AXT Inc (AXTI) 2016 Q4 法說會逐字稿

Good day, ladies and gentlemen and welcome to the AXT fourth quarter 2016 earnings call. At this time all participants are in a listen-only mode. Later we will conduct a question-and-answer session and instructions will follow at that time. (Operator Instructions). As a reminder, this conference is being recorded. I would like to introduce your host for today's conference, Gary Fischer. Sir, you may begin.

Thank you, Terrence and good afternoon, everyone. Before we begin, I would again like to remind you that during the course of this conference call, including comments made in response to your questions, we will provide projections or make other forward-looking statements regarding, among other things, the future financial performance of the Company and our ability to control costs and improve efficiency, increase orders in succeeding quarters, improve our competitive position as the market improves, our ability to meet market demands for our products as well as other market conditions and trends including expected growths in the markets we serve.

We wish to caution you that such statements deal with future events, are based upon management's current expectations and are subject to risks and uncertainties that could cause actual events or results to differ materially. These uncertainties and risks include, but are not limited to, overall conditions in the markets in which the Company competes, global financial conditions and uncertainties, market acceptance and demand for the Company's products and the impact of delays by our customers on the timing of sales of products.

In addition to the factors that may be discussed in this call, we refer you to the Company's periodic reports filed with the Securities and Exchange Commission and available online by link from our website for additional information on risk factors that could cause actual results to differ materially from our current expectations. This conference call will be available on our website at axt.com through February 22, 2018.

Also before we begin, I want to note that shortly following the close of the market today, we issued a press release reporting financial results for the fourth quarter of 2016 and the year ending 2016. This information is available on the investor relations portion of our website at axt.com.

Thank you and now turning to a review of our fourth quarter results. Total revenue in the fourth quarter of 2016 was $20.3 million compared with $21.9 million in the third quarter of 2016. This is above the high end of our expectations which were $18.5 to $19.5 million. In an effort to understand the underlying trends of our business, many of you have asked us for a bit more granularity in the way we report revenue. Therefore beginning this quarter, we will also break out revenue results for substrates and for raw materials.

In Q4 2016, revenue from substrate sales was $16.0 million compared with $17.0 million in the prior quarter. Revenue from our raw material joint venture was $4.3 million in Q4 compared with $4.9 million in Q3. In the fourth quarter of 2016, revenue from North America was 6%, Asia-Pacific 71%, and Europe was 23% of total revenue. In the fourth quarter, we had no customer that generated more than 10% of revenue and the top five customers generated approximately 35% of total revenue reflecting again our diversification of both products and customers.

Also, for the entire year of 2016, we had no customer that generated 10% or more of revenue. Gross margin in the fourth quarter was 37.1%. This is an improvement from the prior quarter and is the result of a favorable mix and good progress in manufacturing efficiencies and yield improvements. Total OpEx in the fourth quarter was $5.2 million. This is up a little from the prior quarter but in general we have continued to keep OpEx relatively flat quarter-to-quarter.

Total stock compensation expense was $298,000 for the fourth quarter of 2016 which was $5,000 included in cost of revenues, $246,000 in SG&A and $47,000 in R&D. Operating profit for the fourth quarter of 2016 was $2.3 million compared with $2.7 million in the previous quarter. Interest and other income in the fourth quarter was a net charge of $274K. This net number consists of four categories. One; interest earned of $106K.

Two; foreign exchange loss of $95K. Three; equity accounting on our unconsolidated joint ventures of a loss of $558K. And four; other items equaling a gain of $274K. For Q4 of 2016, we had a net profit of $2.2 million which is $0.06 per share. This is north of our guidance which was a range of a profit of $0.02 to $0.04 per share. By comparison we had a net profit of $2.2 million or $0.07 per share in the third quarter of 2016.

Cash and cash equivalents and investments increased nicely in the quarter by $6.4 million, moving us from a total of $47.3 million as of September 30, 2016 to $53.7 million as of December 31, 2016. Depreciation and amortization in the fourth quarter was $1.3 million and CapEx was about $900K. Accounts receivable net of reserves were $13.5 million at December 31, 2016 compared with $18.4 million at September 30, 2016.

Net inventory at December 31 was $40.2 million compared with $38.7 million in inventory at September 30. Ending inventory consisted of approximately 44% in raw materials, 50% in work in progress, and 6% in finished goods. The spread between the three buckets remains consistent. That covers the quarterly comparison and allow me it briefly highlight the entire fiscal year which is our calendar year.

