艾克爾 (AMKR) 2017 Q1 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the Amkor Technology First Quarter Earnings Conference Call.

My name is Michelle, and I will be your conference facilitator today.

(Operator Instructions) As a reminder, this conference is being recorded.

I would now like to turn the call over to Greg Johnson, Vice President of Finance and Investor Relations.

Mr. Johnson, please go ahead.

Thank you, Michelle, and good afternoon, everyone.

Joining me today are Steve Kelley, our President and Chief Executive Officer; and Megan Faust, our Chief Financial Officer.

Our earnings press release was filed with the SEC this afternoon and is available on our website.

During this conference call, we will use non-GAAP financial measures, and you can find the reconciliation to the U.S. GAAP equivalent on our website.

We will also make forward-looking statements about our expectations for Amkor's future performance based on the environment as we currently see it.

Of course, actual results could be different.

Please refer to our press release and other SEC filings for information on the risk factors, uncertainties and exceptions that could cause actual results to differ materially from these expectations.

Please note that the financial results discussed today are preliminary, and final data will be included in our Form 10-Q.

And now I would like to turn the call over to Steve.

Good afternoon, and thanks for joining the call.

Today, I'll discuss our first quarter results, our second quarter forecast and our key growth initiatives.

In addition, I will discuss Amkor's wafer-level packaging strategy, including our pending acquisition of NANIUM.

Our first quarter operating results came in as expected, with revenues up 5% year-on-year.

Although mobile communications demand was seasonally slow, most other end markets stayed healthy.

Excluding a noncash foreign currency loss, we achieved breakeven profitability in the quarter.

EBITDA was $154 million, and free cash flow was $17 million.

In the second quarter, we expect a 9% sequential and year-on-year increase in revenue.

Over half the sequential growth should come from the mobile communications market.

Our expertise in advanced SiP, MEMS and wafer-level packaging makes Amkor a natural choice for smartphone IC designers, who must continually find ways to squeeze more circuitry into smaller spaces.

In the second quarter, we also expect increased demand in other markets, including automotive, industrial, consumer, computing and networking.

Moving on to our key growth initiatives.

In Greater China, Amkor revenue grew 60% year-on-year.

Nearly all of that gain came from advanced packages and test.

Many of those packages are manufactured at our Shanghai factory, which offers 8- and 12-inch wafer bumping, wafer-level CSP, MEMS, through mold via and through silicon via technologies.

Our technical support teams in Shanghai, Shenzhen and Beijing work closely with customers to ensure high-yielding, high-volume success in both assembly and test.

Our advanced SiP business grew 7% year-on-year and accounted for nearly 20% of first quarter revenue.

RF, memory and sensory applications drive most of our advanced SiP volume.

Our ability to manufacture complex advanced SiP products will become even more important in the future, since package-level integration, in many cases, will offer the most economical way for our customers to miniaturize.

This is particularly true as leading-edge silicon geometries shrink to less than 10 nanometers.

There are 2 basic platforms for advanced SiPs.

The first utilizes a thin substrate or laminate.

Nearly all of today's advanced SiP products are built on a laminate platform.

The second platform is wafer-based SiP, so named because the packages are constructed using 12-inch carrier wafers.

This high-density integrated fan-out technology results in a super thin product, typically employed in high-pin-count smartphone IC applications.

Looking forward, we believe that laminate-based advanced SiP designs will continue to proliferate, and that wafer-based SiPs will become a significant part of the overall market.

SWIFT, short for silicon wafer integrated fan-out technology, is our brand name for wafer-based SiP.

We have installed a SWIFT prototype line at our K5 facility and are working on joint development projects with a number of customers.

Automotive IC has accounted for 26% of our first quarter revenue.

We are the #1 OSAT for automotive ICs due to our long history and investment in the business, our geographic and financial stability and our ability to meet automotive quality specifications.

And moving to my final topic, Amkor's wafer-level packaging strategy.

This strategy is critical to our long-term success since wafer-level packages will drive a substantial part of our future growth.

Wafer-level packaging is pervasive in small-form factory applications.

As an example, each iPhone 7 contains 44 wafer-level packages, according to Techsearch International.

This technology is also used extensively in Android phones and has found its way into other space-sensitive applications, such as tablets, smart watches, cameras and fitness trackers.

The primary advantage of wafer-level packages is that they are small and thin.

They're also generally less expensive than substrate-based solutions and are very reliable when manufactured by a first-tier OSAT, such as Amkor.

Amkor has over 10 years of experience in the wafer-level packaging business.

Today, we operate high-volume wafer-based manufacturing lines in 3 countries, serving over 100 customers.

Our total installed capacity is over 300,000 8-inch wafer equivalents per month.

There are 3 basic types of wafer-level packages: wafer-level chip scale; wafer-level fan-out; then high-density fan-out.

