[Non-English content ] Welcome to TSMC second quarter 2019 earnings conference and conference call. This is Elizabeth Sun, TSMC 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.
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 Senior Vice President and CFO, Ms. Lora Ho, will summarize our operations in the second quarter 2019, followed by our guidance for the third quarter. Afterwards, Ms. Ho and TSMC CEO, Dr. C.C. Wei, will jointly provide the company's key messages. Then TSMC's Chairman, Dr. Mark Liu, will host a Q&A session where all our executives on stage, including TSMC's Deputy CFO, Mr. Wendell Huang, will entertain your questions. For those participants on the call, if you do not yet have a copy of today's 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 conference presentation.
Usually, I would like to remind everyone that today's discussions may contain forward-looking statements, which are subject to significant risks and uncertainties, which could cause 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, Ms. Lora Ho, for the summary of operations and current quarter guidance.
Thank you, Elizabeth. Good afternoon, everyone. Thank you for joining us this afternoon. My presentation will start with financial highlights for the second quarter followed by the guidance of the third quarter. Our second quarter revenue increased 10.2% quarter-over-quarter as we already passed the bottom of the cycle of our business and began to see demand increases. Gross margin increased by 1.7 percentage points sequentially to 43%, mainly due to the absence of photoresist defects material incident and a slightly more favorable foreign exchange rate. Total operating expenses represented 11.2% of net revenue, lower than 11.9% in the first quarter. Operating margin increased by 2.3 percentage points sequentially to 31.7%. Overall, our second quarter EPS was TWD 2.57, and ROE was 16.2%. Let's take a look at revenue by technology.
7 nm process technology accounted for 21% of wafer revenue in the second quarter, 10 nm was 3%, and 16 nm was 23%. Advanced technologies, which are defined as 16 nm and below, accounted for 47% of wafer revenue, up from 42% in the first quarter. Let's talk about the revenue contribution by platform. Smartphone increased 5% quarter-over-quarter to account for 45% of our second quarter revenue. HPC increased 23% to account for 32%. IoT increased 15% to account for 8%. Automotive increased 3% to account for 5%. Digital consumer electronics and others went down slightly, accounting for 8% to 4% each of our wafer revenue. Moving on to the balance sheet. We ended the second quarter with cash and marketable securities of TWD 765 billion, an increase of TWD 4 billion from the last quarter.
On the liability side, current liabilities increased by TWD 244 billion, as we accrued about TWD 259 billion for 2018 and for first quarter 2019. 2018 cash dividends of TWD 8 per share will be paid today, and the first quarter 2019 dividend of TWD 2 per share will be paid out in October. On financial ratios, accounts receivable turnover days decreased seven days to 42 days. Days of inventory decreased three days to 76 days, primarily due to lower days in working process inventories and finished goods, both of which resulted from higher shipments out of inventories built in first quarter 2019. Let me make a few comments on cash flow and CapEx. During the second quarter, we generated about TWD 118 billion cash from operations, and spent TWD 116 billion in capital expenditures.
As a result, we generated free cash flow of TWD 1.4 billion, and our overall cash balance increased TWD 4 billion to TWD 650 billion at the end of the quarter. In US dollar terms, our second quarter capital expenditure was $3.75 billion. I have finished my financial summary. Let's turn to third quarter guidance. Based on the current business outlook, we expect third quarter revenue to be between $9.1 billion and $9.2 billion, which is an 18% sequential increase at the midpoint. Based on exchange rate assumption of $1 to TWD 31.0, gross margin is expected to be between 46% and 48%. Operating margin is expected to be between 35% and 37%. This concludes my financial presentation. Let me follow by making a few comments about the profitability, CapEx, and the cash dividend. First, about the profitability.
Let me make some comments on our second quarter and third quarter, and the overall profitability outlook. Our second quarter 2019 gross margin improved by 1.7 percentage points sequentially, mainly due to the absence of the photoresist material incident from first quarter, and a slightly more favorable foreign exchange rate. The reason second quarter revenue is slightly above the high end of our guidance, but gross margin is at the low end, is because the pace of cost improvement at N7 did not meet our plan in second quarter. We expect cost to gradually improve towards the plan starting from the third quarter. We have just guided third quarter 2019 gross margin to improve by 4 percentage points sequentially at the midpoint, mainly as we expect a higher level of overall capacity utilization. Our gross margin in first half 2019 was primarily impacted by a lower capacity utilization rate.
As our business and utilization rate improves in the second half of this year, we believe above 50% is still a good target for our gross margin going forward. Regarding 2019 CapEx planning, at the beginning of this year, we have guided our 2019 CapEx budget to be between $10 billion and $11 billion. Over the last three months, we have seen an acceleration in the worldwide 5G development. We believe this will lead to an increase in demand for our 5 nm and 7 nm technologies beyond the level we forecasted three months ago. We are therefore working closely with our customers for the most effective capacity planning for our N5 and N7. We expect our 2019 CapEx is likely to exceed the high end of our guidance range.
We are currently evaluating our 2019 CapEx plans and expect to provide you a more detailed update during our October earnings conference. My last comments is about the cash dividend distribution. We have communicate our dividend policy earlier this year. We will have sustainable cash dividend per share on both an annual and a quarterly basis. In addition, as our free cash flow increase, we will distribute about 70% of our free cash flow as cash dividend. TSMC's AGM in June approved the board's approval of TWD 8 cash dividend per share for full year 2018, and the revision of the Article of Incorporation to adopt quarterly dividends. The board then approved TWD 2 per share dividend for the first quarter 2019, which will be distributed in October 2019. Therefore, TSMC shareholder will receive a total of TWD 10 cash dividend per share this year.