For the fiscal year 2016, revenue was $81.3 million, up from $77.5 million in fiscal year 2015. In 2016, revenue from substrate sales was $65.6 million compared with $58.2 million in 2015. Revenue from our raw material joint ventures was $15.7 million in 2016, compared with $19.3 million in 2015. Gross margin for the fiscal year 2016 improved meaningfully to 32.4% of revenue up from 21.7% of revenue for fiscal year 2015.

Our profitability in fiscal year 2016 also increased meaningfully. We achieved net income of $5.6 million or $0.17 per share compared with a net loss of $2.2 million or $0.07 per share for fiscal year 2015. That is a swing of $7.9 million year-to-year moving from a loss to a profit which, I guess to use non-financial terms, is pretty cool. This concludes our financial review.

I'll now turn the call over to Dr. Morris Young for a review of our business.

Thank you, Gary. And good afternoon, everybody. Our positive Q4 results capped off a year of growth and diversification of AXI. Revenue came in ahead of our expectation driven by a growing number of exciting technology applications. In our vision, we continue to demonstrate improvement in our business model. Our gross margin performance is the highest we have seen in many years. These are partially the result of continued leverage for a number of cost improvement programs implemented over the last 18 months. As well as good progress yields.

We also benefited from an optimal product mix in the quarter. Our profitability also exceeded our expectation in the quarter and with continued (inaudible) we generated healthy positive cash flow. As we enter into 2017, we encouraged by the evidence that the market is evolving in our favor in several strategic areas that can provide further growth and expansion opportunities.

The first is indium phosphide. Indium phosphide bounced back nicely in Q4. Overall we continue to benefit from the demand for PON equipment which is expected to rise to more than $43 billion by year 2025 according to a December report by Grand View Research. Growth in this market is being driven globally by the increasing adoption among telecom carriers for fiber to the home and fiber to the office applications.

As we have noted in Q3 and even into October, we saw some weakness in the demand for indium phosphide in certain geographies such as Taiwan and China which is likely the results of too much inventory. Improvement in the second half of Q4 is encouraging, but a bit more time perhaps is needed to sort out their inventories. Looking forward, we believe that the industry transition from the 2.5G to 10G power will provide further growth opportunity in the second half of 2017.

Now turning to data center connectivity. We continue to see growing emphasis on indium phosphide base silicon photonics technology. This is likely a function of both the massive growth in data driven by cloud computing, internet exchange, mobility and video streaming as well as the constraint of Moore's Law which is driving the industry to seek alternative ways to achieve performance improvements. In 2016, was an important year in the commercialization of silicon photonics and the turning point for market adoption as many companies made investments in the technology.

We want to give you a little color of this to frame out our positive perspective. In addition to early players in the market such as Intel, IBM, Cisco and HP, the interest of the companies focus on the development of silica photonics continues to grow. Sienna, Juniper Networks, both made acquisition in 2016 to bring silica photonics in house. Indicating their continued proliferation of this technology and its growing adoption.

Acacia Communications, a maker of coherent transceivers, based on a silica photonics IC will allow this transmission had a successful IPO in May and we are seeing a number of start-ups moving closer to launching products. From a product perspective, Intel, Mellanox and (inaudible) launched silica photonics devices during the last year. And Intel announced that Microsoft Azure will be an early adopter for its hyper scale computing environment and in Q4 Broadcom announced the Tomahawk ll Switch Series whose roadmap is expected to include a device with 25.6 terabytes per second capacity by year 2020.

This is likely to require optical interfaces to move data on and off the chip and across the board. And could represent an other inflection point for silica photonics technology. In total, all of these data points indicate that the indium phosphide technology is moving beyond the research phase to market adoption. Our own revenue from this application continues to grow and early indication from our customers suggests that they are planning for increased demand for 2017.

One of the positives in our business model is a long trajectory in the market applications. The data center upgrades, expansion is generally expected to be a multi-year of the cycle. As such we are ready to meet the volume and the specification requirement for this exciting market opportunity. AXT has been a leader in the development of high performance indium phosphide substrates for more than 20 years and we believe that our deep expertise in these demanding materials is allowing us to make an early market lead.

