[Non-English content ] Happy New Year to everyone, and welcome to TSMC's fourth quarter 2019 earnings conference and conference call. This is Jeff Su, TSMC's Deputy Director of Investor Relations, 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 CFO, Mr. Wendell Huang, will summarize our operations in the fourth quarter of 2019 and the full year of 2019, followed by our guidance for the first quarter of 2020. Afterwards, Mr. Huang and TSMC's CEO, Dr. C.C. Wei, will jointly provide the company's key messages. Then TSMC's Chairman, Dr. Mark Liu, will host the Q&A session where all three executives will entertain your questions. 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 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 over to TSMC's CFO, Mr. Wendell Huang, for the summary of operations and the current quarter guidance.
Thank you, Jeff. Happy New Year, everyone. Thank you for joining us today. My presentation will start with financial highlights for the fourth quarter and a recap of full year 2019. After that, I will provide the guidance for the first quarter of 2020. Fourth quarter revenue increased 8.3% sequentially to TWD 317 billion, driven by high-end smartphones, initial 5G deployment, and HPC-related applications using TSMC's industry-leading 7 nm technology. Gross margin increased 2.6 percentage points sequentially to 50.2%, thanks to a higher level of capacity utilization and continuous cost improvement, partially offset by an unfavorable foreign exchange rate. Total operating expenses increased by TWD 3.6 billion, reflecting higher development activities for 5 nm and 3 nm, as well as opening expenses in preparation for 5 nm ramp. Operating margin increased by 2.4 percentage points sequentially to 39.2%. Overall, our fourth quarter EPS reached TWD 4.47, and ROE was 28.9%.
Now let's take a look at revenue by technology. Seven-nanometer process technology continued to ramp strongly and accounted for 35% of wafer revenue in the fourth quarter. Ten-nanometer was 1%, and 16 nm was 20%. Advanced technologies, defined as 16 nm and below, accounted for 56% of wafer revenue, up from 51% in the third quarter. On a full year basis, 7 nm contribution increased from 9% in 2018 to 27% of wafer revenue in 2019. Ten-nanometer was 3%, and 16 nm was 20%. Advanced technologies accounted for 50% of total wafer revenue, up from 41% in 2018. Now let's take a look at revenue contribution by platform. Our fourth quarter revenue growth was driven mainly by smartphone and HPC. Smartphone increased 16% quarter-over-quarter to account for 53% of our fourth quarter revenue. HPC increased 6% to account for 29%. IoT decreased 4% to account for 8%.
Automotive remained flat and accounted for 4%. On a full year basis, smartphone and IoT led the growth with 12% and 33% respectively, while HPC, automotive, and DCE decreased 8%, 7%, and 8% respectively. If we exclude cryptocurrency from both years, HPC would have grown mid-single- digit in 2019. Overall, smartphone accounted for 49% of our 2019 revenue, HPC 30%, and IoT 8%. Moving on to the balance sheet.
We ended the fourth quarter with cash and marketable securities of TWD 583 billion, flat versus the prior quarter. On the liability side, current liabilities increased by TWD 96 billion as we increased TWD 33 billion in short-term borrowing, mainly for hedging purpose, TWD 51 billion in payables to suppliers, and TWD 13 billion in dividends payable. On financial ratios, accounts receivable turnover days remained at 41 days. Days of inventory decreased 10 days to 55 days due to higher wafer shipments during the quarter.
Now let me make a few comments on cash flow and CapEx. During the fourth quarter, we generated about TWD 203 billion in cash from operations, spent TWD 170 billion in CapEx, and distributed TWD 52 billion for first quarter 2019 cash dividend. We also increased TWD 36 billion in short-term loans for hedging purpose. Overall, our cash balance slightly increased TWD 3 billion- TWD 455 billion at the end of the quarter. In U.S. dollar terms, our fourth quarter capital expenditures reached $5.6 billion and total $14.9 billion for the full year. Now let's take a look at the recap of our performance in 2019. 2019 was a challenging year for the global semiconductor industry, given rising macroeconomic uncertainties and supply chain inventory correction, to name a few. However, we're able to grow our revenue by 1.3% year-over-year in U.S. dollar term and 3.7% in NT dollar term.
Gross margin decreased 2.3 percentage points to 46%, primarily because of lower capacity utilization in the first half of the year. Operating margin decreased 2.4 percentage points to 34.8%. Overall, full-year EPS slightly declined 1.7% to TWD 13.32. On cash flow, we spent TWD 460 billion in CapEx, while we generated TWD 615 billion in operating cash flow and TWD 155 billion in free cash flow. We also paid TWD 259 billion in cash dividends, an increase of 25% from the previous year. I have finished my financial summary. Let's turn to first quarter guidance. Based on the current business outlook, we expect our first quarter revenue to be between $10.2 billion and $10.3 billion , which represents a 1.4% sequential decrease at the midpoint. Based on the exchange rate assumption of $1 to TWD 29.9, gross margin is expected to be between 48.5% and 50.5%.
Operating margin between 37.5% and 39.5%. Now I would like to make one more comment on tax rate. In the past, we needed to accrue tax on undistributed earnings, which triggered a much higher tax rate in the second quarter. Now, due to the tax regulation changes, we can offset the tax with our capital investments and no longer needs to incur the tax expense on undistributed earnings. Meanwhile, we're still subjected to the alternative minimum tax.
As a result, we will still have a full-year tax rate approximately 12%, and this will be equally applied to all four quarters of the year. This concludes my financial presentation. Let me follow by making a few comments about near-term demand and inventory and 2020 capital budget. We concluded our fourth quarter with revenue of TWD 317.2 billion, or $10.4 billion , slightly above our guidance, mainly due to better demand from smartphone-related applications than our forecast three months ago.