That also means shareholder will receive at least TWD 10 per share cash dividend for 2020. Going forward, TSMC has set the payment months for the quarterly dividend as January, April, July, and October of each year. This concludes my remark. I would like to turn to C.C. Wei for his comment.
Thank you, Lora. Good afternoon, ladies and gentlemen. Let me start with near-term demand and inventory. We conclude our second quarter with revenue of TWD 241 billion, or $7.75 billion, slightly above our guidance due to a higher demand from HPC and IoT applications as compared to the time when we gave the guidance. Although our business continue to be impacted by the softer overall global economic condition, customer inventory management, and high-end mobile product seasonality, we have also passed the bottom of the cycle of our business and began to see demand increases. Moving into third quarter this year, TSMC's business will be driven by new product launches of premium smartphones, the acceleration of 5G development, and the increasing adoption of our industry-leading 7 nm node by high-performance computing applications. Let's talk about inventory. Our fabless customers' overall inventory is being gradually digested throughout second quarter.
We expect it to reduce to several days above seasonal level exiting the second quarter, leading to an improved inventory environment for the second half of this year. Although a soft global economic condition and trade uncertainties remain, we expect our business to be much stronger in the second half as compared with the first half of this year due to the strong demand for our industry-leading 7 nm technology solutions. The progress of our advanced technologies were on track, and we are very confident in our technology leadership. Over the last three months, we have seen an acceleration in the worldwide 5G development. This will speed up the introduction and deployment of 5G network and smartphone in several major market around the world.
We expect this to lead to an increase in demand for our 5 nm and 7 nm technologies. We are working closely with our customers to carefully plan our capacity to meet their demand, as our CFO just said. I will talk about our N5 status and N3 development. Our N5 technology has already entered risk production in first quarter. Customer tape-out activity are underway, and volume production is scheduled in first half of year 2020. With 80% logic density gain, 80%, and 15% speed compared with the 7 nm , we believe our N5 technology is the most advanced in the foundry industry, with the best density, performance, power, and transistor technology. Our 5 nm technology solution will be the foundry industry's most advanced solution until our 3 nm arrives.
We are confident that a 5 nm will have a strong ramp and be a large and long-lasting node for TSMC. On N3, the technology development progress is going well. We are already engaging with the early customers under technology definition. We expect our 3 nm technology to further extend our leadership position way into the future. I will talk about the ramp-up of N7, N 7+ , and the progress of N6. We are seeing very strong demand at N7 across a wide spectrum of products for mobile, HPC, and IoT applications. Meanwhile, our N7+, which adopts EUV for a few critical layers, has already entered volume production. We expect our customers' end products using N7+ will be in the market in high volume this quarter. We expect strong demand to continue into next year.
N6 provide a clear migration path for second- wave N7 products, as its design rule are 100% compatible with N7, while providing 18% logic density gain and performance- to- cost advantage. N6 will use more EUV layer than N7+. N6 risk production is scheduled to begin in first quarter year 2020, with customer product tape-outs in second half 2020. Volume production start before the end of year 2020. We affirm N7 and N7+ will contribute more than 25% of our wafer revenue in this year. We expect even higher percentage in next year from N7, N7+, and N6, because development of 5G accelerates, and demand from HPC, mobile, and other application continue to grow. Let me talk about TSMC's competitiveness. The foundry business model has proven to be the most efficient model in the semiconductor industry.
As a pure-play dedicated foundry, we collaborate and work closely together with our customers to unleash their innovations to the market and enable their success. We do not have any internal products, and we do not compete with customers. Within foundry, TSMC competes on technology leadership, manufacturing excellence, and customers' trust. Our trinity of strengths enables us to be everyone's foundry. We have the most useful and robust technology offering across both advanced and specialty technologies. We work diligently to protect our customers' technology, extend our leadership, and accelerate our technology differentiations. We are the world largest and trusted provider of logic capacity with an excellent manufacturing track record. We will continue to unleash innovations for all our customers for years to come. Finally, I'll talk about our CFO transition.
After serving very well as TSMC's CFO for the past 16 years, Lora Ho will take on a new challenge as Head of Europe and Asia Sales. Subject to the Board of Directors' approval, Wendell Huang will become TSMC's new CFO effective September 1st. Wendell has been with TSMC for 20 years and have served as TSMC's Deputy CFO and Head of Finance division. He brings a wealth of experience and knowledge of TSMC, and I am confident he will continue the strong tradition of TSMC's finance organization. I'm excited about both appointments and look forward to continue to work closely with both Lora and Wendell in their new roles. 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, so that all the participants have 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 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. We'll start the question first from Credit Suisse, Randy Abrams.
Yes, thank you. Maybe since it's Lora's last time, and we also have two CFOs, I'll start with a financial question. On the gross margin, if you could elaborate just a bit more on what were the triggers for the slower progress on gross margin, as it seems uncharacteristic for TSMC, at least recently, to have a bit of a impact on process ramp-up. Looking forward on the 50%, if you still expect that target toward the end of this year, and if looking into next year as you ramp up N5 and normally have a 2 point-3 point impact, if you still expect 50% reasonable for next year?
When we ramp up any new technologies, we have a series of productivity cost improvement activities on the plan. 7 nm is the same. 7 nm now is very busy. It's a lot of new tape-out. In second quarter, we have set a goal. We did not achieve the goal. That is why, as I explained earlier, the margin fall into the low end of the guidance, okay. As I said earlier, we have back to the track to gradually improve productivity. With the volume continue to come in third and fourth quarter, we are still confident we can achieve our cost reduction and productivity target. As to the 50% gross margin, I think for the whole year, I think the main reason is the lower utilization, particularly the first half of the year.
As I said, if we can maintain the high utilizations, which we believe we can, the 50% is still a good target going forward.
Okay. That implies also fourth quarter of this year?
That is certainly my hope.