As indium phosphide remains an emerging material we expect to see some lumpiness in our sales of indium phosphide product in any give quarter but feel confident that indium phosphide will be a primary source of growth and opportunity in our business for the foreseeable future. Now let's turning to gallium arsenide. Gallium arsenide continues to provide a healthy base of profitable revenue for our business. As expected, sales of semi-insulated gallium arsenide showed some weakness in Q4 due short term to adjustment in our customers but we expect to see some improvement in Q1.

Sales of semi-conducting gallium arsenide were generally flat from the prior quarter, largely supported by demand from our traditional markets. In addition, we're watching with interest the continued market development of VCSEL for 3D sensing acquisitions such as gaming, mobile phones, smart TVs, high speed communications and high power material processors. Recent news indicate that key mobile devised manufacturers are likely to incorporate this technology in upcoming high end cell phone platforms.

Which would lead to numbers between 500 to 600 million units per year. This could make the technology a mid-term and long-term game changer driving a sizable increase in demand for high performance semi-conducting gallium arsenide substrates for the VCSEL applications. We believe that we will see meaningful revenue traction in 3D sensing applications beginning in late 2017 and are preparing our business today for increased demand, including R&D investments, capacity planning, and sales readiness.

3D sensing requires devices with highly precise functionality and consistent reliability. The competitive landscape of substrate suppliers that can meet the specification is limited. Therefore, we expect to be a player in this market and view this opportunity as yet the other exciting growth driver for our industry and for our business. Turning to our raw material business. It's still pretty tough out there. Although the pricing environment is improving modestly as we have seen in the past, a strong pricing environment can have a meaningful positive impact on our business.

While our joint venture in total remains a headwind our bottom line, we believe that each one of our joint ventures has the encouraging business opportunities in this coming year. In closing, this is exciting time for AXT. Our strategy to invest in the emerging market, to diversify our revenue across a variety of applications and to differentiate ourselves through deep expertise in some of the most challenging materials is succeeding.

In essence, I believe we're leading in the areas where the market is headed. In addition, our commitment to improving our manufacturing yields, operation efficiencies and strong financial discipline is allowing us to achieve some of the highest margins in many years. And to deliver profitability and positive cash flow. We're executing well across our organization, and we are positioned for continued growth in 2017.

I want to thank our customers, our shareholders, for their ongoing support of AXT and our employees worldwide for their ongoing dedication to our mission.

This concludes my prepared comments and will turn the call back to Gary.

Thank you, Morris. We believe that the revenue contribution in Q1 is likely to be similar to that in Q4. As such we believe that the total revenues will be in the range of $19.5 million to $20.5 million. In terms of our bottom line outlook for Q1, we're expecting to be in the range of $0.02 to $0.04 profit per share based on 33.7 million diluted common shares outstanding. Okay, this concludes our prepared comments and Morris and I will be glad to answer your questions.

Terrence, please take it over.

Thank you. (Operator Instructions). Our first question comes from Richard Shannon from Craig Hallum. Your line is open.

Hi, Morris and Gary. Thank you for taking my questions and graduations to a good ending to 2016. Let's see. A couple questions from me. First maybe on the guide ants for from me, Gary. You gave us a revenue range similar to what you just gave us, the EPS number that was a little bit lower. Should we infer that the gross margins might be a little lower and/or a little bit higher OpEx or how can you help us resolve those differences?

Yes, I think the gross -- we're very delighted with the gross margin in Q4. But it was one of those kind of nice quarters where the sun and the moon and the earth lined up. So I don't think the gross margins will be 37% in Q1. And I still think it's best for people running models to be in the low 30s until we have a few more quarters where we can sort of see how things even out.

Okay. I think you mentioned -- on the topic of gross margins, I'm curious if you could help us understand the mix versus the efficiency benefits there. It seems like raw materials were lower and I think pricing is still rather low there. Can you talk about the mix versus yield or efficiency improvements to help that?

Yes, let me take a first crack at it. I think it's very difficult to really differentiate it too unless we run very deep models. There are of course a number of things that we're doing. Improving yield is yet another one. Improving efficiency. I mean we reduced head counts in our total operation and we're winning more wafers through the line. And, obviously, some of the product mix do help us.

But for us to separate in what bucket it's a little bit more difficult. But I think in all categories, it's helping us. I mean lastly, I think the lower costs of raw material input which will align also is helping our gross margin as well. So, other than our raw material business, it's very tough but the substrate business is actually getting very good.