Concluding 2019, the semiconductor industry, excluding memory, declined 3%, while foundry was flat. TSMC's revenue grew 1.3% year-over-year in U.S. dollar terms, outpacing both the semiconductor ex memory and foundry industry growth. On the inventory front, our fabless customers' overall inventory continued to be digested throughout the fourth quarter. We now expect it to reduce to the seasonal level exiting 2019. Setting up a healthier inventory base entering 2020. Moving into first quarter 2020, despite mobile product seasonality, our business is expected to be better than the seasonality in recent years, supported by continued ramp of 5G smartphones. I will talk about our capital budget in 2019 and 2020.
We expect the ramp of 5G related and HPC applications to drive strong demand for our advanced technologies in the next several years. In order to meet this increased demand and support of customers' capacity needs, we raised our 2019 CapEx guidance by TWD 4 billion to TWD 14 billion-TWD 15 billion, and we ended up spending TWD 14.9 billion. Our 2020 capital budget is expected to be between $15 billion and $16 billion. Out of the $15 billion-$16 billion CapEx for 2020, about 80% of the capital budget will be allocated for advanced process technologies, including 3 nm , 5 nm, and 7 nm. About 10% will be spent for advanced packaging and mask making, and about 10% for specialty technologies. With this level of capital spending in 2020, we reiterate that TSMC remains committed to a sustainable cash dividends on both an annual and quarterly basis.
Now let me turn the microphone to C.C.
Thank you, Wendell. Good afternoon, ladies and gentlemen. Let me start with our 2020 full year outlook. For the full year of 2020, we forecast the overall semiconductor market growth, excluding memory, to be 8%, while foundry industry growth is forecast to be about 17%. For TSMC, we are confident we can outperform the foundry revenue growth by several percentage point in U.S. dollar term.
Our 2020 business will be supported by strong demand for our industry-leading 7 nm and 5 nm technologies, where we see strong interest from all four growth platform, which are mobile, HPC, IoT, and automotive. Now let me talk about 5G and the HPC as the major long-term growth driver for TSMC. We continue to see strong deployment of 5G networks and smartphones in several major markets around the world. We reiterate mid-teens of penetration rate for 5G smartphones of the total smartphone market in 2020.
We also forecast a faster penetration of 5G smartphone as compared to 4G over the next several years, where silicon content of 5G smartphone will be substantially higher than that of a 4G smartphone. In addition, the significant performance, bandwidth, and latency improvement of 5G network will drive AI application and unlock new usage cases such as real-time response and control across many different types of connected end devices. We believe 5G is a multi-year mega-trend that will enable a world where digital computation is increasingly ubiquitous, which will fuel the growth of all four of our growth platform in the next several years. With 5G driving exponential growth in the amount of big data being generated and continuous improvement in algorithm, a smarter and more intelligent world will require massive increase in computation power. Thus, HPC become another major long-term growth driver for TSMC.
CPU networking and AI accelerator will be the main growth area for our HPC platform. By working diligently to provide the foundry industries the most advanced technologies and making it available to all the product innovators, TSMC can expand the pool of innovators who fuel the semiconductor industry growth. With the successful ramp-up of N7+, and the upcoming ramp-up of N6, N5, and N3, we are able to widen our customer product portfolio and expand our addressable market. We also see growth in networking, thanks to 5G infrastructure deployment over the next few years. With 5G and HPC applications as a major growth driver, we now expect to grow at the high end of our long-term growth projection of 5%-10% CAGR in U.S. dollar terms. Now I'll talk about the ramp-up of N7+, and the status of N6.
As N7 enter its third year of ramp, we continue to see very strong demand across a wide spectrum of product for mobile, HPC, IoT, and automotive applications. Our N7+ is entering its second year of ramp. N7+ is the industry's first high-volume production with EUV photolithography technology, while paving the way for N6. Our N6 provides a clear migration path for next wave N7 products, as its design rule are fully compatible with N7, while providing 15%-20% higher density, which improved power consumption when compared to N7. N6 is on track for risk production in first quarter this year, and volume production before the end of this year. N6 will have one more EUV layer than N7+, and will further extend our N7 family way into the future.
We expect our 7 nm family to continue to grow in its third year and contribute more than 30% of our wafer revenue in 2020. Allow me to talk about our N5 volume production. Our N5 technology is a full node stride from our N7, with 80% logic density gain and about a 20% speed gain compared with 7 nm. N5 will adopt EUV extensively, and is well on track for volume production in first half this year and with good yield. We expect a very fast and smooth ramp of N5 in the second half of this year, driven by both mobile and HPC applications. We expect 5 nm to contribute about 10% of our wafer revenue in 2020. N5 will be the foundry industry's most advanced solution with the best PPA.
We will offer continuous enhancement to further improve the performance, power, and density of our 5 nm technology solution into the future as well. Thus, we are confident that 5 nm will be another large and long-lasting node for TSMC. Finally, I'll talk about our N3 status. We are working with customers on N3 design, and the technology development progress is going well. We have many technology options in development, and we carefully evaluated all the different approaches. Our decision is based on technology maturity, performance, and cost. Our N3 will offer another full node scaling benefit in terms of performance, power, and density as compared with our N5 technology. We expect our 3 nm technology will be the most advanced foundry technology in both PPA and transistor technology when it introduced.
We will announce more details about our N3 technology at our TSMC North America Technology Symposium on April 29th. Thank you for your attention.