Okay, great. Okay. Second question, I actually wanted to ask a bit more on the sales growth outlook you put out. Maybe the first starting point is you originally guided the full year would be growing back in April. I was wondering if you could update that expectation. To dig into the growth, if you could talk a bit about, in that guidance for high teens growth for the different segments, the HPC, auto, IoT, smartphones, how you're seeing each of those segments for third quarter and second half, and if you're also seeing cryptocurrency come back.
Well, for the first part, I think C.C. can answer. Yeah.
Okay. What is the first part?
Sorry, wrong question. First part was if you still think you can grow f or t his year.
Oh, for the whole year?
For the whole year.
Let me give you some kind of a taste. Right now, the uncertainties have really become, in these days, it's really hard to say. I cannot give you a kind of a, again, a very firm number of what kind of a growth we are going to get. I can assure you that the fourth quarter will be better than the third quarter. That's all I can say. Okay. How much? I don't know yet. You can calculate.
All right.
The key is fourth quarter we think will be better than third quarter. I think the second part, maybe Lora can answer.
Randy, did you ask about the by platform growth for third quarter?
Yeah. Actually, if you could give.
Okay.
The third quarter and then an update on the full year.
Okay.
For the platforms.
Okay. Third quarter, as I just guided, there is a sequential 18% growth. We are seeing very strong growth on smartphone, also actually, all platform are going to grow in third quarter across the board, with smartphone growing the most. Also HPC is growing very nicely. IoT is very strong, although it's very small as a basis. We also see automotive coming up to grow. It was kind of low in the first half of the year. Okay.
In terms of the whole year, we expect smartphone will grow single-digit year-over-year. HPC, if exclude cryptocurrency, we will also see single-digit growth. With the cryptocurrency, it's a decline number. Okay. IoT will grow more than double-digit, very nicely. Automotive will be a down platform for the year. Okay. That is the segment analysis.
Just the final clarification. For cryptocurrency, if you're seeing. Like how much of that activity is coming back, and do you expect to target that in any of the new capacity on 7 nm tied to that?
Why don't C.C., will you answer that? Yeah.
The cryptocurrency, recently, the pricing is up, and so we start to see the demand improving. We support the cryptocurrencies by available capacities.
Next question will be also coming from the floor from Morgan Stanley's Charlie Chan.
Thanks. My first question is about your full year outlook. I think there is still uncertainty about the U.S.-China tension. Huawei is still in the Entity List, right? When you plan the full year or even fourth quarter, do you discount that risk? How much do you discount that risk from Huawei? I think there's a different intuition, right? Meaning the biggest impact to the industry could be the 5G infrastructure from the U.S.-China tension. You are seeing that infrastructure team is accelerating, right? Can you give us some sense that what you see differently about the infrastructure markets?
All right. The effect of Huawei being in the Entity List and impact to TSMC, we do see some impact, but not direct business between Huawei and TSMC. Actually, HiSilicon and TSMC. Because we already announced that we continue our shipping practice, because we follow the law, so we continue to ship. The infrastructure, actually, the 5G development actually accelerates, and we see a very strong demand from that. As I said in my statement, in the many countries, that they speed up the 5G deployment, and we see the increase of the demand on our leading-edge 7 nm.
Yeah. The thing is that, for example, being in the Entity List, Huawei still has some restriction to update, for example, Google's mobile service. That has impacted their overseas smartphone demand. Right. According to your breakdown, the smartphone business still can see single digit year-over-year growth. I'm wondering, have you discounted the potential downside risk if that Entity List issue is going to last longer?
Actually, I don't want to specifically pinpoint one customer only. All I can say is, the second half, the new smartphone launches, especially the premium grade has been the seasonality phenomena for us. Always the second half. Acceleration of the 5G a ctually enhance this kind of increase. That's all I can say.
Let me add a little bit, yeah. Of course, the second half, there are still a lot of uncertainties, right? The geopolitical trade policy changes still go with the time. I think the second half of the biggest growth momentum come from the new smartphone launches. On Huawei's factors, yes, we think we did the discount.
Okay.
Since Huawei's ban, many things happened, too. The downside is generally the smartphone market becoming uncertain, in addition to Huawei themselves. The whole smartphone market is suppressed, I think. Secondly is the trade barrier. Trade uncertainties still prevail all sectors. From all sectors, particularly the industrial and the consumer, I think we still see the momentum still not coming back. Of course, the 5G momentum picking up, that is also new for the world, globally, in U.S., in China, in Korea, in Japan and so forth. Those all factors combined, we think we try to make the best judgment, and I think that's the conclusion C.C . Just delivered.
Thanks. That's very helpful. My next question is to Lora. It's going to be a tough question before you transfer to another role. Follow up Randy's question, right? Compare your third quarter revenue scale versus the revenue scale in the first quarter 2018. That quarter, you made a 50% gross margin, right? That means you have a higher scale, but the gross margin, I think, is 3 percentage point or 4 percentage points than that quarter, right? I'm asking you whether there's any structural issue. Also, more specifically, your 7 nm is in the second year, right? Supposedly, it shouldn't cause any margin dilution anymore, especially you are saying that 7 nm fab is quite busy, right?
Can you explain why revenue scale is higher, fab is busy, second year of the new node, but the gross margin is below a few quarters ago? Thanks.
Actually, utilization is still the manufacturer, if you compare on a year-over-year basis. Of course, there are still some dilution for N7, but this dilution will start to diminish it as we have a much strong demand in the second half. There's a little bit of a product mix issue, because we have portfolio product mix. We have talked about this. The certain technology, capacity here is low. For example, 28 nm, that has some impact on our overall corporate margin. Those are the things I can think of that are related to the margin changes.
We will go to the line now. Operator, please have the first caller on the line.