Okay. Fair enough. One or two last questions for me on your comments, Morris on 3D sensing. You talked about pickup a little later in 2017. It seems like we've been hearing from other companies who have a relationship with this space talking about a pickup potentially maybe before that especially since they're a little bit later in the manufacturing cycle here. Curious if you could lay out your comments there and why you would initially see something earlier than later in 2017, and any suggestion of the potential size here relative to your current business would be great to hear as well. Thanks.

Yes, Richard, we are excited about these opportunities and so far we already sending quite a bit of samples and sample quantities to our customers for evaluation as well as qualification. But we have seen some protections and projections but we haven't really started to run those numbers into our projection yet. So we would much rather be conservative in terms of projection and be ready for the R&D as well as capacity planning and R&D and the sales readiness for the market. But, to run the model this early, especially is a rapid ramping I believe if the market were to adopt this technology potentially the bottom can run up very fast. Right now we're just not.

Let me add something just to underline what Morris said. We're actually making great progress on low edge pit density yields. We've got some really smart scientists working on it and some manufacturing managers and leaders and it's been very encouraging. So we're not sort of sitting around sort of waiting for the bells to ring. We're being very proactive and we're making great progress. So we're ready and when the market rings our bell, we will be able to deliver. On the other hand, we're still waiting for that to get to that point and until we cross the line, we think it's better to be cautious.

Understood. Appreciate that detail. Last quick question for me, guys and I'll jump back in the queue. Curious do you have any update on the timeframe for the eventual move of your factory in Beijing?

Yes, we have been to a selection of site process and we think the move for our factory is we project is probably going to be end of 2018 or early 2019 event. And let me just add one more thing. We also plan -- we're watching this 3D fencing market development very carefully and it's exciting too. And if this market were to really evolve into a major event for the 3D sensing going forward, then a move could be a blessing for AXT because we are planning to build a factory to accommodate this growth.

We are acquiring enough land that we can expand our footprint over the years going forward if needed. And we'll build enough initially in the footprint so that the building is fitted out and we can add equipment and capacity quickly. And that's one of the strengths. We've proven than in the indium phosphide growth. We think it's a differentiator for us that we have an entrepreneurial ability to move fast to add equipment which can grow the capacity.

For instance the land we're looking at is approximately 80% larger than the land we presently own. And the factory because we're going to design in a very single step rather than we're accumulating all the old factory in our (inaudible) facility which we acquired over the last seven or 10 years which by definition is less efficient. So we expect to build out a very efficient factory with more capacity. Of course we will be prudent not to overbuild and then, -- but we want to be ready for this market opportunity if it were to really develop into reality.

The key is to get the permits and the infrastructure done in advance and then, which is not very expensive. And then we can move equipment in line as the market demands.

Okay. Thanks. That makes sense. I appreciate all that detail, guys. All the questions from me. I appreciate it. Talk to you later. Thanks.

Take care.

And our next question comes from Edwin Mok from Needham and Company. Your line is open.

Hi everyone. This is actually Arthur on for Edwin. Congrats on the great quarter and thanks for taking our questions. Maybe just to dive a little bit deeper into the 3D sensing market. Have you guys initially begun to size the opportunity? If it becomes adopted, what kind of opportunity can we expect and do you think a double digit growth rate can be achievable or is it too early stage to predict?

Well, I think the interesting thing is that right now it's zero. So any growth is going to be double digit, or infinite. But I think the market is very exciting. I mean I've seen some reports in saying it's supposed to be at least two chips per phone in some models and potentially three chips per phone on this VCSEL so two large VCSEL and one small one but some phones may only require two. But we believe the early stage will only adopt into the high end phones so I've seen some reports in saying per phone the VCSEL opportunities between $2 to $4.

But you've got to discount it back into the substrates and because there's outer layer which is [epi grower]. I think the bottom line is I think the gross potential or the market opportunity definitely is very substantial and great opportunity for all the gallium arsenide players especially for those who can meet those low EPD requirements. But how big it is, I mean it really depends upon how fast you think the adoption is going to be. And -- but smartphone is not the only applications. There are mobile devices which requires it. Also autonomous cars. That also requires.

So the way I envision this VCSEL business is it provides the machine vision to understand its environment to enter into its computer data processing capability to understand where everything are so that you can measure depth, height as well as volume. So I think it's a great opportunity to enable machine to have eyes so it can develop intelligence. So I think really how big the market is I think one needs to study a little bit more. I think the volume is definitely, it's a great, great opportunity.

