite a lot, no doubt about it. That's due to some of the reason I can foresee for the future. First, the 5G ramp-up is much faster than 4G as we expected. Second, TSMC actually is expanding our customer portfolio. In the same time, we're also expanding our product portfolio. So put all the factors together, we have a good reason that we increase our CapEx this year and probably next year.
Okay. That basically suggests that 5% to 10% is not the growth. Maybe just given the higher growth in 5G and accelerated build out, could you talk a little bit about going to next year? Do you expect 5G and smartphones to be the main growth drivers? Or do you feel that HPC momentum is going to be even faster than smartphones and that some of your customers are starting to take a lot of market share?
To answer your question, actually it's both. We expect the smartphone to grow faster than... not in terms of unit, but in terms of silicon content. The HPC also grow. As I said, we are expanding our product portfolio and also our customer portfolio. The addressable market is increased. That's all.
Okay. Just a follow-up question. At the beginning of the year, we had a more conservative view on 5-nm build out. Obviously, that view has changed. Just to calibrate, should we expect now that 5-nm revenues next year are likely to be higher than, let's say, 7-nm last year, first year of 5-nm?
I don't want to say exactly what is the percentage. Let me say that now we have more optimistic than six months ago. Six months ago, I believe what I said is that we want to be very careful and a little bit conservative in building the five-nanometer capacity. Now we change to be more aggressive on the 5-nm capacity build-up because as I said, we work closely with customers in both all the applications, like a smartphone, HPC, even IoT and automotive.
Okay. Thank you.
Next question will be coming from UBS, Bill Lu.
Thank you. First of all, Mr. Huang, congrats on the new post, and looking forward to working with you. Going back to CapEx, given the big increase, can you give us some guidance for modeling depreciation?
Depreciation for this year would be flattish compared with last year, because the big chunk of the increase happens in the fourth quarter. As to depreciation in next year, based on what I just indicated, our capital expenditures for next year, it will be higher. However, the detailed number or more specifics, we plan to discuss it in January next year.
If I look at when the new capacity is going to come online, is it mostly, say, Q2, Q3? Is that the right way to think about it?
Well, if you look at towards this time, your estimate is very good, Q2, Q3 time.
Okay, great. Thank you. By the way, for 7-nm , obviously demand looks quite good. Can you give us some guidance for 7-nm as a percentage of revenues by end of next year?
Higher than this year.
Okay, thank you.
Next question will be coming from Citig roup's Roland Shu.
Good afternoon. Thanks for the very good result in 3Q. I think the first question is that you talk about the next year for 5G smartphone will be the key growth driver. You talk about not only because of unique growth but also for the semiconductor content increase. Do you have the number for the dollar content per 5G smartphone compared to the 4G?
That is CFO to answer this one.
The answer is we don't have the number.
If you compare to the 4G, where does the increase come from? Is that because from this bigger die size or because of more semiconductor features? You probably have an assessment on that, right?
We do have some assessment, to be frank with you, because we know the die size, we know that the number of the chips inside. I give you some of the feeling. We are using the more advanced technologies, so the die size increase on the leading edge is not so much. It did increase because the more functionality, you can expect the people put the AI application inside or those kind of things. More important, at the last time, I believe I already mentioned that the camera, for one example. Now you can see more and more camera in a smartphone, especially the high-end smartphone, and also the more resolution. The other one was the silicon real estate to form 20 MP-40 MP. More camera, higher resolution, more pixel.
Also you need a lot of Power Management IC to control the power consumption because of 5G, they consume a lot of power. Your Power Management IC has to be more advanced. You put more Power Management IC inside. You are talking about more channel in the communication. The RF front end, the transceiver, everything, not only the die size was increased actually to adopt the more channel. It would move into the higher leading-edge technology also.
Thanks. I guess for the high-end phone, probably, the dollar content will increase more. For your total 5G smartphone, you need focus to be about mid-teens of the total shipment. Do you have this 5G smartphone breakdown by high-end and low-end?
We'll talk about it in January.
Okay. Thank you. My second question is, now we are raising N7 plus 5G. Is the N7 plus gross margin going to be very different from N7 we are enjoying now?
It's very close.
Okay, thank you. How about for N7+, is it going to take seven to eight quarters to bring up the gross margin to corporate average?
Let me take that. We look at the N7+ , and N10 as the big nodes. From that point of view, N7 nodes actually has reached corporate margins.