The first question comes from the line of Gokul Hariharan from JP Morgan. Please ask your question.
Thanks for taking my question, and apologies for my voice. First question. There's been a lot of discussion recently about core competition from one of your foundry competitors. Could TSMC talk a little bit about a roadmap on 5 nm and especially on 3 nm , where one of your competitors is looking to introduce a new transistor structure like Gate-All-Around. Is TSMC going to stick with FinFET at 3 nm as well, or TSMC is also likely to move to a nanosheet or a Gate-All-Around structure? That's my first question. Could we talk a little bit about what is the activity level TSMC is seeing at 5 nm capacity? I think previously you mentioned that 5 nm is seeing. You're building l esser 5 nm capacity for next year.
Could we also talk a little bit about how are customers thinking in terms of choosing for 6 nm versus 5 nm , especially the customers who are on 7 nm today? Thank you.
All right. Gokul, since, I guess because you have a flu, so we can't really hear you quite well. Let me try to see if I understand your questions. First, your question is with regard to the competition within foundry. You ask whether we can compare our technology roadmap of 5 nm and 3 nm versus that of the competitors. Your second question is with respect to 5 nm, where you asked that we seem to be a little bit more conservative about 5 nm capacity build plan some time ago, now we are a bit more aggressive. You are asking us if we have seen any differences in customers' demand for 5 nm.
C.C., would you answer the first question?
Yeah. Yeah.
That's right. Thank you.
The first question?
First question is 5 nm and 3 nm . We think our 5 nm is very competitive, and the first one in the industry in that geometry.
What be the risk production right now and the volume production, the first half in the next year. This continue to be the same situation. About the 3 nm , let me clarify a little bit. We have evaluated all the possible options, and come with a very good solution for our customer. We continue to work with our customer to define the spec, to define the approaches, and to meet their requirement. I will update you that our choices next time.
Let me add that. Actually, our five nanometer is a full node stride from our 7 nm . Our 3 nm is another full node stride from our 5 nm . This is very different than our competitors' road map. If you compare their numerical, 3 nm it's probably closer to the 5 nm. Secondly, on the 5 nm capacity build. The second question is?
Yeah.
Yeah, five nanometer capacity.
What have you seen customers mention?
Okay, C.C. then.
Okay. The 5 nm , in the last quarter, when we communicated with you, we say that we are going to be a little bit conservative, and to work with customer to plan the capacity that necessary to support them. In these three months, because of a speedup of the 5G's development, we are working with the customer again. Probably we change our conservative attitude to become a little bit more aggressive. To meet our customers' demand, right now this is a new development, all because of 5G AI's progress.
Okay, thank you.
All right, let's come back to the floor. Next question will be coming from Goldman Sachs, Bruce Lu.
Good afternoon. The last couple months has been pretty exciting in terms of the macro environment. I think there are some changes in terms of customer behavior. Typically, we somehow synchronize their production together with their sales. If they have some problem in terms of their sales, they somehow change their production plan. There may be some time lag. Last couple months, we start to see that customer is willing to piling up more inventory. There are some disruption in terms of their sales plan. In terms of production plan, you don't really see a lot of meaningful changes, at least for TSMC. How do we forecast this kind of thing moving forward? It seems to me that the macro environment is so dynamic, as analysts, we have a lot of difficulties, right? We can't just follow the tweet, right?
Can you help us how to predict that kind of customer behavior in terms of the production planning moving forward?
That's an interesting question. You ask us how we estimate or forecast a customer's inventory or how they pile up their component. Let me tell you that TSMC only receive the PO we put into production. Okay? We did not see that strange customer's behavior say that they are piling up of the product in expectation of something happening. No, we did not see that. That's in our daily life. Again, that means to stress one point, we receive the PO, we do the production, we did not see a very strange kind of a phenomenon.
Don't follow the tweet. As the fabless inventory, I think Lora just reported, is coming down.
Yeah, one of the important customer, they don't provide the public information.
That's right.
That become the biggest swing factors, right?
Most system companies, we have the inside look. The company you just mentioned, yes, indeed, we do not know their inventories. From their orders stream, we don't see any abnormal flow.
The second thing is management keep on mentioning about 5G's acceleration. Can you give us some granularity about what kind of revenue contribution coming from 5G in the second half of 2019 or 2020, either from smartphone side or from infrastructure side? What is the revenue exposure to TSMC at this moment?
Lora, do we have a 5G specific pointed out?
We don't. It's difficult to differentiate how much demand is coming from 5G. We track the demand on node basis, also on platform basis. 5G may be related to HPC, some of the smartphone as well. We do see those two segment are growing very strongly.
Yeah, let me add some color to it. Actually, the networking processor, FPGA product, even some CPUs, also base stations, smartphone, that all are in 5G. We see a very strong demand in the second half of this year.
I think the question you're asking about exposure of 5G.
What can you elaborate what that means?
Well, basically is the big study says the strongest demand as management just mentioned. Basically, we try to get some granularity or we try to quantify how much growth we can expect moving forward. As I said, it can be any kind of product, but the bigger revenue contribution still comes from the base station and the smartphone side. That's the key area we are trying to focus on.
Yes. The smartphone adopt implementation of 5G on smartphone, we see actually stronger than when we were at 4G's ramp. Okay, that's the information you get. That is the opportunity for us as we consider it, yeah.
Lastly, can I squeeze in, can we still maintain that five years revenue CAGR's guidance?
Sure.
I love that.
Okay, that's the confirmation. Next question will be coming from Citigroup's Roland Shu.
Hi, good afternoon. With the 7 nm reach 25% of total revenue this year, 7 nm will be the biggest node ever in revenue to TSMC this year. Next year, C.C., you expect 7 nm will be growing even more, more than 25% of total revenue next year. It means that your 7 nm will be even bigger. C.C., you also said 5 nm going forward will be a big and long-lasting node and also will be growing bigger than 7 nm. Are you still maintain this view with that this 7 nm is going to be very big next year?
