Axcelis Technologies Inc (ACLS) 2017 Q1 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the Axcelis Technologies call to discuss the company's results for the first quarter of 2017. My name is Terrence, and I will be your coordinator for today. (Operator Instructions) As a reminder, this conference is being recorded for replay purposes.

I would now like to turn the presentation over to your host for today's conference, Mary Puma, President and CEO of Axcelis Technologies. Please proceed, ma'am.

Thank you, Terrence. With me today is Kevin Brewer, Executive Vice President and CFO; and Doug Lawson, Executive Vice President of Corporate Marketing and Strategy. If you've not seen a copy of our press release issued earlier today, it is available on our website. Playback service will also be available on our website as described in our press release.

Please note that comments made today about our expectations for future revenues, profits and other results are forward-looking statements under the SEC's safe harbor provision. These forward-looking statements are based on management's current expectations and are subject to the risks inherent in our business. These risks are described in detail in our Form 10-K annual report and other SEC filings, which we urge you to review.

Our actual results may differ materially from our current expectations, and we do not assume any obligation to update these forward-looking statements.

Today, Axcelis reported first quarter 2017 financial results. Revenues for the quarter were $86.9 million with gross margins at 40% and earnings per share of $0.29. Our systems mix in the quarter was 67% mature foundry logic and 33% memory. Both market segments continued to be strong, driven by data storage applications and the Internet of Things. The geographic mix of our system shipments was Korea 43%, China 37%, U.S. and Europe 13% and the rest of the world 7%.

Chinese market activity is currently driven by mature technologies supporting IoT devices. We expect memory activity in China to begin in 2018. Purion products continued to gain momentum during the quarter. In Q1, Purion accounted for 93% of our systems revenue. We placed Purion products in 7 new fabs. This included 3 customers who bought Purion products for the first time and 4 Purion systems that were first in fab tools at existing customers.

Also, today, we shipped the 100 Purion system, a significant milestone for Axcelis. Our business outlook remains positive. In the second quarter, we are forecasting revenues of approximately $100 million, gross margins of approximately 38%, operating income of $12 million to $13 million, and EPS of $0.30 to $0.35. Our total cash should increase significantly to approximately $100 million.

In the second quarter, we expect a systems mix of approximately 40% mature foundry and logic and 60% memory. For the full year, we expect the system segment split to be weighted toward the mature process technology market. In 2017, we still expect our market share to increase to between 22% and 28%, driving total company revenues within a range of $320 million and $380 million. We expect to provide an update to this range in July at SEMICON West.

As we discussed in our last earnings call, we plan to invest more aggressively to grow revenues, to capitalize on strong market conditions. This will include incremental investment in R&D to develop additional Purion product extensions targeted at specific customer requirements. We will also selectively invest in SG&A to address our growing customer base, especially in China, and to expand our reach into leading edge logic customers and the Japanese market. Demo activity at our Advanced Technology Center in Beverly, Massachusetts is very strong, with both new and existing customers investigating emerging applications and novel products.

We expect this increased demo activity will lead to additional customer evaluations and further penetrations. This activity is key to achieving our long-term target business model. The 3 key assumptions in our model are: a strong foundation built on continued strength in the mature process technology segment and our Service business; a robust memory cycle, driven by sustained NAND activity and cyclical strength in the DRAM market; and growth from new opportunities, including the Chinese memory market, penetrations into leading edge logic and reentry into the Japanese market.

Before Kevin reviews the details of our first quarter financials, I would like to reiterate our 4 key objectives for 2017. First, we will focus on new customer penetrations with Purion. The expanding number of customers investing in IoT, 3D NAND in China has broadened the CapEx opportunity, providing new prospects for Axcelis. Second, we will also drive to grow our share of business at existing customers as they increase their capacity in 2017; third, we will focus our marketing and engineering resources on enhancing Purion and on developing additional product extensions to drive revenue. And fourth, we will execute against detailed gross margin improvement programs.

Now I'd like to turn it over to Kevin to discuss our financials.

Thank you, Mary. In Q1, we delivered solid financial results, driven by our higher system sales and strong gross margin performance of 40%. Gross margin in the quarter was fueled by a favorable mix of Purion XE, Purion product extensions, and a higher percentage of revenue from the mature process technology segment. System margin improvement continues to be driven by a strong focus on Purion value engineering, supply chain optimization and factory-lean programs.