Okay, thank you. Just a follow-up. I know you don't comment on the ASP, but for the same amount of the wafer shipment on N7+, is this going to contribute more revenue upside to TSMC?
We just mentioned we don't-
No, I talk about revenue. I don't talk about ASP.
The same thing.
Okay, thank you.
Next question will be coming from CLSA's Sebastian Hou.
Thank you. My first question, I want to ask about the CapEx intensity. With the hike in this year's CapEx, currently they give that the CapEx spend will go up to, you showed over 40% this year. How do you see this trend into 2020? Is there any updated guidance on the CapEx intensity going forward versus prior guidance?
Okay. Yeah, as you just mentioned, we can calculate the CapEx intensity this year will be over 40%. At the same time, as we indicated that next year's CapEx, although preliminary, is about the same level as this year, and we also see strong demand for our business next year, our CapEx intensity will be lower next year. From what we can see at this moment, it will then gradually come down to probably between 30% or 35% level after next year. You are now saying going to 30%-35% range in 2030 year?
In a couple of years after 2020.
That's potentially next year. You're just saying you don't want to talk about that right now.
Right. The potential of capital intensity, you mean?
Lower to mid 30% by 2020.
It's lower than 40%.
Okay. Thank you. Second question is on the gross profit margin. You told us a couple or two quarters ago, first half this year, when TSMC powered a table saying that we are past the bottom, we're seeing the strong recovery second half this year. Also, expect the gross margin will go back to 50% level second half. Indeed, now we see the ASPs already coming back, revenue coming back from second half as expected. It looks like the gross margin is lagging a little bit compared to the revenue direction. Can you explain what is this factor that driving gross margin recovery behind the revenue recovery? Thanks.
Yeah. I think you're talking about the third quarter gross margins. Whenever there's a big jump up in utilization quarter-over-quarter, there will be a negative hit on gross margins. It's called the inventory valuation. I believe we mentioned about this some time ago. Yes, you can imagine the utilization in third quarter is much higher than that in the second quarter, and therefore, there is an element of negative inventory valuation happen in the third quarter. The amount is slightly over one percentage point. Okay. Now, if utilization remains similar level, that factor will not happen.
Okay. When we're looking at the fourth quarter gross margin guidance, I know the high end will reach the 50%, but it still looks like it ranges between the 48%-50%. I think there's still a chance for it to go back to 50%, but it looks like the midpoint doesn't reach that yet. I think not just about the 2P, but also the fourth quarter.
Yeah. As I mentioned earlier, our structural profitability remains the same, i.e., if we reach 90% of utilization, we are targeting 50% of gross margins.
Okay. Thank you.
Next question will be coming from Morgan Stanley's Charlie Chan.
Thank you. Congratulations for a very strong result. Sorry to keep coming back to this question. Since you are raising kind of a long-term CapEx outlook, including next year, why don't you just revise up your revenue CAGR assumption? I think 5%-10% revenue should be too conservative. Can you comment on that part?
It's not about time to change the target. I would say still 5%-10%, but it's in the upper side.
Okay.
I can say.
All right. Thanks. Yeah, another one on this capital intensity question is about the payback period, right? Because for me, I still cannot add up why your structural profitability can remain the same, whereas your capital intensity increased so much, right? In other way, can you just comment about the payback period for your 7-nm and 5-nm investment? Is that becoming longer going forward?
I remember this topic was discussed last time. I think our answer was that we don't look at payback period. We do look at return on invested capital, and from that part, we don't see that big a difference with the advanced technologies.
Okay. Thanks. My next question is about the advanced packaging. I remember in a few recent quarters you commented advanced packaging should outgrow the front-end business. First of all, is this remains the same trend? Also, how about the potential margin dilution from the packaging business?
The forecast on the advanced packaging business, the growth rate is still faster than the silicon growth rate. The wafers revenues growth rate. It stays the same.
Okay. Still, that statement still valid?
The gross margin, that's another consideration. The gross margin of the back-end business actually is lower today, still lower than the wafer margin. We look at it, whether it's a good business to go or not on two factors. One, we really want to support our customer to improve their system performance. We have to do it because TSMC is the only one company right now that can support customers' advanced packaging. Second, actually, is the CapEx intensity on the back end, and that's the advanced packaging business, is smaller. The asset turnover is better. Put all you all together, we still think it's a very good business to pursue.