Yes, I can give you two perspectives. Number one is I did some Beta testing because my boss asked me to, so I showed the Lenovo phone that has Tango in it to a couple of my adult sons and they like to do the gaming stuff. I think gaming will be a strong appeal for this application. The other perspective is this. The potential here, I mean there's no sense trying to limit the words, but the potential here is significant and since we are a small cap company, if the market accepts this application, it's going to move our needle for sure. So we just don't know how much yet and we don't know how soon and I think you guys will be probably more accurately at figuring it out than we can because you're talking to more data points.

Got it. Thanks for that color, guys. So in terms of VCSEL, can you talk a little bit about your competitive position with regard to VCSEL and if you could kind of share what type of share you have relative to your competitors?

Well, basically there are three legitimate players in this market other than us, there is a German substrate maker and there is a Japanese substrate maker. I think the other big requirement for this application is it requires, because of the VCSEL, it stands for lasers and lasers hate defects because a dislocation or defect underneath a laser will kill the device and since you're running a ray of lasers, up to 100 of these lasers per chip. So too much defect in the substrate will kill the device.

Such that the defect density requirement is a key barrier for entry. So I think the ways to achieve this consistent low defect density at an affordable cost and a big volume I believe now is the race. But since this market potentially is going to be so big, I believe it's going to benefit all the industry players between the three of us. And who's going to win? I mean we hope we are.

One of the things we've been promoting at AXT is that we know VGF since we are the first entry into the market was the VGF product. We know VGF crystal grows for 30 years and VGF is known to be able to reduce defect density in substrate. So, we think we have a leg up. But of course we do have a lot of respect for our competitors and the race has not even begin yet so we are all lining up at the starting gate. We hope we can win but if not, at least it will benefit the whole industry for sure.

Thanks for that color, Morris. Maybe just one last question for me. How much did indium phosphide grow in 2016 year-over-year? And based on the trends that you have outlined in your prepared remarks on silicon photonics and fiber to home and office, can we expect that this type of growth can be sustainable in 2017 and beyond?

Yes, so let me give you some color. I think we have said before in 2014 and 2015 we were growing more than 50% year-over-year for two years. And with the beginning of 2016 we started out pretty good and then sometime in June or even earlier, all of a sudden all the towns market seems to stop and we heard from our customers and our customer's customer, there seems to be an overbuild of inventory because in the two previous years people were just expecting too much, everybody double order or triple order.

And we believe they're in the process of weeding out the inventory and in fact I heard one of the market intelligence said, even one of the distributors in China bankrupted so they are dispensing out and using up the inventory. I think the important thing to remember is PON is a very necessary device to enable the fiber optics to be connected. Everywhere you need to communicate with indium phosphide light, PON is needed. So I don't think that market is going away.

It's just having an inventory adjustment. With that said, we did slow down in 2016 and we are in the low double digit growth for 2016 with the impact of the slow-down of the PONs market but we do believe that sometime in 2017, that should recover. With that said, I think also the silica photonics saw a very vibrant 2016. There are a number of our end customers announcing they have new product introduction and they are telling everybody the street and market they are preparing for market growth for 2017.

So we think that the silica photonics business will continue to grow for 2017 and sometime later this year, perhaps when the PON market, the inventory adjustment is over with, that will add on to the growth opportunity for 2017. So to answer your question, in our model, we do believe 2017 will give us a better gross percentage-wise than 2016 on indium phosphide.

Okay, thank you, guys. Congrats again on the great quarter.

Thank you.

Thanks.

And our next question comes from Tom Sepenzis from Northland. Your line is open.

Hey, congratulations on the quarter. I just want to follow up on if you were to win some of the VCSEL business in the second half of the year. Where would you be building that product? Would that be in the current fab?

Yes.

And what is the timing that you will complete the new fab?

We think the soonest we can probably finish the construction of the new factory is probably the first quarter of 2018.

And how long are you allowed to stay in the current facility?

We think we have a good set-up with local governments. We like to run the two parallel line with the old location as well as the new location for a period of time to allow our customer to qualify the new facility before we move on to the -- shut down the old facility and move on to the new facility. And don't forget, that's only required for gallium arsenide. Indium phosphide, actually, there's no pressure to move although in our own mind once when we move to the new facility, we think we may as well move indium phosphide as well.

Eventually in the future, yes.