Okay, thank you. This concludes our prepared statements. Before we begin the Q&A session, I would like to remind everybody to please limit your questions to two at a time to allow all the participants an opportunity to ask questions. Questions will be taken from both the floor and from the call. Should you wish to raise your question in Chinese, I will translate into 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'd like to remove yourself from the questioning queue, please press the pound or the hash key. Now let's begin the question- and- answer session.
Our first question will come from the floor, Credit Suisse, Randy Abrams.
Yes, thank you. I guess with the trade war eased a bit, there's still a couple maybe political issues out there, so I wanted to just start with that. The first one, if you could give some color on TSMC's U.S. content and just if maybe the calculation, there's talk about the content threshold lowered to 10%. There's some talk the equipment may be excluded from that calculation. Second, with more of the geopolitical concerns about some of the military technology, you've talked in the past about not needing to or having more scale with the fabs in Taiwan. If any new considerations on the fab location?
Yeah. Happy New Year, everyone. On the possible further tightening up of export control from the U.S. government, I think they haven't really announced what the specifics about the rules. Everything I say here can only is a speculation. In particular, I don't want to comment on particular customers. One thing for sure is our business profile is massive. We are everyone's foundry, and we will deal with each customer fairly and equally.
Secondly is, we have been, and we will follow the law and regulation. Upon the regulation being effective, we will carefully study and evaluate product by product our eligibility in the export. We really have a very sophisticated export control system. As you might know, every product is calculated automatically. Every product is different in terms of their content. It's really difficult to describe to you generally what is the content percentage is. I can just tell you that whatever you read on newspaper is not true, okay. We are prepared to deal with this new export control regulation.
Yeah. No, thanks for clarifying, especially that it's not a node by node, which some of the press was speculating. The second question, I just wanted to ask on the higher CapEx, where it came in the high end of guidance and a higher range of CapEx for 2020. If you could talk maybe the areas where the incremental increase, both just from an investment where that new spend is coming, and maybe what changed on the demand side versus a few months ago to lift the budget. The second part, because of this higher base, how you're looking at the base over the, say, the following year, because now we have a two-year higher elevated spend, if, say, some moderation from there?
Yes, we did increase substantial amount on the CapEx. Let me give you some color why we did that. Actually, we expect the mobile phone and HPC, these two segment, probably grows above 20% this year. With another two segment, two platform, automotive and IoT, probably in mid-teens. Put all together, we had to increase the capacity. We work with the customer to fulfill their demand. That's a result of why we increase our capacity. Now you question about our growth?
Yeah, the second question was where the spend, say, like five, seven versus back end.
Oh, okay. A little bit on seven that we announced it last year, and then most of them is in five and then prepare for three. 10% in the back end and 80% in the leading-edge technology, like seven, five, three all together.
Okay. The follow-up I had was just the moderation of CapEx, if you expect that to maybe moderate from the very high level next year. In the back end, 10% would be TWD 1.5 billion for CapEx. I'm curious, maybe between back end and mask, if really over TWD 1 billion back end CapEx.
You're right. I mean, the back end, including the mask.
Okay. I guess of that, do you think the back end CapEx is over TWD 1 billion investment on that side?
I don't.
No.
Okay. Thanks. Okay. Could you talk about the moderation for.
I think, Randy, the last part of your question is just talking about our CapEx this year and last year is at a higher level. Looking out the next several years, where do we think the CapEx will be?
That would depend on growth, right? Yeah, if we enjoyed a good growth in these two years. If it's a success introduction of our N3, I mean, the CapEx probably will not drop. Yeah.
Great. Thank you.
Thank you. Let's go on to the next question. We'll take it from the floor. Citigroup's Roland Shu over here.
Thank you, and Happy New Year. First question, I just follow- up on the CapEx. With this $15 billion-$16 billion CapEx spending, how many of the total capacity increase is going to be this year and how about last year? With this almost TWD 15 billion CapEx spending, how many percent of the CapEx increased last year? Thank you.
Wendell will-
Do you have the number?
Discuss this question.
You're asking about the capacity increase?
Yes.
Right. Okay. The increase mainly come from advanced technologies, right?
Okay. Mm-hmm.
For 2020, we're looking at mid-single- digits capacity increase.
Last year, a low- single-digit number.
Okay. How about the total depreciation is going to be this year?
It will increase by high teens in 2020.
Okay. Thank you. Second question is for the gross margin. For your first quarter revenue guidance, U.S. dollars is down slightly. Apparently, in first quarter, we have fewer working days. It means that your utilization in first quarter definitely is going to be much higher than 4Q. Also, for 7 nm last quarter, we said gross margin has already reached corporate average. Also, I believe in first quarter, we probably won't have the inventory revaluation. All of this, is your gross margin guidance 48.5%-50.5% a little bit conservative? Thank you.
The utilization, we expect it to increase a little bit. Of all the factors that you just mentioned, there is also a potential factor, which is foreign exchange rate impact. There are six factors affecting our profitability. The development and ramp of our advanced technology, pricing, cost, utilization, technology mix, and foreign exchange rate. When you put all these together, that is how we came up with the guidance.
Okay. Okay. Thank you.
Thank you, Roland. All right, next question will come from Goldman Sachs, Bruce Lu.
Happy New Year. I try to look as more granularity color for the 5G penetration rate. Management just mentioned that spot mid-teens in 2020. Can we have more color in terms of that, whether it's sell in or sell through, or what kind of geography or distribution, what kind of distribution between high-end and low-end for the 5G?