Wafer-level CSP is a fan-in technology, since all of the bond pads or bumps fit within the 4 corners of the die.

Like all wafer-level packaging, it's manufactured in a clean room environment.

Wafer-level CSPs are built while the die are still in wafer form.

Now wafer-level fan-out packaging, also known as low-density fan-out, was invented many years ago, but only went to high volume last year.

Wafer-level fan-out makes sense for ICs where the die surface area is too small to accommodate all the bond pads.

The fan-out package essentially enlarges the bondable surface area by building the border around the die using low-cost molding compound.

Wafer-level CSP and wafer-level fan-out are complementary technologies.

Customers can toggle between the 2 package types as their die sizes shrink or grow.

The third wafer-level package is high-density fan-out, which we call SWIFT.

SWIFT-based SiPs are roughly 40% thinner than laminate-based SiPs.

They're ideal for large, high-pin-count die in space-constrained applications.

This is a new technology just introduced last year.

We anticipate ramping SWIFT to high volume in the 2018, 2019 time frame.

Amkor is the leader in wafer-level CSP and high-density integrated fan-out technologies.

With the acquisition of NANIUM, we will have an equally compelling proposition in the low-density fan-out area.

NANIUM is generally seen as the fan-out technology leader as well as a very capable manufacturer, having shipped over 1 billion fan-out packages.

Their yield and quality performance are generally considered to be best in class.

In summary, wafer-level packaging is a large and fast-growing market, driven by the form factor and performance requirements of the smartphone market.

Only Tier 1 OSATs like Amkor with engineering expertise, deep pockets and competent service organizations can compete effectively.

We will continue to invest to support our leadership position in this market.

Megan will now provide more detailed financial information.

Thank you, Steve, and good afternoon, everyone.

Today, I will review our first quarter results, then provide some comments about our second quarter and full year activities.

We delivered record first quarter revenues of $914 million, up $45 million year-over-year.

Our business was essentially breakeven for the first quarter and in line with guidance.

We generated approximately $850 million of EBITDA in the last 12 months.

Earnings in the quarter were reduced by a noncash foreign exchange rate loss of $11 million or $0.05 per share related to the remeasurement of our balance sheet liabilities denominated in foreign currencies.

Most of this loss was due to our Korean severance plan liability.

We have started to increase our Korean won-denominated assets to provide a natural hedge against future movements in the won.

First quarter gross margin of 15.6% included approximately $8 million or about 90 basis points in costs related to our factory consolidation activities in Japan.

Operating expenses in the first quarter were $118 million and included labor, depreciation and other cost associated with our new K5 facility in Korea.

These costs are being classified as research and development expense until we commence high-volume manufacturing.

At March 31, we had total debt of $1.5 billion and debt to EBITDA of 1.8x.

This is a significant improvement from our ratio of 2.4x at the end of 2015.

Our liquidity is solid with $614 million in cash and around $350 million in available undrawn loans.

We generated $17 million in free cash flow for the first quarter and expect to be free cash flow positive for the full year.

Turning to the second quarter.

We expect to complete the sale of our K1 factory in Korea in May.

The sales price is approximately $140 million.

And we expect to recognize a pretax gain of $110 million in operating income.

After tax, the K1 sale will contribute approximately $80 million or $0.33 per share to our second quarter earnings.

We are currently evaluating a variety of uses for this anticipated increase in liquidity, including further investment in our business and the repayment of debt.

Driven by our strong Q2 outlook of around 9% sequential revenue growth, we expect operating income to double, excluding the K1 gain.

Q2 includes approximately $9 million in cost related to our factory consolidation activities in Japan.

We plan to complete these activities before year-end.

Operating expenses are expected to be around $120 million.

As in Q1, R&D expense includes labor, depreciation and other cost associated with K5.

By the end of 2017, we foresee quarterly operating expenses to return to more traditional levels of around $100 million as we bring up production in K5, and these R&D costs transition to cost of goods sold.

Heading into 2018, we expect that the incremental fixed cost associated with K5 will be offset by the savings from the closure of K1 and the completion of our factory consolidation initiative in Japan.

We expect our 2017 full year effective tax rate to be about 25%.

Our full year 2017 forecast for capital investments remains unchanged at approximately $500 million, give or take 10% depending on demand.

And finally, our pending acquisition of NANIUM is expected to close during Q2.

Our Q2 guidance does not include any financial results of NANIUM or the effects of purchase accounting.

With that, we will now open the call up for your questions.

Operator?

(Operator Instructions) Our first question comes from Randy Abrams of Crédit Suisse.

Yes.

The first question I wanted to ask just on the outlook for high single-digit growth in second quarter.

Maybe just characterize the component that might be from, say, market share gain or Amkor-specific, if you see specific drivers or opportunities ramping up.