Yeah. Last one, this is a quick one maybe for CFO. You revised up your CapEx, right? What is the free cash flow trend in the coming quarters? I guess, we want to know the next potential quarterly dividend hike. What would be the timing?
Let me just answer it by saying that we will not lower our dividend quarter-on-quarter basis. The investor will get at least the same amount, if not more, compared to the previous quarter or compared to the previous year.
How about the free cash flow projection? Do you have the number for coming quarters?
Yes, we do. I don't think that really matters to your question.
Okay, thanks.
If I may, I would like to remind Charlie that today all our shareholders are getting the TWD 2 per share dividends today, going into their bank account. We have already announced that the next quarter's dividend is going to be TWD 2.50, so it has already gone up. Next question will be coming from Goldman Sachs, Bruce Lu.
Can I ask, what was your 5G penetration forecast about six months ago? 5G smartphone penetration forecast six months ago.
Six months ago, we don't think the 5G smartphone will be any significant amount for this year.
No, I mean, six months ago, what's your forecast for 5G smartphone shipment in 2020?
Oh. Single digit.
Basically, you increased from single digit to mid-teens-
Yes
...for the past six months.
Yes.
Can you tell us what's the rationale behind your changes? For the past six months, we still have a lot of concerns with the 5G smartphone, such as NSA, SA, the specs are not finalized, the cost structure is getting a lot higher. Global telcos, they are not as aggressive. Because from single digit to mid-teens, which is a big change.
It's a big change. All I can say is whether it's like NSA, SA or those kind of 5G base stations and installation, we work with our customer, actually, we listen to them. They also have their own customer to consider. Some of the areas in the world, they accelerated the 5G deployment. That's why it result as compared with the six months ago, we did not see this momentum. In this six months, the momentum is going bigger and bigger. I cannot give you the specific number of which country or which region, but we can see the momentum continue to grow. Now our own estimate, it's almost doubled.
The concern is still there from a lot of investors, such as NSA or SA, because you still cannot see the clear indication at the end of the day from the telcos, right? That's why we are a bit surprised to see this kind of meaningful changes.
Yeah. I have no way to comment on the carrier, how they think about this kind of 5G deployment. They are moving ahead. For the 5G applications, especially in some of the big countries. That's what we see.
Okay. My next question is that we always rely on management to comment about the fabless inventories. You always talk about that. Moving forward, your top two customers, it might account for 30%-40% of your business pretty soon. They might account for 70%-80% of your advanced geometry. They don't really report their inventory anyway. Can you give us some other indicator for us to judge the industry growth or judge the inventory level for the industry?
Judge for the industry level, the fabless company is still one of the big factors that we consider. As for my customer, actually, for TSMC, the more importantly, we really work with our customer closely. I'm not going to give you the exact number of my big customers' inventory, but we work with them. Believe me, we are working closely work together, so we understand their strategy and their inventory.
That's why I said I wanted you to give it some heat, right? We cannot just tell my investor that we have to trust TSMC, even though I say that all the time.
You can trust TSMC. No doubt about it.
Thank you.
All right. I think this is probably about the right time that we go to the lines for questions. I think there are quite a few analysts waiting on the line for questions. Operator, could you please go to the first person on the line? Thank you.
Yes. We have a question from Brett Simpson from Arete Research. Please ask your question.
Thanks very much, Elizabeth. I had a question really on China. I guess in the last couple of years, we've seen your business double with Chinese customers. I guess at the moment, it's pretty clear you're going through a very healthy inflection point with Chinese customers at the moment. Can you talk about how you see this part of your business evolving over the next one or two years? I guess from a planning perspective, are you concerned that the rise of your China business comes at the sacrifice of other customers, particularly U.S. companies? Thanks so much.
Brett, your question is with respect to the business we derive from China. You have observed that our business has doubled in the last couple of years from China, and you'd like to see the outlook for our business in China in the next one to two years. The strong growth coming from China, whether it is coming at the expense of our other customers in other regions.
Well, we can see the strong growth from China because that's a very big market, especially in the semiconductor area. We are happy to see that growth. TSMC is offering the most leading-edge technology to support our customers in China. To be exact, we are going to grow with the China market. At the expense of other customers? The answer is no, because we support all the customers with all our strength and our capacity.