First, let me say again, it will be very big next year.
All right. Why we are so upbeat on the 5 nm, because we forecast the ramp-up will be faster than, in terms of our revenue, faster than 7 nm. We expect that our 5 nm solution to all the customer is very competitive. The 5G AI, again, that will be benefited the 5G, the 5 nm, of course, and the 7 nm.
Yeah. For next year, first you are going to ramp 5 nm very fastly, you still have a very big 7 nm. For next year, I think that, what is the overall growth outlook for you for next year?
Let's wait for next year, I give you the answer. All your statement are true.
Okay. Thank you. Second question actually is also related to this 7 nm and the 5 nm . Now your seven plus is entering a mass production. How is the EUV availability and the productivity for your 7+ so far? Has it reached the mass production need now?
The EUV, we don't see any problem in production. It's going very well, on schedule, and we are very happy about it.
How does this EUV productivity or reliability compare to immersion in the same stage?
Same stage.
Same stage. That's a long time ago.
So far it's on our schedule, everything according to the plan, that's all I can say.
Okay. Just an outlook view. For next year with your very fast 7 nm, and the 5 nm ramp, is this EUV going to be a big dilution to the gross margin next year?
Oh, that's a tough question. All I can say is that EUV will perform as we scheduled. All right. It will be very important in our cost reduction path.
Okay. Thank you.
All right. Let's go back to the line. Operator, please have the next caller on the line.
Next question comes from Mehdi Hosseini from SIG. Please ask your question. Your question.
Yes. Thanks for taking my question. A couple of follow-ups. On the CapEx item, how should I think about a higher than expected CapEx in 2019? Is that more of a pull-in from 2020 or your capital intensity is going to remain at a higher level looking forward?
We are still working on the CapEx number. We're going to report to you next quarter. In terms of direction, we are seeing more CapEx requirement both for 7 nm , also for 5 nm . For 5 nm , we need to pull in the tool to meet customer's request, okay. In terms of capital intensity, you will probably see in the short-term period, you will go higher than what I have guided at a 30% level. In the longer term, we still believe the 30% level is the right level of CapEx intensity.
Thank you. Just a quick follow-up on your view on Q4. I get a sense that there are a still uncertain environment. What is the one end market that is the most valuable when you think about the trend into Q4?
Mehdi, you are asking what is the one end market that is most vulnerable or valuable? I did-
Is that makes it more challenging to-?
Most challenging. Okay.
To forecast the Q4. Yeah.
Well, I think the most recent Japan and Korea dispute probably is the most uncertain one for the fourth quarter.
Sure. How should I think about its impact on a specific end market? Is that more of a broad-based, unknown fact? Is the unknown more of a broad-based macroeconomic trends, or is it specific to one particular end market?
It is so uncertain that we cannot pin down specific product that will be impacted. People talk about smartphone itself could be impacted. It's breaking down many supply chains. It is including display or other electronic components. That is the difficult part for us to make a estimation.
Okay, great. May I ask you one more question?
Okay. All right.
Sure. I'll make it quick. I think there's no doubt that 5G opportunities are enormous. I just want to better understand how TSMC is planning. Carriers refer to a commercial end market or commercial aspect of 5G as more meaningful than consumer on handset. On the other hand, your customers, semiconductor companies, highlight the opportunities in the smartphone. As you plan for your capacity, leading edge capacity plans, it's very long lead time, how do you think about 5G and its impact on consumer, which is more of a handset, a smartphone, versus commercial aspect, which could have an impact on your HPC or other segment?
Well, actually, let me answer the question because the 5G, what we expect is not only on the consumer product like smartphone per se. I do think the 5G will affect every area in our platform. For example, without 5G, the autonomous driving will not be possible. It's very important for automotive also. It's very important for the high-performance computing also because that's one of that networking processor, everything connected. It also important for the IoT because of that where you collect all the data, and with the 5G, it's a multi-channel, with a very small latency, with very high speed, and all the data collection need to be analyzed. I think in terms of 5G's effect on the business and also on TSMC's capacity plan is enormous.
What TSMC doing is right now working with all the customers to plan for their business and plan our own capacity. That's all I can say.
Let me add some color for this.
Thank you. Yes.
5G, I think, if you look at next year, the biggest business influencer should be in the smartphone. Secondly, will be the high-performance computing, which compose of the networking and other infrastructures. As far as automotive and the IoT c onsumer, we have a high expectation for that. I think it will take some time before the usage model get implemented in the market. That will come to us later.
Your next question comes from the line of Bill Lu from UBS. Please ask your question.
Hi. Thanks for taking my question. First of all, I just want to say thank you to Lora for all the help over the years and patience in answering all of the questions. Thanks a lot and best of luck. My first question is again on 7 nm . If I go back and look at TSMC for 28 nm, that was TSMC's most successful node of all time, and that coincided with the big smartphone ramp. If I now look at 7 nm, you've got the move from 4G to 5G, but you're also adding HPC. If I look at capacity for 7 nm, is it reasonable to assume that it could be bigger than 28 nm?
Mm-hmm. Yep.
7 nm .
Yep.
7 nm at this same stage, is this bigger than 28?
Are we?
Yeah.
Oh, okay. My numbers say yes. I don't think so.
I think revenue-wise, 7 nm was definitely much bigger than 28 nm, not necessarily the capacity.
Necessary. Yeah, not capacity.
Okay. Can you give me a sense for, ultimately, how big 7 nm capacity could be relative to 28 nm?
It's smaller, but it's close. At the same stage.