Well, I can only say that the 5G's penetration is higher than the 4G, and you know a few countries that they are moving faster than the other area, right? That all we know. We are making the judgment, look at the installation of 5G infrastructure, and we look at each country's adoption, and we do our own estimate. That's why we come up with the mid-teens penetration.
It's more from the top-down perspective instead of bottom-up from each product line. Is that right?
We do both.
Make a judgment.
If you do both, can you give us some color about distribution between high-end?
That's our company's competition.
Bruce, do you have a follow-up?
Yes, of course. Can I double-check that the assumption of mid-teens penetration is based on there is no change in terms of De Minimis rule?
That's a good question. The change of De Minimis rule is still speculative, our forecast will assume the business as usual.
Understand.
I think, yes, the number we currently forecast does not include the De Minimis rule title change. For whatever the export control is coming up, we think the 5G's momentum will continue. If any interruption, it will be very short-term. After going through the supply chain changes and share exchanges, I think the momentum will just as strong, yeah.
Okay, your second question?
The second question is regarding the management. You used to mention that TSMC becomes more important in terms of geopolitical situations. Does that change your equation in terms of building out the fab, the cost structures? In the past, we only care about the manufacturing cost as a main cost factors. Given the current situation, do you think that you have to put that into your consideration as well? I.e., you know management keep on saying that building a factory outside of Taiwan is a lot more expensive than doing that in Taiwan. With all the geographical risk that build a factory outside of Taiwan becomes like, you see the increasing pressure on that. Do we expect any changes on that?
Yes, the cost in Taiwan is the lowest among all regions across the world. We have been studying it continuously, and that decision is made to the best interest of our customers. Yes, the geopolitical is evolving, but we still listen to our customers as the priority. At this point, our customer, when asked to be manufacturing in a higher cost region, their answer is, "We cannot be competitive this way." To maintain the competitiveness of our customer, currently, this is the fab layout we're having. In the future, right now it's too early to say, and still our customer prefer we have the lowest cost production sites and doing business with us.
What if they are willing to pay for a higher price?
They may be special products, but by and large, it's unlikely.
Understood. Thank you.
Thank you. All right. Let's move to the line. We have quite a few callers on the line. Operator, can we take the next call from the line, please?
Sure. Your next question comes from the line of Gokul Hariharan from JPMorgan. Please ask your question.
Thanks. Happy New Year. My first question is on margins. Given that we are looking at a very strong demand pickup, could we talk a little bit about any change in the view on longer term gross margins? Could we see a meaningful improvement in gross margins beyond the 50% range that we have been in the last four to five years? Also, could you talk a little bit about the margin dilution impact in second half of 2020 from 5 nm? Is it likely to be more modest given there's a very strong 7 nm demand also through the course of this year? I had a follow-up question. Thanks.
Okay. Thank you, Gokul. Let me just repeat your question to make sure we got it right. Your first question is asking about sort of, with given the strong demand pickup, is there any change to TSMC's long-term gross margin target? Why could it not be beyond or above 50%? Your follow-up or addition to that is looking at this year, could the margin dilution from our 5 nm ramp in N7+ be more modest given the continued strong demand of our 7 nm?
That's right, thanks, yes.
Right. We continue to use 50% gross margin. We think it's still a very good target. Of all the six, I just mentioned the six factors that will affect our profitability. One of them actually relates to the ramp of new nodes. The ramp of every new node, we will see margin dilutions. This relates to your second question, and indeed, we are seeing a margin dilution, in the second half of this year, from the N5 ramp. Is it going to be better than before? Yes, but only slightly. Okay. The other factors that I just mentioned include the foreign exchange rate, which is really uncertain for anybody to guess. At this moment, we believe the 50% gross margin is still a good target for us.
Okay. Gokul, do you have a second question?
Yeah. Could you talk a little bit about what we are seeing in the N-2 , N-3 nodes? Should we anticipate any recovery in the situation in 28 nm, given overall foundry growth seems to be rebounding? Second part is on 12 nm and 16 nm. Given a lot of the smartphone customer seems to be migrating to 7 nm and even 5 nm because of the adoption of 5G, would there be any challenge to backfill 12 nm and 16 nm as these customers migrate to more advanced process nodes? Thanks.
Okay. Let me just repeat to make sure we understand again. You're asking about our N- 2 and N- 3. First part of your question is, do we see any recovery in the demand for 28 nm given the overall strong demand that we see this year? The second part of your question is that, with a lot of products going from 12 nm and 16 nm very quickly to using our 7 nm and our N5, will we have challenges or difficulty to backfill 12 and 16?
All right, let me answer 28 nm first. With the strong market growth, 28 nm, we did see a little bit better than we expected. We have reiterated saying that 28 nm, the capacity has been overbuilt in this industry. The utilization is still below our average in 28 nm. We expect it will be improved in next one to two years when we develop a new specialty technology for all our customer to utilize it. We start to see the sign because of from a new tape outs, we can be sure that one to two years later that the utilization rate will go back to company's average.
Also 12 and 16?
12 and 16, we did not see that. Today is still very strong demand, and they continue to be a very high utilization rate. It all because of we continue to improve the technology, and so it's being utilized. The first wave smartphone, HPC, and now is IoT automotive. Okay.
Okay. Does that answer your questions, Gokul?
Yes. Thank you very much. Thanks.
Okay. Thank you. We will take the next question from a caller on the line, please.
Yes, sir. Next question comes from the line of Bill Lu from UBS. Please ask your question.
Yeah. Hi there. Thank you very much, and Happy New Year. Question on 5 nm. I remember when 7 nm was ramping, you gave us numbers on the number of tape outs. I am wondering if you can help us with 5 nm either also by giving us number of tape outs or maybe just comparing it to 7 nm and what is higher or lower?