Oh, okay. Maybe just to follow up on 5-nm specifically, I guess maybe you can talk about whether you think the ramp of 5nm will be similar to prior nodes in wait times or revenue. How should we think about the ramp up, specifically as we get through into the second half of 2020? On your CapEx increase, I think you've said in the past, you plan to grow. You've been quite consistent. You want to grow your top line 5%-10%, and your capital intensity to support that growth would be around about 30%, maybe low 30s as a percent of sales. Now you're stepping up your capital intensity significantly, you're not changing your growth outlook. Can you perhaps just explain your thinking there? Thank you.
First part of the question is with respect to 5-nm ramp, whether the ramp profile next year will be similar to our prior nodes ramp profile. The second question is related to the substantial increase in the CapEx, because in the past, TSMC has indeed said that we could have TWD 10 billion-TWD 12 billion CapEx to support a 5%-10% CAGR. Our CapEx is substantially higher than the TWD 10 billion-TWD 12 billion, and therefore, does that mean that it's going to support a much higher CAGR on the revenue growth?
Well, let me answer the first question first. The 5 -nm ramp for next year, certainly as compared with the six months ago, we are right now is a more aggressive and more optimistic about it. Hopefully, because we spend a bit money, hopefully that our ramp up in terms of revenue will be much faster than 7 -nm. In next January, we are going to talk about it more. So that also answers the second question that we spent, we increased the CapEx quite a lot, of course. From TWD 10 billion-TWD 11 billion to about TWD 14 billion-TWD 15 billion. With that money we spent to buy the tools, to prepare everything, we do expect that our growth is what go beyond 5%-10%. Right now, we are not ready to change the long-term, five years target yet. However, we are working on that.
Okay. Thank you very much.
Operator, please go to the next caller on the line. Thank you.
Yes. Your next question comes from the line of Mehdi Hosseini from SIG. Please ask your question.
Yes. Thank you for the opportunity to ask a question. A couple of follow-ups. When you make a reference to 5G. How should we think about opportunities from the networking, specifically base station? Thanks for the detailed color on the smartphone units. You expect mid-teen penetration. How would you quantify opportunities from the networking, specifically base station? I have a couple of follow-ups.
Mehdi, your question is that although we have talked about 5G smartphone units for next year, but you would like for our management to talk about the 5G base station business next year.
Yes.
Actually, the networking is fundamental for the 5G infrastructure. Because it's a shorter wavelength, so that you can expect that the base station will be much more than 4G base station on the SA or even for the NSA implement. We expect the networking process will be much higher. The opportunity will dramatically increase. For the executive, for the number, we don't have an exact number for right now.
Is networking included in the HPC category?
Yes.
Okay. The reason I ask this question is, let's say, if your customers are building too many 5G smartphone, could opportunities in networking actually help offset any downside risk to excess inventory built on a smartphone?
I think maybe you are asking whether or not the increase in demand in the networking side will offset the higher inventory or the inventory correction on the smartphone side. Is that your question?
Yes. If there is an inventory, I'm trying to better understand how networking could offset any risk of inventory built.
I did not catch what you say the relation between the networking or the base stations are set up with the 5G smartphones inventory.
Sure. Let me clarify. I think one of the concern among investment community is, yes, there is a very strong build plan for the 5G smartphone. The risk is these phones are too expensive and perhaps there is a risk that your customers' build plan are too aggressive. In that context, could opportunities on the networking side be large enough for TSMC to help alleviate any downside risk on the phone or excess inventory on the phone side?
I see. Let me repeat your question. You say that my customer are making the smartphone quite a large number. If the base station or networking did not catch up, is that the smartphone going to be inventory? The answer is no, because of the 5G application will be implemented, and you bought a 5G smartphone, you still can use a 4G functionality, by the way. It won't be a kind of inventory as you mentioned about it, provided you get not enough 5G base station.
Mehdi, I think what C.C. is saying is that you don't have to worry too much about 5G smartphone inventory because there will be sufficient demand to take up those products.
Yes. Thank you. Just a very quick follow-up. In the past, you have talked about advanced packaging accounting for high single-digit of the overall revenue in 2019, and that mix would grow into the teens in the next decade. Are those targets remain intact?
The advanced packaging used to account for about high single digit percent of TSMC's revenue. Mehdi's question is, will that continue to increase to maybe high teens of TSMC's revenue?
It will continue to increase, but it's not a high teen. It's still the high single digit, but the growth rate is higher than the silicon wafer's growth rate.