Oh. Okay, great. Maybe I could ask that a different way, which is, if you look at 7 nm demand, let's say a year from now, what do you think is the split between HPC, smartphones, and maybe the other platforms?
Bill is asking, a year from today, if we look back and look at the 7 nm revenue, what will be the split between HPC, smartphone, and other platforms?
We don't have the specific number. If you look at today's HPC and smartphone percentage ratio, I think it will be probably similar, although a fraction of the smartphone may not be the leading edge. For those two sectors, both sectors are closer to the leading edge. Both of those segment will go to 7 nm . Probably smartphone will still be bigger than HPC. It's a similar ratio, I think, as the business we have today.
Great. My second question is on 5 nm. A quarter ago, I asked Dr. Wei about the cost per transistor at 5 nm. I heard some of the feedback from customers is that maybe it is not coming down as fast as expected. I think Dr. Wei's answer was that TSMC is working on it. I am just wondering if you can give me an update on that, whether costs have seen improvements or what the status is.
Excuse me, you are asking the device improvement or is it demand improvement?
No, sorry. I think if you look at cost per transistor for 5 nm, it is not coming down as fast as previous nodes. I am wondering if you have an update there.
No, I do not have any update, except that right now we see a much more stronger demand.
Okay, great. Thank you very much.
Great. Thank you, Bill. Let's come back to the floor. Next question will be coming from Daiwa's Rick Hsu.
Yeah, hi. Good afternoon. Just one question from me. I think Lora just talking about this year's revenue growth by different platform. Like smartphone up a single digit, high performance single digit up, and IoT up double digit. If I recall my memory, this set of guidance looks the same as that you provided earlier this year. Can I fairly assume by your total revenue, are you guys still keep the same guidance of the revenue for the whole year will be up slightly on you?
There is still uncertainty on fourth quarter. What I can say now is fourth quarter revenue will be higher than third quarter.
Okay, fair enough. Just one quick follow-up. Can you also update this year's global semi outlook and also GlobalFoundries?
Next, you mean 2020?
This year.
This year.
This year's foundry actually is a little bit negative.
Semi exclude memory.
Yeah.
-3%. Foundry, - 1% this year.
Okay. Thank you so much.
We have implicitly answered your question.
Thanks.
All right. Next question will be coming from Credit Suisse, Randy.
I just want to follow up. There's been a lot of press, and maybe it's noise about customers evaluating other foundries. You are talking about strong 5G HPC. I'm just curious if you're factoring in any offsets. Like in the past, you've had a couple of major customers go back and forth between foundries. As you look at the next two years, do you see that much? How do you see your market share, and do you see any offsets to that strength over the next couple of years?
Well, we do see the strong competition. Let me assure you that our technologies are leadership, and also that our manufacturing is excellence. With our customers' trust, I believe we are going to maintain our market share and increase the market share. Did that answer your question?
Yes. Well, within that, because we know about some projects, like customers like AMD that have shut down or switched from GlobalFoundries. It's more if you also see any offsetting drags that we should factor in, just so we don't get too far ahead of ourselves. If there's anything like we saw a few years ago, we had a mobile customer shift foundries. If you see any offsets or from a market share it still looks like. I think in the past you've said you have pretty much all the customers, but if you still view that on the.
Yeah, we factor in what the possible competitor's sales approach. We factor it in.
Okay. If I could ask on the nodes, actually two of them. One on, as the 7 nm matures in the next couple of years, how you now see the variation of 7+ nm and 6 nm. Because I think in the past you saw that as becoming mainstream. If you still think the 7 nm evolves that way, where most of the customers move to the EUV version, 7+ nm or 6 nm. For 5 nm for next year, how concentrated do you see that both in terms of customers and then also if it's concentrated smartphone or you see a lot of HPC?
Let me answer that. A little bit ask about the 7+ nm or 6 nm, right? To continue the 7 nm node. Actually, we would believe that most of the customer, especially the second- wave customer, will adopt the 6 nm. Okay? Because you're 100% compatible with the 7 nm that reduce their burden and redesign all the IPs. The 6 nm will be a very sweet spot for them to continue with their 7 nm route. That's one. Okay. 7+ nm this year actually is, we provide a better performance and better density for some of the customers that have been adopted, and they will continue to grow. The majority, as I said, will be going to 6 nm. As for 5 nm will be adopted by a lot of platform that including the smartphone, the mobile, including the high-performance computing and, let me see what else. IoT? No, we did not see IoT yet.
At the beginning, it will be mobile and HPC.
Thank you.
Next question will be coming from CL Securities, Sebastian Hou.
Thank you. My first question is a little bit follow on Randy's questions. If we narrow down Randy's questions to very leading edge technology, the market share. Let's say sub 10 nm market share in the next three years. What's TSMC's expectation on that? Are we going to maintain or further increase market share here?
I would believe we will maintain. We will have a very high market share. Very high.
In terms of the very high, like 95 is high, 90 is high. From 95 to 90-
You are going to nail down the number?
I want to know the direction.
I think direction probably our 16 nm is higher than our 28 nm.
Right.
Our 7 nm will be higher than.
16 nm.
16 nm and on. Yeah.
All fine. What's your expectation on 5 nm versus 7 nm?
At the beginning, those were the two.
5 nm is too early to say, but we target to be higher than 7 nm.
Okay. The second question is, can TSMC talk about the possibility of building a fab or acquire a fab company in the U.S.? Would this be out of the pure geopolitical concerns or any other consideration?
Of course, geopolitical concern is everybody's concerns. However, if you want to answer how can we solve the concern, it's not that simple. It's not building a fab outside Taiwan or any other country can solve that problem. However, we are always open to build a fab overseas, provided we can provide the same cost structure to our customers and to our investors. Far in the U.S., we have talked to the industry and to see whether that is a viable approach, is a good approach to our customers given the cost differences and if given in addition to the local subsidies. Still, current supply from Taiwan is still the best solution for our customers. We are open to that, but we are not in a hurry to make a decision. Of course, we also see some unproductive facility overseas.