Since volume ramp of a full Purion product line in Q1 of 2015, system-standard margins have increased 1,240 basis points on a rolling 4 quarter average. Improving Purion maturity, platform commonality and product line extensions are also contributing to higher gross margins. In Q2, we expect gross margins to be approximately 38% based on product and segment mix. As we have discussed many times, gross margins can vary from quarter to quarter as a result of product and service mix, segment mix and a timing of the evaluation to a recognition. With a strong diversified market anticipated in 2017, we've seen opportunity to firmly establish Purion across a broader customer base. To ensure we capture these opportunities, we expect to continue to support new evaluation tools for our Purion product line.

We're also investing and updating some of our core business systems and adding headcount and infrastructure to capture the fast-growing Chinese market. China represents a significant opportunity for us, especially in our core markets of memory and mature process technologies. In fact, 3 of our 7 new customer placements this quarter were in China.

Finally, we are developing additional Purion product line extensions, which we expect will capture new opportunities and help drive gross margins. As a result of these planned investments and increased variable compensation, we expect our combined SG&A and R&D to be approximately $26 million in Q2, which I believe will be our peak spending level for the year. I expect full year spending to average approximately $25 million per quarter.

At this point, we are aggressively pursuing revenue growth during the strong period of expansion in the semiconductor industry. I believe the incremental spending is needed to capture opportunities we have targeted. That being said, I will continue to tightly manage expenses as I had in the past to ensure they are aligned with revenue growth and gross margin improvement and objectives.

Now I'll review the first quarter financial results. In Q1, we finished above guidance and above consensus estimates across the board. Q1 revenue finished $86.9 million compared to $69.4 million in Q4. Q1 system sales were $55.3 million compared to $37.7 million in Q4. Q1 CS&I revenue finished at $31.6 million compared to $31.7 million in Q4.

Q1 sales to our top 10 customers accounted for about 85% of our total sales compared to 80% in Q4, with 4 of these customers at 10% or above. Q1 system bookings were $97.2 million compared to $34.8 million in Q4, with a Q1 book-to-bill ratio of 1.71 versus 0.88 in Q4.

Backlog in Q1 finished at $48.2 million compared to $17.5 million in Q4. Q1 combined SG&A and R&D spending was $24 million compared to $21.1 million in Q4. SG&A in the quarter was $14.1 million, with R&D at $9.9 million. In the second quarter, we expect SG&A and R&D spending to be around $26 million, which includes a previously discussed investment and variable compensation expenses. Gross margin in Q1 finished at 40% compared to 38.9% in Q4. In Q2, we expect gross margins to be approximately 38%, down slightly, due to product and segment mix.

Operating profit in Q1 was $10.7 million compared to $5.9 million in Q4. Q1 net income was $9.5 million or $0.29 per share compared to $4 million or $0.13 per share in Q4. Q1 inventory ended at $115.6 million compared to $113.9 million in Q4. Inventory turns, excluding evaluations, finished at 2.1 compared to 1.6 in Q4. Q1 accounts payable were $26.4 million compared to $25 million in Q4. Q1 receivables was $67.1 million compared to $50.6 million in Q4. Q1 total cash finished at $76.2 million compared to $77.7 million in Q4.

We expect Q2 total cash to finish significantly higher at approximately $100 million. Overall, I'm very pleased with our financial performance in Q1. Purion sales drove significantly higher systems revenue and strong gross margin performance provided solid operating profits.

We continue to execute to our cost out roadmaps, which is moving system gross margins higher, and we continued to penetrate new accounts of Purion products as shown by the 7 new customers we gained in Q1. And finally, in case I need to say it again, I will be closely watching our expenses and looking for ways to cut cost. But now is the time to take advantage of the strong market, and invest a little more to capture new customers and gain market share.

I will now turn the call back to Mary for her closing comments.

Thank you, Kevin. The industry is in a period of unusually strong growth, driven by the growing IoT and data storage markets. During this period, Axcelis' revenue growth will be the result of broader customer base, increased spending in both the memory and mature process technology segment, and penetration of the leading-edge logic segment and Japanese market. Our gross margin improvement initiatives remain on track, and our costs are controlled and aligned with our revenue growth objectives.

Axcelis has provided our customers with a strong alternative in ion implant. And as they have come to know Purion, they are choosing Axcelis. As I highlighted last quarter, they choose Axcelis for several reasons.