Okay, thank you. Is cryptocurrency stable? Should we be concerned that maybe some of this growth is driven by crypto? Do you see crypto as pretty stable and with minimum downside risk?
Well, right, you are talking about cryptocurrencies mining. We did see the Bitcoin's prices increase starting from this year until now. Let me state TSMC's policy and strategy. Cryptocurrency mining is a business with available capacity. That's all we say. We are not going to add in more CapEx for that.
Great. Thanks so much for answering the questions.
Thank you. Operator, could you please go to the next caller on the line? Thank you.
Your next question comes from the line of Sangam Iyer from Consilium. Please ask your question.
Yeah. Hi. Thank you for the opportunity. I just wanted to understand, given the recent developments that are happening in Hong Kong, and the elections that we are heading in Taiwan next calendar year, do you see that you might have to revisit the CapEx plans given the trade wars and the intensity of the trade wars that are increasing?
You think about the trade war between the two big countries, is that going to affect TSMC's CapEx plan?
Yes, in terms of the.
In terms of Hong Kong situation or what you think about it. Let me say again, TSMC build capacity working closely with customers to meet customers' demand and our own judgment. We did not put that kind of trade tension in the world into consideration. Although we think that any trade tension or trade war between any countries will have a negative impact to the semiconductor industry.
Right.
That is New York.
Hypothetically, if we were to assume that in the election that comes in, presidential election in January, if things were to go towards a party that's in more favor of a one state kind of a thing, would that have an incremental impact in terms of how the trade war is seen from TSMC's perspective?
No, we don't think so. Even we have a general election in the next January, that won't affect TSMC's strategy. Because we do the business according to the demand, according to the market situation, and have very little impact from the politics in Taiwan.
Okay. Good. Thank you.
Yeah.
Sir, finally. Yes, to a large extent, yes. That was helpful. One more thing, sir. In terms of the 5G CapEx that we have been talking about, there have been, on the high speed HPC segment, there have been delays in terms of the chips at the next node coming through from your large customers. When you see that 5G rollout will be pretty strong, and also the HPC side things will pick up strongly, is there any indication from your customers that the launch on the next node is going to be pretty soon? What gives us the visibility here?
Let me try to see if I understand your question. You are asking us about the visibility of our customers' demand for our next node technology, how much visibility we have with respect to seeing the demand from them for the next node.
Oh, next node.
Right.
You are mentioning about the 5-nm and the 3-nm?
5-nm for now.
Yeah.
Let's say 5 -nm. Judging from we increase the CapEx for the 5 -nm, you know our position and you know our forecast on the 5 -nm business. It will be good, and I cannot tell you how many percentage more. I think the 5 -nm TSMC going to do a very good job and going to have a high market share.
Great. Thanks a lot, sir.
Thank you.
Thank you. Operator, let's move to the next caller on the line.
Your next question comes from the line of Susie Chu from S&P Global. Please ask your question.
Hi. Thank you for taking my questions. I just have a brief question follow-up on the China issue. I was wondering, is there any change in the numbers of your Chinese clients in recent years, especially in this year? If it is possible that we can get a specific number of the Chinese customers or the percentage of the Chinese customers as of the total clients. Thank you.
The number of customers in China actually is more than 100. Some of them changes all the time. There are some bigger ones and smaller ones. I'm not sure if this is your question.
Yeah, sure. Is there any change of this number in this year?
The number of customers generally increases.
It increases. May we know the portion of the Chinese customers in terms of your overall customers or the percentage?
Oh, you mean the revenue or number of customers?
The number of customers.
We have over 400 active customers globally. If you look at the China, if it's somewhere over 100, that gives you an idea.
Great.
Does that answer your question?
Yes. Great, thanks. Very helpful. Yeah, that's all my questions. Thanks.
Okay. Thank you. Do we have any questions on the floor? Okay, let me come back to the floor. First of all, we will have a follow-up question from Citigroup's Roland. Roland Shu.
Thank you. I would like to follow up on the inventory question. However, this is for your own inventory. I look at in 3Q, your inventory level actually have been decreased to TWD 97 billion. Even though we had a very strong revenue in 3Q and a very upbeat guidance in 4Q. The wafer in process should increase. However, overall inventory decreased a lot. Does that mean that our finished goods and also even for our materials like wafer, inventory decreased a lot?