We don't want to increase the excess capacity for the industry. The possible approach of building a greenfield fab acquisition is probably better for the industry. That's our current consideration, but we don't have a definitive plan today that we are going to have that in the U.S., at this point.
Okay. Can I just conclude that TSMC preference is to if you're going to do this, your preferred option would be buy rather than build?
At this point, yes.
Okay. Thank you. Dr. Sun, can I add one more question?
Follow up later.
Okay. I'll come back later.
Yeah. Next will be coming from Citigroup's Roland Shu.
I would like to switch gear to 8 in . You load your 8 in at the utilization, I think probably below corporate average in first half. How do you see your 8 in demand or utilization in second half?
The second half will be better. Actually, it's because of smartphone seasonality. The 8 in utilization rate will be much higher than the first half.
How do you compare with your corporate average?
Let me give you some taste that some of the segment actually is fully loaded.
Okay. This is the demand across the board. It's not just for certain customers rush order.
Oh, it's a demand across the board because of I'm talking about the product segment.
Okay. Thank you. Second question is how about your Nanjing Fab? Can you give us a color, an update for your Nanjing Fab?
Nanjing Fab progress very well. The loading is very healthy, and we continue our plan. We are going to build a 20,000 wafers capacity over there.
Any plan to above this 20,000?
Right now, 20,000 is our plan.
Okay. Thank you.
Let's go back to the line. Operator, please have the caller on the line. The next one.
Next question comes from the line of Xiux i Zhu from Standard and Poor's. Please ask your question.
Hi, thank you for the presentation. Just have very two brief questions. The first one is about your year-over-year drop of your net profit, net income. Could you please give us a specific reason for the drop, the 17.6% drop in the net income? The second question is about your outlook on the mobile sales in year two. You said it will be the main driver of your revenue growth. I was wondering, would that be overall increase of the shipment of the mobile that you are thinking about, or is there any some other reasons such as the technology upgrade over there? That's all, thanks.
First question, of course, is to explain what's the reason behind the year-over-year drop of our net income in this year, I guess, or second quarter. The second question is, the outlook for mobile, the growth that we indicated, was it because it was driven by a quantity increase or it is because of technology migration, leading to a revenue increase?
Let me answer the first questions. On year-over-year basis, if we look at the second quarter this year and the second quarter last year, the margin difference is mainly, there is about 5 percentage point margin difference, mainly from the lower utilization this year. First half of this year has been very weak, including the second quarter. Utilization is the main reason.
Okay.
A little bit of product mix as well.
What is the second question?
The mobile increase.
The second question is smartphone.
Yeah, smartphone.
Revenue increase.
Revenue increase, whether it's quantity or technology migration.
The second half's revenue increase due to the smartphone, firstly because of seasonality, of course. That as compared with the first half, second half, the smartphone is much better. Although we say that the whole year in the smartphone unit, we forecast a drop, TSMC's revenue still grow because of what? It's because of silicon content increase. Second one is because we gain the market share through our customer. We still forecast the whole year the smartphone will increase in revenue, especially the second half will be much stronger than the first half. Did I answer the question?
Yes. Great. Thanks.
Thank you.
All right. Operator, let's have this next caller on the line, please.
Next question is from Yu bao Hua from Tianfeng Securities. Please ask your question.
Hello, can you hear me?
Yes, we can.
Hello?
Yes, we can hear you.
Okay. Can I speak Chinese, please?
Yes, you can speak Chinese.
Okay. Thank you. [Non-English content]
Let me translate to English first. Yu bao had observed our June revenue, which is a strong growth, and he want us to comment on the second half business outlook.
[Non-English content]
Lora, can you speak in English too for the record?
I have just gave the guidance for the third quarter, where our revenue in third quarter will be $9.1 billion-$9.2 billion. That is 18% sequential growth. As to the fourth quarter, we have said we believe our fourth quarter revenue will be higher than third quarter, although there are still some uncertainties. We did not provide a clear guidance on the fourth quarter yet.
Okay. Thank you. [Non-English content]
[Non-English content] Yub ao's question is he likes to know the breakdown among our smartphones, HPC, and IoT, among our growth platforms, the revenue breakdown.
I will not provide a revenue breakdown by platform for the whole year, but I have just said that our smartphone will grow single digit. It still accounts for the biggest part of our revenue for the whole year, follow by an HPC, and then the rest of them are much smaller, like IoT, automotive, and DCE, and others.
Okay. Thank you.
Thank you. We still have a caller on the line. Operator, let's go to the next caller on the line. Thank you.
We have another question from Mehdi Hosseini from SIG. Please ask your question.
Yes. Thank you. Just one quick follow-up. Before, I also want to express my gratitude to Lora Ho and wish her the best of luck in her new endeavor. Going back to a question that came up last earning conference call, had to do with the SOI, I just want to revisit the topic and better understand how you're planning for some of the challenges that 5G brings, such as the lowest power consumption, and whether SOI is going to be included in your roadmap.
Well, we did develop some of the technology for the RF circuit, RF technology using the SOI wafers. We don't do the logic technology on FD-SOI. Let me make it sure that everybody understand what I'm going to deliver. We don't do FD-SOI for logic, for conventional logic technology. We do use SOI wafer to develop RF technologies. For the RF front- end, for example, that we are doing. Did I answer the question?
Got it. Yes. Thank you.
All right. Let's come back to the floor. Next question will be coming from Morgan Stanley's Charlie Chan.