Advanced technology is first on the list. The Purion product family provides them significant advantages in angle dose and energy control, derived from our scanned spot beam technology. The common Purion end station and ELS4 Source provide them cost of ownership advantages.

Second, our focus on customer-specific applications provides them with implant solutions tailored for their needs.

Finally, there is the innovation benefit, which by promoting competition, allows customers to gain significant technology advances and engineering expertise.

Customers measure the innovation benefit through improvements in productivity, yield and cost of ownership. They maximize this benefit by splitting their business evenly between 2 strong implant suppliers. We believe that with Purion, we are poised to take advantage of this strong market and deliver on our target business model, yielding attractive earnings for our shareholders.

With that, I'd like to open it up for questions.

(Operator Instructions) And our first question comes from Craig Ellis from B. Riley Financials.

Congratulations on the very good execution to start 2017. The first is just a clarification on the first quarter's results. The first quarter revenues were about $7 million or 8.9% above what was expected about 3 months ago. Can you help us understand where the business wound up being stronger, versus early year expectations?

So what we found as we moved through the quarter is that there was increased or increasing strength from both the memory market as well as the mature process technology segment. In particular, we saw some additional strength coming from DRAM as certain customers transitioned through a technology shrink. So those are mainly the 2 areas.

And was that multi-customers? Or just a single customer on DRAM?

It was a single customer.

Okay. And then, switching to the longer-term dynamics with revenue. Thanks for giving us a view on the full year. I had a couple of questions related to that. I'll try and tie it all together. So at a $320 million to $380 million range for the year from the second quarter's guidance for $100 million, the low end of the range would be down 22% half on half, the high end of the range would be plus 11%. The midpoint's down about 5%. Can you help us understand the factors that would either result in a low end of range outcome for Axcelis, or a high end of range outcome for Axcelis?

So at this point in time, we're only guiding one quarter, given the fact that our customers continuously make changes in the timing of their investments. That's -- and we did say that in 2017 our revenues were going to be between those goalposts of $320 million and $380 million. At this point in time that's really as much color as we're going to give on guidance. But we will update the revenue range in July in SEMICON West. That said, we feel very good about the dynamics of the market. And we believe the semiconductor industry is in an unusual period of growth. NAND continues to be strong, based on end-use market drivers, and those drivers would imply that the market will remain strong into next year. DRAM pricing is currently holding, which is a good sign. But the spending will be metered based on the timing of shrinks and competitive pressures along the DRAM makers to invest. And then, IoT investment is also robust around the world, particularly in China and Europe. And again, we expect that to continue into next year. So at this point, the market dynamics seem to continue to be strong. But again, given where we are today, we're just not going to give any additional guidance out through 2017.

Fair enough. I'll take my last one with Kevin, and then I'll jump back in. Kevin, can you just clarify, were there any evaluation tools rev rec-ed in 1Q? And what's the assumption in 2Q's 38% gross margin?

So there were not any tools recognized in Q1. What we're expecting, in the second quarter, Craig, is that if you look at the mix of both the products and kind of the segment, that's where we're seeing a little bit of change in gross margins. In the investor pitch, we have that slide that highlights the differences, really, between the 3 products, between the HCM and EXE. And it's really just movement in that $100 million revenue range in terms of the products that are filling in there. So that's why we're down slightly. As we've been talking about full year, margins are continuing to improve. We've got a lot of initiatives in place, we're making good progress on them. So I think the key takeaway is we're continuing to lower the cost of our products. So we're going to continue to improve the margins across each product. And the thing that will move us from quarter to quarter will be the mix of those products, and potentially, whether does an eval recognize or not in a quarter.

Okay. That's helpful. But just one follow up to that though. There's no doubt the company is making very good progress on a number of underlying drivers to gross margin. But I think on last call, the company noted that there would be potentially 3 or 4 evaluation tools rev rec-ed this year. So does the first half commentary mean that, there's 3 or 4 tools to rev rec in the second half? Or is there a different view at this time?

Yes. So we have -- Craig, we have several eval tools out there right now. And we will have additional tools going out. Some of these will close out before the new tools go out, but we actually are going to have some additional tools out there, increasing the total numbers before things start closing out. Most of those tools are planned in the Q2 through Q4 time period. So they're kind of scattered throughout the year, how it is right now.