So we think, Tom, that the facility is the first step of course. The overall move of the equipment will be done in stages so we'll never -- we're never going to turn everything off and then move it and then turn it back on. We'll move it, like, 25% at a time. So we'll never be down in terms of production. And we can do things like build inventory for key customers and things like that so that there's a safety net for them. We don't have an exact timeline yet but our discussions with the government agencies, we've talked about 2018, first half of 2019 and that's kind of what we're circling.

Okay. So I guess what I'm trying to get at is in terms of the volume, potential volume with the VCSEL win in the second half of this year, do you have the current capacity in the existing fab?

Yes.

And is that making conversations at all difficult with any potential customers?

Yes, we believe that we have looked at our yield and looked at our number of firms we should able to meet that demand. But of course I mean right now the numbers that we're seeing that we talk to our customers are only preliminary and we haven't really got any firm order. I mean once we got it, then we're going to make a hard calculation. And the other thing is I do believe that once the qualify us, they require us to deliver from this facility with the existing equipment for at least the first set of the product launch which potentially will last for two years.

So Tom, there's a silver lining sometimes in the ebb and flow of businesses. And as you may recall, AXT added quite a bit of capacity in the ramp of gallium arsenide in the 2009, 2010, 2011 stage and then when the silicon insulator cannibalized some of these sockets, we took some of that capacity down. And guess what? We still have the furnaces and stuff. So we think we're in good shape. That's not an issue at all.

Great, thank you. And then the thing we didn't talk about today because it's still a couple years out is the indium phosphide and 5G cellular power amplifier, potential market there. I was just hoping you might rank the opportunities as you see them, like the top three in terms of potential revenue growth over the next three years in terms of silicon photonics, 5G or VCSEL as what order do you think those come in in terms of potential growth?

That's a very good question. I think in my mind, I think this silicon photonics or indium phosphide, that's a very steady and has a great potential and should be a steady business, although it's very hard to gauge because silicon photonics was just introduced late last year and because it's coming from a very small volume base to begin with, one never knows how big it can grow. Although we've seen some number that's projected to grow 50% year-over-year for the next four or five years. I mean that comes from a very large big company projection. So, if that were to grow 55% year-over-year potentially can exceed the PON business opportunities.

Exactly.

I think that's a very solid business. And it's very difficult penetrating through that business. VCSEL I believe with this low EPD requirement, it's a great opportunity, especially it's a commercial product. So the volume could be just tremendous. I really think so. And it really -- it's our responsibility to capture how much of this opportunity that is presented to us and we're very anxious. But of course at this point, we don't want to build up expectation too much because you hear they're going to launch and then the next year you say, wow, maybe they're only going to launch with a limited version and maybe only 10% of the product will include that feature.

So, we think that eventually we will be here but just about the timing and volume. As far as 5G is going concerned using indium phosphide, I think definitely it's going to be three or four years away from the business. We are still develop our 6-inch indium phosphide capability and if that business were to come, that's going to overwhelm everything else. But since it's going to be four or five years from now, we're going to need to discount it very heavily. The expectation at least.

Yes, Tom, just in kind of a concise summary. We think the indium phosphide has a high barrier to entry. We see the optical network PON business going on and on even though it took a breather at the end of the year. And we think silicon photonics could exceed that product line in terms of that application. So indium phosphide is like gold for us. It's in the bank. We know how to do it. We're already doing it.

We could also do low EPD but at least for this year we're really expecting indium phosphide to continue to be the driver. In 2018, we think it could very well be a three-horse race. Silicon photonics, the PON business and the VCSEL business and in each one, we probably don't care which horse wins because we have all three horses.

That's great. Congratulations again. My last question, the VCSEL business being gas-based, would you expect gross margins to come back down if that was to equal indium in terms of revenue? Or does that have a higher gross margin than the traditional gas base?

We don't really talk about individual product gross margin. But so let me characterize this product. This EPD requirement so-called defect density, it's very stringent. The EPD, let me give you an example, it's very technical but EPD requirements for very good laser quality material is less than 500. This product by the end of figure out specification is less than 100. Okay?

Not only that, on top of it is the size of this crystal is actually six inches. The larger the crystal is, the more difficult it is to make this low defect density. The reason why I'm elaborating on all this is it's more difficult then the yield is lower for just about everybody. So the race among us is who can achieve higher yield in shorter time to get a higher gross margin?