You are right. Actually, let me If suppose that the process actually enlarge, so the inventory should increase, but it's decreasing. It's because of some of the wafer, we work with our customer at the beginning of this year. We produce it in early stage to make sure that we don't have a wafer counted at the end of this year, and then we ship it out. The inventory actually decreased quite a lot because of that.
Means the finished goods?
Yes. Finished goods.
Okay. How about for the wafer? I mean, for the raw wafer inventory.
The raw wafer almost stays the same because our TSMC's business is increasing. Actually, the raw wafer, we require a lot of amount in preparation for the good business. It stays the same in terms of base. I'm sorry.
Okay. For the very upbeat outlook, next year for this 7-nm and 5-nm, are we securing all of these wafers for our production next year?
You bet.
Okay. Thank you. Okay. The second question is that, you announced that your EUV tools have been reached at production maturity. How about for the infrastructure? Means about other components like the photoresist, chemical, photomask, or even for this inspection tool, chemical, and materials. Yes, we are going to have very best trend on dynamics because of very strong demand from customers. Are there any gating items for this EUV infrastructure, would be a potentially a risk?
So far, we do not see any gating item. All the infrastructure, actually, TSMC, we are prepared. We produce our own chemical. We have a large number of masking capacity and everything. Even photoresist, those kind of things, we have been taking into account. We are ready. Actually, we are in a high volume production for the EUV lithography technology. For next year, you will see even higher volume, and I can assure you that we are all prepared.
Next question will be coming from Credit Suisse, Randy Abrams.
Okay. Yes, thank you. I want to ask, because you kind of hinted the utilization still below 90%. To get to 50%, you need to get utilization up. Could you talk about the 28 node prospects, how you see that? Also, 16 and 12, some of the applications likely move on, some of the mobile and graphics. Your view on utilization backfill to hold up that technology node.
Let me answer that question. On 28 -nm this technology node right now, I believe in the industry is over capacity. The utilization is very low, not to TSMC's expectation. We prepare a lot of the revenue technology, specialty technology, let me say that. We expect two years later from now, it will be again in the high utilization, but not today. It takes time to work with the customer and to utilize the capacity. For 16 and 12 node, today we are still in a very healthy utilization, but we are prepared. If our competitor is continuing to increase the capacity just like the 28 -nm node, we expect a couple of years later, something like that, it will be over capacity again. Right now we are preparing all the specialty technology like RF, like even some of the ISP or something like that.
It won't happen to TSMC.
Okay, the second question, just to ask about the segment revenue. I think into third quarter, mobile and HPC were both going to be strong. The HPC was good, but it wasn't quite as strong. I'm curious, just in that mix, maybe factors for HPC. IoT was quite strong, so maybe some of the things you're seeing in that. If you can tend to give an outlook for fourth quarter by segment.
In the first quarter, of course, every segment is an increase. The HPC itself consists a lot of different market segments. That including network processing, including GPU, including accelerator, and including all the cryptocurrency mining. We expect that HPC will grow significantly in the fourth quarter.
Could you give the other segments?
Other segments?
Any lagging and any picking up some segments.
The IoT, the third quarter of the IoT increased quite a bit. In the fourth quarter, probably it will be level or slightly decreased. While all other segments will continue to grow.
Great. Thank you.
Next question will be coming from UBS, Bill Lu.
Yeah. Thank you. If you look at the foundry industry historically, what we typically see is a big node followed by a smaller node, right? Essentially what you're saying is a big 7-nm and a big 5-nm. Can you maybe just discuss what you're seeing that is different now versus before? Clearly, TSMC is addressing new markets. I think the demand drivers are different now versus before. I'm curious as to what you're seeing that is different now.
You mentioned there's a big and then followed by a small node. I don't really catch what you mean, but let me say that TSMC introduced a 16 -nm and then 12-nm, and then 10-nm, and then 7-nm, and then 5-nm. Let's talk about 7-nm and 5-nm. We are looking at the new market in the 5G, because the 5G, the requirement on the speed and the power consumption reduction is quite a lot. It's not the same experience that we had before. After the seven nanometer, all the customers asked us to develop the technology to meet their requirement, which is a higher speed, lower power consumption. It must be a full node, kind of, improvement. It's not to say that a 10% improvement will be good enough. Look at our seven nanometer.
We have a 7-nm, we have 5-nm, then we have a 6-nm. The 6-nm being introduced as kind of for second wave product. The leading wave product always at a full node and much in ahead, with about two years of cadence that we expect.