And our next question comes from Edwin Mok from Needham & Company.

This is actually Arthur on for Edwin. Mary, maybe just to kick things off with you. I know you're not ready to provide second half commentary, but just based upon what you're seeing now versus 3 months ago. How would you characterize any shift in your view on the second half? Are you seeing better sentiment around memory? Is mature logic looking a little bit better? Any color there would be helpful.

While I don't know that we've necessarily seen any significant change in the year over the last several months. I did mention that we saw an uptick in DRAM. But we all know that in the DRAM -- DRAM tends to be very cyclical, and it's just really a function of when some of our customers are either going to shrink, or as I said before, for competitive reasons, decide that they need to make investments to keep up with their friends so to speak. So that's something that we're just going to have to watch very closely. For us, though, there really hasn't been any change in what we see in either the NAND market or the Internet of Things. And I think even at our last call, we were very, I'll use the word bullish, on the mature process technology segment. It's actually, as we said, accounted for a significant portion of our systems revenue, although it was less than 50%. In the first quarter, it will shift to be a stronger percentage of our systems revenue as we move throughout the year. So we're still feeling very good about things, and we're still very confident about the dynamics that we're seeing in the market. As I said, we're just really not positioned to give any guidance any further out than Q2.

No. I understand. And then, I think, in the past, you've talked about material modification as an area of opportunity for Axcelis to penetrate leading-edge foundry logic. Can you provide an update on any progress in this effort? Or when we can start to think about this as a driver for Axcelis business?

Well, Arthur, this is Doug. We've got a lot of activity, as Mary mentioned in the script, in the demo area. A lot of that is with leading logic customers in general. Leading-edge logic customers are focused on material modification. So again, the Purion H and the capabilities of the spot beam and several other capabilities that the tool has, are the things that are driving the demo activity. The fact that the tool now is installed in a lot of places running both memory and trailing as logic, is giving them much more confidence to be able to come in and look at the tool for production. So as we've said in the past, we think, we will penetrate leading-edge logic over the next 12 to 18 months, and somewhere in the 7 to 5 nanometer node. And it's likely to be on a material modification application.

Got it. Thanks for that color. And then, Kevin, maybe moving on to you. You had previously outlined a target of achieving 4% gross margin by the end of 2017, and I know you provided some color to a previous question on the put and takes for gross margin. But given that you've delivered that goal, essentially, on 1Q and you think about the progress you made on driving lower cost. Is it fair to say that there could be a potential for upside to that 40% target?

Yes. So I'm at -- at this point in time, I think all I can say is that we're going to continue to make progress. And depending how the mix lands from quarter-to-quarter, I think that's going to be -- that's going to really determine what number is in front of this. You can see the impact in Q2 goes down a little bit. It's not because our cost has increased it's really because of the mix. So it's a little difficult to predict what Q4 is going to be at this point in time. And we certainly have a mix in front of us, but a lot's going to move. So again, I think what's important is we're continuing to drive lower cost, and we're making a lot of gains with the initiatives we have in place.

And the target business models haven't changed.

Yes. I mean -- yes. So I mean, that's a good point. I mean, target business models hasn't changed. In the investor presentation that's going up on the web today, and it might be already up. We've updated our models. So we've got some gross margin ranges in there, along with operating expense ranges versus revenue ranges. So we haven't changed what we're doing for this year, I guess, what I would say.

And our next question come from Christian Schwab from Craig-Hallum.

Great quarter guys. Great guide. I haven't had the chance, so I just pulled up here quick on the website. But are you still thinking that kind of in the 1.5, 2-year time frame that we can drive $460-plus million of business and $2.80 in earnings? Is that still our more longer-term objective?

That's -- the models that we have in our investor presentation are models that we believe, we can achieve at the minimum. They are certainly target models that we are targeting for. So we definitely believe they're within our sights.

So Christian, we've updated the model this quarter, and basically, we've given a little more visibility looking out longer. So now you'll see a column that's for $450 million, and one for $550 million in revenue, and the appropriate gross margins and percent OpEx and so forth.

All right. Excellent. Thanks for the -- for added clarity. I think the rest of the questions I had have already been answered.

And our next question comes from Brian Chin from Stifel.

Mary, first, I'd be a little surprised that the implant market is growing 20% to 40% year-on-year. So it would be helpful if you could discuss what you currently think the baseline growth rate for the overall ion implant market will be in 2017?