Okay. Bill, sorry, you broke up a bit at the end. I think I understood your question. You want to ask us if we can give you some comparison of the number of tape outs at 5 nm versus our 7 nm at a similar stage.
Well, the 5 nm tape outs.
That's right. Thank you.
Yeah. 5 nm tape-outs is a little bit less than 7 nm compared at the same stage of the time. However, the most important thing is that the high volume tape-out is almost equal. We expect that our 5 nm ramp is very fast and smooth, and it will contribute about 10% to this year's revenue.
Okay.
Sorry, just to finish on my first question. Is the expectation then that the 5 nm will be, in terms of wafer capacity, as big as 7 nm?
We are building the capacity right now to meet the customers' demand. Very high demand. That's all I can say.
Great. Second question is a follow-up on Gokul's question on 28 nm. As you develop these specialty technologies, can you talk about what kind of applications are expected to use these new specialty technologies?
You want to repeat or?
Yeah, sorry. Okay. You're breaking up a little bit. You're asking Bill, on 28 nm, we talk about developing specialty technologies for 28 nm. Your question is what type of applications will be used or these specialty technologies targeting.
Well, let me be a little bit specific. We are developing the 28 nm into 22 nm geometry, and the ultra-low power is one of the directions we are working on, which can be applied to a lot of IoT devices, and also applies to some specialties such as CMOS image sensor and all others. Okay.
Okay. Thank you, Bill. Let's move on. We'll take one more question from the line, and then come back to the floor.
Your next question comes from the line of Brett Simpson from Arete Research. Please ask your question.
Yeah. Thanks very much, and Happy New Year, everyone. I just had a question on China. China was more than 100% of your Q4 sales growth on a year-on-year basis, and sales more than doubled in 2019 from China despite the headwind from crypto. Can you maybe talk a bit more about the region? What's driving so much growth, and how should we think about China growth specifically in 2020? Thank you.
Okay. Brett, please allow me to repeat your question. Your question is about our China business. You point out that it was 100% of our fourth quarter growth and grew quite strong in 2019 despite the drop-off in cryptocurrency. You want us to comment on how we should think about China as a percentage of sales and future growth drivers going forward. Is that correct?
That's right. Thanks, yeah.
You want to-
China is about 20% of our business and has been stable around that number last year, this year. What's changed is, I think the last year we see the China growth particularly strong. Other regions such as U.S. probably growth is less. That is the disjunction for this way. We continue going on, I think we expect to maintain this level.
Well, let me add some color to it. The same thing, China's major business with TSMC is also still 5G and AI. The same thing. Two years ago, probably we have some kind of a big increase in the cryptocurrency, but right now, it's become normal situation.
Okay. Brett, do you have a second question?
Just to follow- up on that question. Looking at 2020, can you maybe just sort of help us with the drivers for growth from China? Do you think it'll be mainly 5G smartphone related? Is it going to be quite a meaningful contribution from other markets like HPC? Any more color would be very helpful.
All right. It will be 5G related. Both are smartphone that will be increased and also the networking that's in the HPC area. That's the two major area that China's business that will be increased in 2020.
Okay. Brett, do you have a second question?
Yeah. Thanks. The second question is really targeted at HPC. Can you maybe talk about how much of your 7 nm capacity is running HPC at present? I think you were planning to ramp that in the second half of 2019, any more color on the portion of 7 nm running HPC would be helpful. How does this scale as we look at 2020, the use of 7 nm for HPC? I think you mentioned 5 nm HPC chips would actually ship in 2020. What would this be used for? Any more help there would be great. Thank you.
Okay. Brett, let me just repeat your question. The first part is how much of our 7 nm capacity is for HPC products? How much was this last year? How do we expect this capacity for HPC to scale in 2020 for 7 nm? The second part is for 5 nm, what types of HPC products or applications are being used on 5 nm?
Well, we do not disclose the capacity breakdown for a specific node. In terms of revenue under HPC, in N7, actually, all I want to say is that it continue to grow, to increase. We expect that this momentum will continue in the next few years. For N5, as we said, is driven by mobile phone and HPC. Still, the two biggest increase in this year.
Okay, thank you. Let's come back to the floor, open the floor for questions. Next question will come from Morgan Stanley's Charlie Chan.
Thanks. Happy New Year. First question is about your future technology developments, right? Next year, I'm not sure if you are going to introduce so-called 5 nm pro process. Can you comment that a little bit? If there's an extension of the current 5 nm, does that mean positively to your gross margin improvements? That is the first question. Thanks.
Well, we continue to improve the performance of each node. Next year, you're talking about the 5 Pro?
Yeah.
Okay.
5 nm Pro
Whatever.
Yeah.
Okay. It will be better than this year's 5 nm, that's for sure. All right? All the major customer will use it.
The gross margin improvement, that will be the same as the previous node. It takes about seven quarter.
Six to eight.
Six to eight?
Yeah.
Okay. Six to eight quarter to reach the company's average. Okay.
Okay. I would assume that 5nm gross margin will continue to improve into next year.
That's for sure.
Okay, thanks. Next is, I'm really interested in your 3 nm, right? Currently, what is the visibility, meaning the cost per transistor? Do you think you can really reduce the cost per transistor level at 3 nm? I'm also curious, given this assumption, do you think HPC or mobile will be a bigger user for your coming 3 nm, for example, CPU or smartphone AP? Thanks.
Per transistor cost, I believe continue to reduce. Okay. That's for sure.
Okay.
Who is going to use it? That's a major question. Again, still, high-end smartphone and HPC will be the users.