Next question will be coming from JP Morgan's Gokul.
Thank you. If you compare the change in expectations in the last six months, would you characterize almost all of it as market growth, faster market growth for 5G, HPC? Or is there any meaningful change that in your market share expectation for some of your future nodes as well?
Well, compared with six months ago, we are really kind of conservative at that time, because of our 7-nm utilization is quite low, so we become conservative. In these six months, a lot of things changed. Let me say that, first, the 5G's momentum is larger than we expected. The second one is, we also at the same time, because our technology offer to the customer, we expand our customer's portfolio. Because of the performance, again, we also expand to new product portfolio. Now we look at the future, we are more optimistic than six months ago, much more.
Any increase in market share that you expect compared to the feedback?
No, I cannot comment on that, but it will be increasing, right? If you expect.
Okay. Quickly on the breadth of customers on N5. I think around this time one year or nine months before N7 ramp up, you were talking about roughly, I mean, 50 to 100 tape outs on N7. Could we talk about what is the breadth of number of customers or number of tape outs on N5? How do you expect that to progress? I know that number of tape outs is not equal to number of revenue or wafer volume, but just want to understand the breadth of it.
Well, let me say that tape out is one thing, that see how popular it is. The more important is the high volumes of products that tape out. Right now we have many high volume products that tape out, and that is a main reason why we increase our CapEx. High volume, and also in addition to smartphone, we have more market segment that we enter at the end.
Okay. Just one quick follow-up. If you think about N7 family, N7+, N6, is it your expectation that pretty much most of the N7 current customers will eventually end up using N6, N7+? Will most of your capacity eventually be EUV enabled for N7 as well, or it will still be some EUV, some still using the current N7?
We still expect some using the EUV and some still stay in the N7, because for those customer in the N7, one day, in a two-year cadence, they move to N5 already, so they have no reason to go back to use that N6. As I said, for the second wave of the product, we're using N6 with the benefit of low die cost and better performance.
We should expect eventually majority will be N6.
Yes.
Next follow-up question will be coming from Morgan Stanley's Charlie Chan.
Thank you. First of all, I want to follow up Bill's question on the future technology. You mentioned that there are lots of benefit from power consumption, from performance, from 5 -nm, et cetera. How about the per transistor cost? I remember some of customers talking about, they don't see the benefit from the transistor cost saving going forward. Do you think that is true statement?
I still believe the per transistor cost is decreasing, because of right now, geometry is smaller and smaller. Although for TSMC, we more pay attention to what customer need. Because of they need the speed, we give them speed. They need low power consumption, we give them the low power consumption. Higher density is always the one that we are moving into the next generation's technology. Per transistor-wise, I did not do the very detailed calculation, but I still think that the transistor, the cost is lower.
Okay. Thanks. You also mentioned that for those new tape out, actually volume is quite big. Can I interpret in other way, meaning those smaller customers, they cannot really afford those future technology unless they have a very big volume. Is that the right thinking?
Not really. That depend on their product. You asked question about the transistors, of course, let me give you some kind of a taste. We improved the logical density by 80%. I did not charge my customer 80% more. You know that the per transistor is lower. Again, in the market, I would say that a product to be successful is more important than you really calculate this improvement, this cost, that cost. Product has to be competitive in the market. That people want the price that they deserve. Okay? Not the small customer cannot afford it. Actually, some of the small customer, they are working on the CPU kind of a performance. They need a very high performance technology. That they are working with TSMC.
Okay, thanks. The next is on EUV. Over the past three years, there were some kind of challenge in terms of power, throughput, pellicle, et cetera. Now do you see any kind of new issues, new challenge for EUV or from now on it's like a blue sky?
Okay. In TSMC, EUV extreme ultraviolet technology is now in the production stage. Are we happy with it? Not yet. We are still improving the availability. We have a output power of 250 W as we expected. Now we can operate the tool with a 250 W consistently. However, there's still something that we need to improve so that we can improve the throughput, we can improve the availability, so we can reduce the cost. We continue to improve.
In the interest of time, we'll just have the last question that will be coming from CLSA's Sebastian Hou.
Thank you. I want to follow up, first point, I want to follow up on EUV. In oversized ASML, which you talked about on supply constraint, on the EUV tool. It looks like the guidance, really no expectation for next year EUV unit growth. Doesn't seem to be amazing either. I wonder what would that impact TSMC's, or TSMC, have TSMC secure what you need for this TWD 14 billion-TWD 15 billion CapEx this year, next year?