Yes. Brian, this is Doug. I'm not sure where you're getting the number of the implant market growing. The implant market this year is up from around $850 million last year to $950 million. We expect to grow, in 2018, to around $1 billion. And so the growth is coming from a couple of places. One, is new wafer starts in 3D NAND is certainly driving it, and the activity in China. So in addition to the normal flow that you will get as companies are expanding their capacity, we've got a lot of new customers. And then, the IoT is driving implants, a lot of upgrades of existing older implanters in those fabs. And so that's really where the growth comes. But it's still -- it's a $1 billion-or-so plus or minus market.

Okay. Great. Yes, I was just using the math that Mary provided for the range of revenues this year. And saying that, obviously, that you'd have to chalk the difference up to market share gains.

Oh. Yes -- no. The market -- we are -- yes, it's all market share gains. Yes.

Yes.

Great. And could you also provide any rough parameters on what the DRAM versus NAND? What was DRAM versus NAND driven in the quarter? And how you might expect this to trend in Q2 and into the back half of the year?

So if you take a look at the technology breakout for our shipments in Q1, memory accounted for about 33% and it was almost evenly split between DRAM and Flash. It was 16% DRAM, 18% or 17% Flash. So that was the breakout. And then, if I look at our bookings, although, again, we've always talked about how our bookings tend to be the same book and ship in the same quarter. Again, very evenly split in terms of what we booked in Q1, it was about 21% -- it was 21% DRAM and 21% Flash. So I think that just gives you some sense that, as I said before, the DRAM was actually much stronger and kind of caught up with our expectations for NAND during the quarter.

And Brian, I guess, as far as the second half goes, we can't give anything more than the break down that we gave, which is that we think the mature technologies will be a little bit ahead of the memory technologies in terms of the segment mix for the full year. So we see the, the first quarter and the second quarter are almost inverse of each other. And then -- but it ends up pretty balanced at the end.

Okay. That's helpful. Maybe playing off of that for a moment. You've talked in the past about implant applications for things like image sensors, silicon carbide power devices as nice growth opportunities for Purion EXE and Purion M. Both of these markets seem to be on pretty solid footing. Can you discuss how material these revenues are currently? And -- or could be later in the year?

Well, both of those fall into our mature markets -- our mature process technology categories, which typically have been roughly half of our revenues the last couple of years, and basically, will be again this year. The image censor piece of that is reasonably significant in that almost every foundry in the world wants to be able to have the capability to build image sensors, and that requires high energy tools. The power device market is a little bit more of a specialty market. There's a smaller number of players, but it is a really nice -- good margin opportunity for us in the Purion M area.

And our next question comes from David Duley from Steelhead Securities.

I have a couple. When you look at your sequential guidance into Q2, I think you got, like, roughly $14 million of an improvement in revenue. Could you help us understand and you've kind of already answered, but help us understand of that $14 million, where a majority of that incremental revenue is headed? I guess it's probably towards the memory market from what your characterizations were. But also curious if you can, like, talk about on the product front, if it's Purion M and Hs? Or if it's the XEs and the other product line extensions that are accounting for a majority of the incremental revenue growth?

Yes. So I mean, we're not going to give you exact numbers, but there continues to be both the mature process technology as well as memory spending into Q2. And it's really a mix of all 3 types of Purion products that we're shipping. We've had a lot of success recently with the Purion M, particularly for some specialty applications with hot silicon carbide, the high-energy market has been extremely hot, not only for memory, but also for image sensors. And then, the Purion H is actually going into both markets that we serve as well, memory and mature process technology. So right now what we're seeing is we're seeing strong markets across market segments and geographies, Dave.

Just let me follow up before you give a little bit more color there. Clearly, the biggest part of the market is the Purion H, right? So in order to take a lot of market share, I would imagine, you have to see incremental shipments in that segment of the market. Maybe I'm wrong. But maybe, just talk about, are you achieving -- do you think you're achieving your goals in the segments of the markets as you thought you would as far -- in achieving your market share gains?

Yes. Dave, we are. This quarter we landed Purion in 7 new fabs, 3 of those brand-new customers. It's a mix of Purion XEs and Purion Hs. It's a mix of memory and mature technologies. We absolutely are meeting the objectives of getting those tools out there and into new customers. The Purion XE, in the high-energy market has been very strong, driven a lot by the image sensor requirements, and by the custom logic demands with the foundries and some of the IDMs have. So that is, that has been a strong performer for us.