I'm curious more from the difference in growth rate versus foundry, if you think it's more an issue of wafer bank inventory.

We're seeing the package and test kind of, you could say, outgrow, whether there might be more overhang.

I'm curious if it's market share or maybe a difference in the profile in the supply chain, that's driving some of your growth relative to foundry, which you're guiding flat to down in the second quarter.

Yes, Randy.

Let me just make a comment.

We take a look at the increase from Q1 to Q2, about half of it's coming from mobile communications.

And the other half is broad based across all our markets.

If I zero in on mobile communications, I think the biggest distinction for Amkor is our participation in the Galaxy phone.

We have the reasonable content on that phone.

Obviously, that's ramping in production this quarter.

So that maybe helping us relative to some other participants on the supply chain.

But we're also seeing an increased activity from the other ecosystems in the mobile communication space.

The other markets, I think it's just a collection of things.

In some cases, the customers are having more success in the market in Q2.

In some cases, we're having more success at those customers, and it's really hard to generalize on any other markets.

Okay.

And the second question, were you -- I appreciate the details on the fan-out and wafer-level packaging.

If you could talk a bit more about the project timing, kind of what you're seeing.

It sounds like it's a 2018 event, but what you're expecting for the market?

Like, if you see a meaningful, say, high-end mobile or if it's other applications that may move toward the SLIM and SWIFT architecture?

And if you see that it's a sole source or you may be, like, sharing with a TSMC info?

And the other side I wanted to ask at how you see the capital intensity and the profitability, if -- as this business ramps relative to, say, advanced packaging or flip chip or relative to the mainstream business.

Okay.

So the timing, Randy, I think that we'll see continued qualifications through the next year.

I think we'll probably see a first ramp starting late '18 or, perhaps, into '19.

I think it's a relatively new package and people need to work with it and really get comfortable with the package before they ramp.

Obviously, there's some exceptions do that, and TSMC with their info package and their single customer today.

But I think it'll ramp in a gradual way.

As I look regarding the profitability -- let me just give you some insight into how we're going to ramp K5.

So K5 is the factory where we're going to install SWIFT.

So today, we have a prototype line in K5.

And we have a number of customers we're working with on joint development SWIFT projects.

The next step is to start building more capacity in K5.

And so the way we're going to do that is to add capacity.

But in the interim, before SWIFT ramps to high volume, we'll start building other packages there that utilize the same equipment that we're installing for SWIFT, for example, bumping, wafer-level CSP, low-density fan-out product and SiP.

Part of the benefit of acquiring NANIUM is we could take their very high-yielding wafer-level fan-out process and also install it in K5, which we intend to do.

And so the basic objective is try to start generating revenue in K5 before SWIFT goes to high volume.

Okay.

And if I could clarify the startup cost just tied to that K5.

It sounds like it's a production ramp-in for -- like, for the next few quarters, it's OpEx.

OpEx would come down.

With the cost of goods sold, you'd have the offsetting impact.

And then just clarify it's a -- that K5 will be more -- I might have misheard it, fourth quarter is when the volume ramp or is it actually coming up more sooner?

I'll comment on the volume ramp.

I'm going to turn over the other questions to Megan.

But we expect some low-rate production Q4 this year, but we don't see the SWIFT production starting until Q4 of next year, perhaps 2019.

So Randy, with respect to the cost associated with K5, you're correct.

Those are currently in OpEx.

And then as Steve mentioned, as the business ramps to high volume, those costs will then transfer to cost of goods sold based on the timing of those production activities.

(Operator Instructions) Our next question comes from Sidney Ho of Deutsche Bank.

Steve, you talked about half the growth in next quarter and your guidance will come from mobile.

How would you assess the current state of inventory in the smartphone supply chain?

How much of a snapback do you expect in Q3 given some very strong tailwinds that we're aware of?

And maybe another way of asking the same type of question, do you expect communications revenue in the second half of this year to be better than last year?

And conceptually, I would think it should be, but just want to get your sense -- get your opinion on it.

Okay.

So 2 core questions there, one about inventory in Q2 and the second about my view on the second half.

Let me just answer in order.

It's difficult to understand exactly what the inventory is in the smartphone supply chain.

But I was in China a couple of weeks ago and talked to a number of participants in that supply chain.

The general feeling was that the inventory was being burned off and that things were starting to turn around in Q2.

But again, I don't have any specific numbers to back that up.

It's just a general impression I had based on my visit 2 weeks ago to China.

I think the other positive thing in Q2 is the ramp of the Samsung Galaxy phone.

We have the content, and I think the phone has been well received.

So we think that's going to drive a lot of volume for us this quarter into next quarter.

So as I look at second half, again, I'm going to talk about this in 2 parts.

The first part, smartphone, and the second is the general market.