I don't want to discuss the gross margin openly because I think some of our customers may be listening in and we have a very good gross margin, the lower the price. I think this is a challenge for the industry whoever can achieve this lower need this specification with good yield is going to benefit. And we believe we're well positioned for this race. Is that a good answer?

That's a great answer, thank you very much for the clarity. Appreciate it.

And our next question comes from Dave Kang from B Riley. Your line is open.

Yes, thank you. Good afternoon. First a couple of questions regarding your indium phosphide business. Is this still about 30% of revenues, total sales?

Yes. About there.

Around there? Okay. And then, with indium phosphide, you talked about silicon photonics or PON business. It sounds like PON is still bigger than silicon photonics so is it fair to Chinese PON is maybe half of indium phosphide and maybe silicon photonics is maybe 20% to 30%, something like that roughly?

Yes, Dave. It's very difficult for us to really identify what is what. We know some of the customers that we work with and they are using silicon photonics and we can identify. But some of the substrate we sell to, they don't tell us what they are using it for. Especially for those who we selling to the epi grower. The epi grower for their proprietary reasons, they don't like us to know. And the reason why we know some of the designated for PONs business is because, again, we work with some special specific customers.

They say well this is the Chinese market and it's mainly for PONs. We know that. But there's a lot of big brackets of it we say we don't know, okay. That could be either for detectors, avalanche photo detectors or PIN diode or it could be for lasers. And whether you're going to use the laser in what fashion, whether it's long distance or metro or for PON business, it's very difficult for us to tell.

Got it.

So to answer your question in general, we believe PON business is bigger because it's used everywhere and silica photonics is gaining traction.

Got it. And is it fair to assume that -- I assume indium phosphide is the biggest component, bigger than gallium arsenide or semi-insulating and semiconducting, is that correct? But then overall gallium arsenide is bigger than indium phosphide?

That's correct. Yes. Semi-insulating and semiconducting if you combine the two it's still bigger than indium phosphide.

Got it. And then Gary you talked about investing in R&D and CapEx, can you provide any quantitative flavor, how much we're talking about in terms of R&D and CapEx for 3D sensing.

I don't have it broken down in CapEx. We usually spend, we say $4 million to $6 million a year, but it's really been on the low side of that in 2016, it was below $4 million in CapEx. I would expect in the next couple years, we will be above $4 million but probably not above $6 million. There's nothing specific to say that, okay, we're buying CapEx equipment for low EPD. When we buy it furnaces for gallium arsenide, it's a furnace for gallium arsenide. We might tweak it a little bit but it's not like it's uniquely just you can only do low EPD in this furnace.

We're running a lot of experiment through R&D activity in the crystal growth area to improve our yields and to achieve those very stringent requirements. I think we are. But it's not showing up I think amazingly in our R&D budget but I think we have achieved quite a bit.

We've assigned more people to it but we didn't necessarily go hire ten people. We just took some really smart experts within the Company and turned them loose. So that's why it's hard to say, oh, we added ten people and they each cost $100,000 and that's a million dollars. It's not like that.

Got it. A couple more. Regarding maybe you talked about this already but I think I missed it. Regarding your first quarter revenue, although, can you just go over the assumptions of various components? Whether, something is going to, whether indium phosphide will be flat, gallium arsenide -- how should we expect gallium arsenide whether they're going to be up or down, any color regarding your opponents?

From my perspective, so I don't provide numbers but I've seen the budgets for the projection. We have sort of a set way of coming up with this these protections, okay. First of all, all visibility is really traditionally very short. Hopefully we're going to improve with demand of the product. A lot of our customer just pull the material potentially in the last minute.

But I think in Q1 projection, I think semi-insulating gallium arsenide is going to rebound from the fourth quarter because fourth quarter really, some of the customers are saying maybe they have the inventory so that inventory is over with and they're going to resume ordering. So semi-insulating is going to be up a little bit. Indium phosphide actually is going to be down a bit from the fourth quarter. I think semiconducting I think is sort of flat and germanium is having a good quarter because satellite activities is fairly strong.

That may be up a little bit.

Yes, that may be up a little bit.

And that's raw materials?

Raw material we think is going to be sort of flat and down slightly.

Okay. Got it. That's helpful. And last question then, this is just a theoretical question. Let's say a new phone with a 3D sensing will launch in September, then when do you see the orders for that? Theoretically speaking?