The answer is yes, we secure whatever we need. We work with ASML very closely, and we are ordering all the tools that we need.
Which means the supply constraints could affect someone else and not us more.
I don't want to comment on that.
Okay. My second question is that it looks like this CapEx step-up of this year and next year, and when you think about that, the 28 -nm will also solve very strong demand with computing build capacity. A couple years later, it turns into the overutilization, overcapacity. Although there's some factor that's because of the competition and lag is catching up. How do you see this risk going forward for 7-nm and 5-nm, and have you considered that whether it will be strong enough, second wave or third wave of the dimension backfill capacity when your current customers migrate into 3-nm or 2-nm or whatever it is, three years from now?
To answer your second question first, we do have confidence that a lot of products will fill up the TSMC's capacity. Now, a few years later, if you compare with the 28 -nm to 7-nm to 5-nm, it's not a good comparison because of 7-nm and 5-nm with the EUV, that technology barrier is much higher than you can expect from the High-k Metal Gate. Anyway, I don't want to comment on my competitor.
Right. Just last one. I promise, last one. When we look at the CapEx setup this time, it looks like the company has to stay around like TWD 10 or TWD 10 billion of CapEx, ±TWD 1 billion-TWD 2 billion for a few years right now. If we go back and look at in 2010, when we work out of the downturn in 2009, also there's a huge step up of your CapEx from 2009 to 2010, then we know what happened to TSMC, it's a golden era for TSMC for a couple of years. Now we have another step up out of the semi downturn in the past 12 months. Do you think it's similarity now for you today and when you make this bold CapEx decision this time?
I know what you ask, let me say that TSMC is getting smarter. Definitely. By the way, the technology barrier is much higher than, as I said, much higher than 28 -nm. We have a confidence that the capacity we build is a result, we closely work with our customer. We decide to increase the capacity at this time with a lot of detailed analysis. I certainly, I have a confidence that we won't repeat the same kind of error that we did.
I wouldn't call it error. I would say that that was a very bright and bold decision-
Oh, thank you.
...in 2010.
Oh, in 2010. Yes.
Yeah.
I'm sorry. Yeah. Those are the 28 -nm, the golden years. Very good. Yes.
I was asking whether we could see another few golden years after-
Oh, you say that for Of course. What you expect?
Well, before we conclude our conference, ladies and gentlemen, I would like to announce that important news that the beautiful lady sit right beside me, Elizabeth, has decided to retire from the company at the end of this year. As you all know, Elizabeth has been with the TSMC as a head of investor relations for 17 years. 17, that's a long time, isn't it? She's also the head of our public relations for 10 years. Well, the beautiful lady has been called in the public as the face of TSMC. TSMC is beautiful also. In her IR role, Elizabeth has won numerous awards and recognitions all over the world as the best IR officer, that in Taiwan, in Asia, and in the world. Many of you have been in frequent contact with her, I presume.
You know that she is a brilliant and enthusiastic and energetic lady, and you can definitely sense how much she loves TSMC. In last many years, she has built a world-class IR team and developed successfully a competent successor who is sitting right there. On behalf of TSMC, I would like to thank her for all her dedication and contribution to the company's success and wish her the best for her retirement. Now, ladies and gentlemen, let's ask Elizabeth to give us a few words.
Thank you, and thank you very much, C.C. I am very privileged to have been able to work for TSMC in the last almost 17 years. Under the leadership of our founder, Dr. Morris Chang, and the current management team, I have witnessed how this company is able to move from strength to strength and still remains true to its mission and its value. I have been blessed with the opportunity to represent this company that I deeply admire in front of the investment community and the press. I have enjoyed every minute of it, and I hope I have served the company well.
You did.
With all my heart, to the investors, to the analysts, and to the press, I want to thank you for your friendship and trust. I very much enjoyed our interactions, communications, and discussions in the past, and I do hope that you will extend your goodwill to my IR successor, Jeff Su, who has been working closely with me in the last four years, and many of you are already familiar with Jeff. I have had the most wonderful, exciting, and rewarding time in my career at TSMC. I'm leaving the company at a time when its future is brighter than ever, and I'm confident that you will continue to derive handsome returns from owning our shares. With that, let me wish you good fortune and good health, and we will conclude today's conference here.
Thank you.