I can't provide you specific guidance, but let me give you some soft color.

So on the smartphone side, mobile communications, we do expect a very healthy second half.

We see continued strength out of the Galaxy into Q3.

Obviously, we have some new iPhone products, which will be launched in the second half, where we have a good share.

And then I think you'll see resurgence from China as well as the inventory issues dissipate.

So I think it's going to be a very good half for the industry and for Amkor on the smartphone side.

The general market is obviously a collection of a lot of different submarkets.

Looking back, we saw the general market strengthened in Q3, obviously, it was a strong Q4, strong Q1.

It continues to be strong.

So the real question is, when does that cool down?

And the short answer is we're not sure.

However, we do check the data.

One of the things we check is the status of the [disde] channel, specifically inventory levels.

And right now, the feedback that I'm getting is the inventory levels are relatively low.

And there are shortages in selected in areas, such as memory and power discrete.

So that's generally a pretty good sign for the market.

The second data point is the overall economic outlook.

There was an April meeting of IMF and they revised the outlook upward for this year to 3.5% global growth and 3.6% for next year.

In general, anything above 3% for global growth is good for semiconductors.

So I would say, we're monitoring the situation, but we don't see a cooling in general market anytime soon.

Okay, that's super helpful.

Maybe just on the general market comments, the -- if I look at Q1, your auto industrial segment was down roughly 7%, maybe just rounding.

And there have been some chatters about inventory build in automotive.

Can you give us some color about whether your revenue decline was driven mainly by auto segment, industrial or some others and whether you have any insight about the current inventory level at your customers or customers' customers in the auto supply chain?

Yes, Sidney.

We did a little research in the automotive market prior to the call.

Let me just tell you what we found out.

We basically looked at all of our automotive customers.

And most of them are expecting continued growth this year.

There are a couple of exceptions, and those exceptions tend to be more exposed to the North American market and less exposed to Europe and Asia.

But if you take a look at the car forecast, both Europe and Asia, particularly China, are forecasting significant increased production this year.

I think North America is flat to down.

The other aspect is electronic content.

And clearly, the content continues to go up.

If you average content now, it's somewhere around $350 per car.

Hybrids are consuming about $600 in electronics, and luxury cars are about $1,000.

So the expectation across the board is for that number to continue to increase for the foreseeable future.

The last comment I would make will be the opportunity for OSATs to participate in the market.

Typically, many automotive customers have built their products in-house.

We're seeing more receptivity to OSAT participation for a few reasons.

One is the reluctance to make further capital investments.

The second is that many of the newer ICs are available only in advanced packages, and that's typically a very expensive back-end capacity build and where we could offer cost advantages.

And the third is a trend towards SiP.

So obviously, we've seen it in smartphones.

But now, we're starting to see that in selected parts of the car.

So this System-in-Package products is really something we're quite good at, and it's hard to replicate that within an OEM customer.

Great, that's helpful.

If I can squeeze one more in.

On the OpEx side -- this is for Megan, I guess.

Your Q1 OpEx is quite a bit higher than what I expected at $118 million.

I had modeled about $110 million.

Is the increase all because of the K5 expenses that you mentioned?

Or did it have something to do with the foreign exchange?

And given the move in currency in the last few months, how should we think about the currency impact?

And can you walk us through the moving parts?

And what is built into your guidance in Q2?

Okay.

So there was quite a few questions there.

So I will try to answer those in order.

And so from an OpEx perspective, yes, we are at about 120 -- $118 million with our Q1 actuals.

And just a reminder, 2017 is really a year of transition for us with the opening of K5, and that's what you're seeing in OpEx.

There's also the closing of K1, which happened in Q1.

And then we also are having our factory consolidation activities in Japan, and that's having an impact on both COGS and OpEx.

So once we complete those initiatives, we'll expect to achieve a more efficient cost structure.

So with respect to what you're seeing in Q1 and the increase in OpEx, some of that is related to K5, but there's also some increase in some seasonal SG&A costs that are in there, so it's not all specific to K5.

As far as the Q2 model that's guiding at $120 million, that's really consistent and flat with what you're seeing with Q1 from an OpEx perspective.

So your question about foreign currency, foreign currency really is not playing at all in the income statement as far as an OpEx impact for what we're estimating for Q2.

Thanks very much, Sidney.

There are no more questions, so I -- we will end the Q&A portion of the call.

And now I would like to turn it back to Steve for his closing remarks.

I would like to recap our key messages.

First, the first quarter played out as expected, no surprises.

We expect to grow revenues 9% in the second quarter.

And our initiatives are on track.

It'll help us to grow revenues, which will, in turn, drive stronger profitability and cash flow.

And thank you for joining the call today.

Ladies and gentlemen, this concludes today's conference call.

You may now disconnect.