Okay.
All right? I cannot be more specific to tell you whom, but that's it.
Can I switch back to some near-term follow-up?
Let's stick with two questions first, and then we get back in the queue, we can come back.
Okay. Sure.
Is that all right? Thank you. Next question will come from CLSA, Sebastian Hou. Over here.
Thank you. Happy New Year. My first question is to follow- up on Chairman's comments about, that you see the 5G momentum continue to be strong, even if there's a change on the U.S. exporting rule. Maybe just some near-term disruption, because TSMC has evaluated the supply chain change. Can you elaborate more about what supply chain change you have seen to give you such confidence that there will be just a near-term disruption, even if there is some change on the exporting rule?
Well, this is a forward-looking analysis. I think some of you also did analysis. Basically, the smartphone, you have to look at this smartphone demand for a year. Look at the 5G penetration per year. Who will be the smartphone supplier? It can change. Who will be their shares? Where the 5G base station being produced, that will change. All these things, it boils down to really the forward-looking smartphone demand. Would that be interrupted? That's the analysis. I think that if any disruption, it will be a shorter- term. Yeah.
Okay. My follow-up question on that is smartphone we understand. If A brands, they lose share, B, C brands will pick up. What about the infrastructure? If there is a disruption on the infrastructure because one of the key infrastructure supplier may not have the critical processors, so which means the whole 5G infra build-out may be slowed down, postponed. What's the point of having those 5G smartphone? With the whole supply chain, whole 5G thesis being postponed.
I think the industry will find there's no business, just one player for too long. Okay. The second, third player will sooner or later come up, it could be pretty soon. In our business, always in a competitive environment. Yes, we have a number one enter the market, but the number two, number three is not too far away.
Okay. Thank you. My second question is on the guidance that the company gave this year, that the foundry industry is growing 17%.
Right.
Semi is excluding memory is about 8%. If I look back historically, usually, I think the formula for TSMC's growth is usually the global GDP, and semi is about 2%, 3% points above that. Foundry, a little bit above that, and TSMC a little bit above that. TSMC get 5%-10% of the CAGR. I think that is very simple formula. Has the formula changed? Historically, semiconductor is highly correlated with the global GDP growth to some extent. If TSMC, or the foundry industry growing faster, is it like the foundry TSMC become the outlier? It's because also that you're implying that the whole economy is growing faster?
This time we have a higher growth rate, right, for foundry and for TSMC. It all because of driven by 5G and AI's application. Whether we can increase our forecast, for example, TSMC always say 5%-10% CAGR, is that our goal? We certainly hope that we can exceed that. This year is still too early to say, but we stay what we said. The foundry industry will be 17%, and TSMC will be better than that.
Sebastian, let me add to this. This year, the formula does change a bit. We put it in the Korean captive, Korean player foundry captive into the foundry. Okay. They do that and constantly. This time we put the Korean players captive into the foundry business. That's why you see the growth quite faster. Yeah.
If we also look at, I think the company also made a comment that you're seeing your five-year CAGR to be at a high end of the 5%-10%. It's not just one year, right? It's just a flash in the pan. I think if you look at the longer time frame, which means I think the correlation with the global GDP, I think that make more sense, more representative. Does that also means that like Semis and tech innovation is going to drive the global economy grow faster in the next couple of years? Is it just because global GDP still grow as the growth like 2%-3%, but Semis foundry TSMC become the outlier. That gap is getting bigger.
We don't want to say it's a outlier. We continue to forecast global GDP still in a normal situation, provided the trade tension between the two big country did not deteriorated. For semiconductor, I want to say that the content of the semiconductor in our life continue to increase, you can see the big example in the smartphone, you can see the big example in the automotive, and you can see that IoT is a big increase also. Now it's changing our world, and that all because of semiconductor content. It's not because of a GDP suddenly grow faster, and it's not because of semiconductor is a outlier. It will continue to be this way.
Yeah. Remember last year, as Wendell said, the semi industry ex memory was a year of decline. Obviously there's a base effect in play as well for 2020.
Oh, yeah. Sure. I'm talking about five year.
Sure.
Okay.
Sure. Yeah. Okay. Thank you. Let's move back to the line, please, and we'll take the next question from the line please, operator.
Yes, sure. Next question comes from the line of Mehdi Hosseini from SIG. Please ask your question.
Yes. Thanks for taking my question. I have one clarification. When you were referring to 5 nm, does that include 6 nm? If it doesn't, what is your view of availability of 6 nm by year end 2020?
Do I have to repeat the question? He said that 5 nm is including N6 or not. N6 is a 7 nm family. We look at the N7+, N6 as one family. N5 is another big node.
Okay. You're still on schedule to have N6 available by end of this year, correct?
Volume production at the end of this year. Right now it's ready for our customers tape out.
Sure. Okay. In terms of just the We're talking about the growth more than 20%. At the same time, one of the leading microprocessor manufacturer based in North America has talked about increased outsourcing, and I just want to get your view, when you look into the longer term, would there be a structural change in semiconductor manufacturing where TSMC would actually be able to grab a higher market share because there would be more outsourcing specifically, from a key company based in North America?
All right. Let me repeat your question, Mehdi. I think you're asking us to comment on the potential for an increased outsourcing from a major microprocessor or CPU vendor. Long-term, could this be a structural change, and the potential for longer-term outsourcing?
Yes.
We certainly welcome that outsourcing continue to grow. For TSMC, all I can say is we develop the technology to meet our customers' requirement. We are confident that we are the best technology leader, and we have excellent manufacturing.