Thanks. I have two follow-up questions, if I may, I may have some industry cross checks. First of all is about the smartphone silicon content increase. Can you give us some color, how much is that for this year? Also, when 5G comes next year, because Mark was just very bullish about the 5G smartphone contribution next year. How would 5G smartphone to help on your silicon content growth in a smartphone? Thanks.
That's a good question, but it's very hard to identify what is the percentage of the silicon content increase. Let me give you some color of it. All right. The first one is, look at your smartphone today. It's a three -camera, or a two- camera, or even four- camera. Look at that. Look at the pixel size, right? You might hear some of the pixel is a 48 megapixel. Those kind of thing come with silicon with it. That's the silicon content, the first one. Actually, there are some other minor things, for example, power management IC. Now, the power consumption is very important. Now every major component inside need a power management IC come with it. One is a screen driver. You need a power management. The application processor, you need a power management. Just a lot of things.
Furthermore, let me give you some of the 5G, you have a lot of different channel. Now even the RF transceiver or RF front end has the die size is bigger. That's why I say that's why silicon content is increasing. Did that answer your question?
Yes.
How many percentage, actually, I cannot identify it.
How significant is that for those modern chip or application processor? Because that relate to your business most. Can you give us some percentage of increase, like 10%, 20% increase in-?
Definitely no. I cannot because you know we are working on it.
Okay.
We know all the minor details, so I cannot give you that sum of the information.
Okay. That's fine. Next question, switch gear back to the gross margin. My question is that, for your leading-edge investment, do you feel like the payback period of those leading-edge investments are getting longer or shorter, right? Because I still want to find out some explanation why gross margin kind of decline year-over-year. Could be leading-edge investment is getting heavier or there is a true price competition from your competitors. Just want to get some thoughts from you on this topic.
Lora, can you give a idea on the payback years?
Okay. Actually, personally, I don't think a payback year means much.
If you look at the various technologies, actually, I check this question. They look very much the same. You know, the payback year are very much the same. I think the ramp profile is very different, and the sensitivity of utilization to margin is different. The more leading edge is more sensitive to margin. Leading edge needs to have very high utilization to secure the margin. That's the key, and we are working on it. Okay.
Thanks. Two things I want to cross-check, if I may. First of all, you mentioned that you pull in some 5 nm CapEx to this year, right? Next year, do you think the CapEx level would be at a range of your kind of annual guidance? Secondly, about the raw wafer price. Do you think you can get a good bargain on the raw wafer price for next year? Thanks.
I think next year's CapEx will have to depend on a lot of things. Depend on overall market conditions, the customers' requirements, so on and so forth. It's probably too early to say whether it's going back to this old range of target. As we get more clear, we will communicate with you. Yeah. The raw wafer price, I think we have done a good job to lock in the price. That's at least also a continued effort. Yeah.
I think, to make sure you understand, Lora just mentioned you will exceed the top range what we gave you today.
All right. Follow-up question will be coming from CL Securities, Sebastian Hou.
Thank you. I have two follow-ups. The first one is, for the past several earnings call, I always feel like TSMC 5 nm tape-out activity or interest level from clients isn't as large as 7 nm at the same stage, well, if I'm right. Now you mention stronger 5 nm demand. Is it more driven by the accelerated 5G deployment or also driven by the EUV cost, or EUV economy improvements?
It's actually driven by the 5G's accelerations.
Okay. In terms of the 5G acceleration, do you see it more from HPC or smartphone?
Both, actually, because of 5G infrastructure is in the HPC area.
Right.
On the smartphone, almost all the premium phone is with TSMC.
Okay. If I understand that correctly, I think so far, most of the infrastructure baseband, 5G baseband is still on 12 nm or 16 nm, the mainstream right now. You see the accelerated 5 nm demand. We will assume that the next uptick will be 7 nm, and now also see 5 nm.
Oh, I see. You are talking about that. Some of the networking process are probably going to the 5 nm, but some will stay in 7 nm, and some of the the baseband, actually, most of the baseband is going to the most leading-edge technology.
Of course, in terms of revenue, smartphone is bigger than the base station. Okay? In the 5 nm, the big players get into 5G smartphone very aggressively, and doesn't take many tape-outs, just a few tape-outs . De mand much bigger capacity than any other products.
Thank you. The second follow-up is, I'd like to hear TSMC's views on the next- generation architecture of transistor. How do you see the FinFET and Gate-All-Around? How to compare that to see the pros and cons of these two architecture in 5 nm and 3 nm? It seems like one of your competitor promote GAA aggressively.
Yeah, we notice that, actually we also have evaluated all the options at their side, just as what I said. We look at the pro and cons, we work with our customer, we choose the most competitive in terms of performance and cost-wise. We choose the most competitive approaches. We work with our customer, actually, closely.
Okay. Would it make sense to assume that so far since 16 nm to 10 nm to 7 nm, the industry standard has been FinFET. Customers will be easier to dual -source if they want to or if they can. Going forward, if there's going to be different route, one doing GAA, one doing FinFET, would it make customers dual- sourcing more difficult?
Actually, let me say that even the FinFET structure is very hard to switch the foundry. It's very hard because our design rule, the architecture, the design flow are all different. At the 5 nm geometry, TSMC still think the FinFET is the best one, although we have evaluated all other options. So far, we're still in 5 nm geometry. Please notice that 5 nm geometries, FinFET is still the most efficient one and most competitive one. For the full node transition to the next one, 3 nm, we are evaluating everything. We talk with our customer when we will define what are the approaches that we are using.
Okay. Thank you.
I think we can conclude our Q&A session now, knowing that we have evaluated all available options and picked the optimal one. Okay, before we end today's conference, please be advised that the replay of the conference will be accessible within four hours from now. Transcript will be available 24 hours from now, both of which will be available through our website at www.tsmc.com. Thank you for joining us today. We hope you will join us again next quarter. Goodbye and have a good day.