Well, theoretically speaking, if it were to launch in September, I think we should start to see some revenue in June because they need some time to build it.

Right. So about one quarter head start?

Yes, I think so.

That's probably realistic, yes.

Okay. That was it for me. Thank you.

Thank you Dave.

And our next question comes from Joe Maxa from Dougherty and Company. Your line is open.

Thank you. I just wanted to follow up on kind of the mix. I'm just thinking more on the raw materials side as far as looking year-over-year, it's been weak as you've stated. And do you expect that to continue to be weak? I'm just thinking about is it more of a volume or a pricing issue that's been weak year-over-year, and how does that look going forward?

Yes, Joe, I think raw material really see it beating in 2016. I think the first half of 2016 the price of gallium probably came down a good 30%, although it started to recover in the fourth quarter but we don't know how firm it's going to be going forward. So I think the raw gallium price, the 49 gallium price has really hit the bottom sometime in October or September and now it's probably 10% to 15% higher than the bottom low but the question is how far will it run or will it go back again?

I don't think it's going to go back but we just don't have no idea. Although some of the other business for our joint ventures we could give you a little bit more color. We have a joint venture which placed into the OLED market and that is going to see a very exciting growth although they are one of our joint ventures so their annual revenue is seen in the neighborhood of very small so although they see very good growth but in our total mix is not going to make -- move the needle so to speak.

And we have yet another joint venture which place into the fluorescent powder for LED lighting. That is also seeing very strong growth but again it's a very small business opportunity. So the main focus probably is on gallium. I think how far it will go from here on is anybody's guess.

But the decline in revenue is primarily due to.

The price.

Pricing. We did furlough one of the gallium factories for about six months in 2016. So that also took the revenue down a little bit.

And that's back up and running?

Partially. Yes.

Okay. And back on the operating expenses, I thought I heard you mention you were going to be perhaps ramping R&D and sales to prepare for the gallium arsenide or the 3D sensing opportunities. What should we be thinking about as far as your overall OpEx especially as we get into I suppose, Q2 in the back half of the year?

Well, if you look back over six or eight quarters, it's pretty flat, it's plus or minus a couple hundred thousand dollars. What I suggest people should think of is there's no big hockey stick but there are some key hires we would like to make this year and so I think it will inch up maybe $100,000 each quarter at most. But it's not going to get to $6 million. I can't see it getting -- it's not going to be a huge jump.

And again, we are -- when we say we're investing in sales readiness, it doesn't necessarily mean we're hiring more salespeople. It just means that we are lining up our ducks. We're making plans. We're making sure we're in front of the potential customers and things like that.

Also I would like to add a little bit of color on this so called R&D low defect density. We are making good progress but I think because we are running our factory and some of our production can be diverted into that R&D run so the material we produce actually are saleable material. So they're not necessarily getting into the R&D expansion bucket. So that's a great advantage for running a factory that we're running a sustainable business.

We have other six-inch substrate requirements so we're just running that at least the aim of developing lower EPD material and the existing customer doesn't mind either and of course by the way, if we achieve very low EPD, they belong to a different price category and we save it for the future.

And Morris, you mentioned expecting increased revenue in the silicon photonics markets in 2017. Any sense on the timing of that? Should we be thinking that's more of a back half or do you expect to see that continue in second quarter and then continue to wrap through the year?

I think it's going to be a continued ramp although if we compare year-over-year, this year's customer requirement is far greater than what we got last year in the first quarter. But however, this quarter of this year we're still being sort of putting a penalty box on the PONs business. So it's the compensation of the silicon photonics is still holding us up. But don't forget, PONs business back in early 2016, it was a greater major business. I think this other phenomena we also said, the industry said that the industry is in transition from the 2.5 G-PON to 10 G-PON.

So, I heard people are saying that they're reluctant to build 2.5G because the speed is lower so they were transitioning to the higher speed 10G-PON and when that business starts to run that should also help us. Because, as I understand it, the 10G-PON requires larger real estate to make the device work better and so larger real estate means more substrate. That means more revenue for us.

Yes. Okay, that's all I had. Thank you.

Thank you.

And at this time I'm showing no further questions.

Okay. Well, thank you for participating on our conference call. As always, please feel free to contact me or Gary Fischer directly if you would like to meet with us, we will look forward to speaking with you in the near future.

Ladies and gentlemen, thank you for your participation in today's conference. This concludes the program. You may now disconnect. Everyone have a great day.