Of course, as a result, we expect that we gain some market share out of it. That is for the future for TSMC's growth. I cannot be more specific than that.
Okay. Thank you.
Let's move on. We have a follow-up question from the line from JPMorgan's Gokul Hariharan.
Thanks, Jeff. Quick question on the high end of the 5%-10% growth. Could you talk a little bit about why only high end of 5%-10% when the CapEx increase seems to be 40%-50% from the last kind of TWD 10 billion-TWD 11 billion kind of CapEx range over the last five years? Any reasons why we are a bit more cautious? Do we feel that the 5G cycle after a couple of years could start to kind of decelerate, especially given this year we are already starting off with a very strong 20% kind of growth?
Just wanted to think about the puts and takes in terms of the high end of 5%-10% and why not stronger than that growth, given the big jump in CapEx and indication that CapEx could stay around these levels even going into the 3 nm era.
Okay. Gokul, let me try to summarize your question. Basically, Gokul is asking why our long-term growth target is only at the high end of 5%-10% when our CapEx has increased 40%-50% versus the TWD 10 billion-TWD 12 billion in the past. He is asking or wondering, is this because we take a more cautious view that 5G cycle may be strong this year and next year, but may slow down after that? Why do we still say at the high end of 5%-10%?
Last year, before we increased the CapEx, we were looking at somewhere in the middle of that 5%-10%. Afterwards, when we see the ramp in 5G deployment, we increased the CapEx, and now we're looking at high end of the 5%-10% range. That is the difference. Also, let me explain this from a capital intensity point of view. Last year, while we increased the CapEx, the capital intensity was over 40%. This year, we think it will be lower than 40%. From next year on, although it's still pretty early, we think it will be somewhere between 30% or 35%, which is pretty similar to the old norm that we used to say before.
Gokul.
Okay. Thank you.
Do you have a second question?
Yeah. Just one more broader question for Dr. Liu. At a board level, Dr. Liu, could you talk a little bit about how the board thinks about TSMC's positioning as a foundry for everyone, everybody's foundry given the broader geopolitical changes that are happening? I don't want to go into each episode in terms of the change in the streaming rules, et cetera. Thinking four to five years out, what are the steps that the board is considering to kind of ensure that TSMC can remain everybody's foundry, even in a more challenging kind of geopolitical environment and a lot more policy kind of risk compared to, say, the last five or seven years?
Okay. Let me try to summarize your question. I think Gokul's asking for Mark to please share your thoughts on, from a board level, TSMC's positioning as everyone's foundry. Of course, we're facing a lot of different geopolitical changes and challenges. Gokul doesn't need us to comment on each one. Generally, how are we thinking about five to seven years out, how TSMC can position ourselves and how we can remain to be everyone's foundry?
Yeah. First of all, currently, we discuss the strategy with the board, the board fully agree with our current strategy. Okay. Of course, this strategy contains several necessary component. First of all, we develop our technology ourselves. All the technology IP and know-how and technology all developed in Taiwan here. Secondly is another necessary element is our technology has to be leaders. When you're technology leaders, people will have to come to you. That's how we maintain to be everyone's foundry. There are exceptions, of course, because of their domestic trade policy that I cannot overcome. Basically, that so far this strategy should be able to play on.
Okay. Does that answer your question, Gokul?
Yes. Thank you very much.
All right. Thank you. Let's come back to the floor and see if there's any follow-up questions from anyone. Morgan Stanley, Charlie.
Thanks for taking my follow-up. First of all, for first quarter, you mentioned that your validation rates are higher, but in terms of U.S. dollar revenue go down slightly. What is the ASP or product mix change here? Thanks.
Charlie is asking that we said first quarter, the utilization rate will slightly increase, but our guidance shows a slight decrease in the revenue in terms of U.S. dollars. Why is that? Does that imply an ASP change?
Last year, fourth quarter or even third quarter last year, part of the wafer revenue come from wafers prepared in the first half of 2019 when the utilization was pretty low. We're pretty much digesting all of those already.
Okay, thanks.
It looks like the year is open, but do you see any kind of next data points, for example, any segment or any customers cutting forecasts or orders recently? Can you comment on that?
No, we did not see that.
Okay. Lastly, it's a little bit subtle. CapEx. I think three months ago, I think the guidance for this year's CapEx is $14 billion-$15 billion. Now it's like $1 billion higher.
What is that additional CapEx for? Is it mainly for 7 nm or 5 nm?
Well, other than the advanced technology, as mentioned earlier, we also increased the CapEx this year for specialty technology as well as advanced packaging. Those are the areas that we are focusing on.
Okay. Lastly, if I may, I guess a market share question. We appreciate that the company provide your assumption for industry growth. I guess first of all, we want to clarify, when you add Korea captive foundry in the comparison, is that a apple-to-apple comparison? Meaning do you include that into last year's revenue base?
Yes. I think we did, right?
Yes, we did.
Okay. In truth, an apple to apple comparison-
Yes
up 17%.
Up 17%.
Okay.
I'm not sure.
China competitors' market share. I think lots of news-
I think that's probably needs to correct it.
Yeah.
I don't think we include the foundry growth in last year, the Samsung captive supply.
Oh.
Yeah. The growth is 17% is particularly high.
Okay. What is a kind of apple to apple comparison then?
I think six points. It will be 11%, if you apple to apple comparison.
Okay, thanks. Based on that kind of 11% industry growth, I think lots of investors are asking whether you're sort of losing market share in China, because China want to push the localization. Some chairs talking about SMIC can take more orders from you guys as 40 nm. Can you also comment a little bit on this front?
Usually we don't specifically answer this kind of very sensitive questions specific on one competitor. Let me tell you that what the newspaper said is not true. All right?
Okay. Clear. Thanks.
Thank you. Yeah. Charlie, also let me just clarify. Last time we did not say that $14-$15 for 2020 CapEx. We said it would probably stay at a similar level as 2019, but we did not specifically say that dollar range. All right, just to clarify. All right, let's go back to the line. Sorry. We have a follow-up from SIG, Mehdi.
Yes, thank you for the follow-up. Want to go back to your commentary about 5G phone and how should I think about opportunities if I were to think about the sub-6 versus millimeter wave? Specifically, do you think that this year opportunities with mobile, you talked about more than 20% growth, is that all going to be sub-6 or is it going to be a mix of the two technologies?
You want to repeat?
I think-
Okay.
Yeah.
Okay. The question is clear enough. If the 5G growth is because of sub-6 or the millimeter wave, my answer is both. Okay. The 5G is a phone and the base station is for sub-6 and the millimeter wave.
Do you have a second question, Mehdi?
Yes. Actually, if I were to have a follow-up, do you think you would actually be building a millimeter wave phone, or for millimeter wave, it's just going to be limited to base station?
Okay. Actually, the phone is much easier to build with millimeter wave plus sub-6. The base station probably will have a much higher number in sub-6 rather than millimeter wave.
Okay?
Okay. Thank you.
All right.
Thank you.
Let's come back to the floor. We have follow-up from Credit Suisse, Randy Abrams.
Yes, thank you. You talked about the good investment in the back end. Last year, I think you said TWD 2.5 billion was the revenue run rate in 2018. If you can maybe give a view of the size of the business now, and maybe what type of growth you're expecting for this year.
You want pay?
Yeah. The revenue size of back end was TWD 2.8 billion in 2019. We're expecting double-digit growth for this year.
Okay.
Mid-teens.
Mid-teens, great.
Yeah.
Okay. If I could ask on the two other areas that don't get as much attention, automotive and IoT. Automotive was depressed last year, but I think you're talking about a pretty big pickup to grow teens. Could you talk the areas, like if it's just cyclical rebound, or if there's certain types of product or components coming back for the automotive where you're gaining content share? For the IoT, where it was very strong growth, it's a big category. If you could maybe center on if there's a few particular pieces within an IoT driving the momentum for that category?
Well, actually, the growth this year most come from the content increase rather than the unit increase. I just mentioned that it will be mid-teens. Mid-teens increase certainly is not the mid-teens units. You don't expect so many car being sold. It's the semiconductor contents increase more important than the unit.
Randy is also asking about for IoT, are there specific areas or segments that's driving the growth in IoT?
Well, wearable is very popular now everywhere.
Both unit and content.
According to our Chairman, is content also increase.
Unit.
Okay. Thank you for that.
Okay.
Next, we have a follow-up from Citigroup, Roland Shu.
Yes. Last time, CEO guided you expected 7 nm revenue will continue grow in 2020. Do you still hold the view?
Yes.
How much growth it will be this year? Last year we have 27%. Is this able to above the highest level of 20 nm , around 34%?
Close.
Okay. Thank you. Also for 7 nm, do you see any competitor with the technology breakthrough and likely to impact or taking your market share in the near- term?
Answer the question. Can you repeat the question?
Can you repeat the question? Sorry.
Yeah. I said, do you see any competitors with 7 nm technology breakthrough and were likely to threat you or take your market share going forward?
We will continue to hold a very high market share. That's all we can say. I don't comment on my competitor.
Okay. Thank you. Yeah, my second question is, you said that you continue expanding your customer and the product portfolio. You have new customers and the products from cryptocurrency in 2018. Last year you have this new CPU foundry also seen as your new customer and technology. How about this year? Do you see any new customer or new applications to contribute to your growth? Thank you.
If this year's growth counting on my customer, it will be engaged last year already. As I said, we continue to expand our product portfolio, and we continue to increase the number of our customers.
I think the content increase along with 5G penetration is the major phenomena, including the leading edge as well as the mature nodes. It's just a widespread of customers or existing customers.
Thank you.
Okay. We have a follow-up here from CLSA, Sebastian Hou.
Hello. My first follow-up is, we heard that the 7 nm demand is very strong at TSMC, and most of the customers are on allocation mode right now. With the step-up of CapEx for this year, I believe some of that also devolved into 7 nm for this year. Do you still expect the similar tightness that your customer may experience on 7 nm by the end of this year?
We do have very high demand from 7 nm, and we work very hard to meet customers' demand. Last year, we announced that we put TWD 1.5 billion more to increase the 7 nm capacity, and we work hard to increase the capacity.
Yes. With 5 nm ramp-up in the second half of this year, the tightness of 7 nm will hopefully can be soothed a bit for the customers.
Okay. Are there any process nodes that TSMC is seeing not growing this year?
20 nm.
20 is already pretty low last year. You're still seeing that not growing?
You say not growing, right? Yeah. Not growing.
All right. even for 16, 12, this platform, TSMC also expect that to go up?
Okay. That's why you say that the newspaper is wrong.
I don't want to so specific.
Okay. Well, that's pretty clear. Thank you.
All right. In the interest of time, we'll see if there's any last questions from anybody. If not, then this concludes our Q&A session. Before we conclude today's conference, please be advised that the replay of the conference will be accessible within four hours from now. The transcript will be available within 24 hours from now, both of which will be available through TSMC's website at www.tsmc.com. Thank you for joining us today. We hope you will join us again next quarter. Goodbye, Happy New Year, and have a great day.