Okay. And as far as the memory business goes, I think you talked about how there was a lot of 18 nanometer conversion activity from, I think, one customer or reference one customer in the DRAM space. Do you expect that strength -- how long would -- I think you started off the year saying that the strength would be 6 months in duration? Do you think that's going to last longer now as many of your -- or maybe just give us a view on how you think the sustainability of the DRAM segment of the business is?

Yes. So Dave, we definitely see strength. Pricing has held up. And as long as pricing holds up, then there will be continued investment. One of the things that we've noted is that DRAM spending is being metered. That's probably the best way to look at it. The same 3 guys that are doing DRAM are trying to spend as much as they can in NAND because there's such a growing market there. So that in itself is helping control the spend, but there is spend there. And as Mary said, half our -- about half of our revenue this quarter was from DRAM -- or of our memory revenue was from DRAM. And we see it being -- continuing to be strong as long as pricing holds up, which based on everything the memory suppliers have said, and their customers have said, looks like it goes through, at least, the third quarter, maybe a little longer.

Okay. So that's helpful. The DRAM spending looks like it might pass or be strong into the -- beyond the second quarter, into the third quarter. And just a final thing on that topic. Do you -- in your thinking about the DRAM business, have you started to include new factories in your thinking yet? Or is that an '18 event?

No. It's definitely -- I would call it late 2017, early 2018 event. So some of it could start this year.

Okay. And final thing from me, for Kevin. Just looking at the press release, I noticed that you when you gave the service revenue and gross, and cost of goods sold, the service margins were negative. So could you talk -- and yet you achieved spectacular margins on your products and still hit your 40% target. So maybe just help me understand why the service gross margin was negative during the quarter?

Yes. So I'll start off by saying our service business, as a whole, if it includes spare parts upgrades, used tools and stuff. It's very accretive overall for the business, and we always say that. Unfortunately, the accounting side of things. When you see service in the breakout in the press release, Dave, that's just the labor piece. It typically does not -- it's not typically a negative event, but it's not a high margin part of the business. And in the particular quarter, in Q1, it really had to do with some new contracts we put in place that were just a little bit more heavily loaded with cost up front, which you kind of take the revenue spread over the term of the contract. But it's a very small revenue number that's associated with that negative margin there.

And our next question comes from Mark Miller from Benchmark.

You talked a little about, in the prior question, how sales and backlog broke out in the first quarter between NAND and DRAM. From what I've heard today, it sounds like things will remain strong, but it sounds like the breakout in terms of orders later in the year will shift more towards NAND. Is that a false assumption?

We haven't really been specific on that. But right now, from a general market perspective, we see the NAND market strong for at least the next year to 2 years. There will be periods where it's real strong and periods where there might be gaps between projects. The DRAM market is going to cycle based on pricing. And the investment that goes when pricing is -- it goes up when pricing is high. So if DRAM pricing stays high through the end of the year, then DRAM would probably contribute more. If DRAM pricing falls in the later part of the year, then NAND would probably contribute more. But it's really too hard to come up with an exact number on that yet, Mark.

Okay. Can you break out by your 3 principal tools first quarter sales and backlog? What percentage for each of the 3 major tools?

No. We don't provide that, Mark.

Okay. Was -- the mix was stronger this quarter than expected for next quarter?

The mix...

We had a -- we did say, in Q1, Mark, we had a strong mix of Purion XE.

Yes. A lot of high energy.

But we did have, in Q1, we had a strong mix of high energy.

(Operator Instructions) And we have a follow up question from Craig Ellis from B. Riley.

I wanted to start, just going back to your comments on geographic dynamics, Mary. That was very helpful. Our research -- and I think it's fairly well known that there's 19 fabs or so that are under consideration, many of them under construction, many in advanced stages. A vast majority of those are mature foundry fabs, some are memory. I think you mentioned that you saw some potential for memory in China next year. What would be helpful is if you could provide some color on how you think Axcelis will fare from a share standpoint in the mature foundry area? And then, secondly, when you look to memory next year, in China, do you see potential for DRAM and NAND? Or is it one or the other? And to what extent is it pilot lines versus production et cetera? Of course, not asking for any customer-specifics, just for what you see in the market right now.

Okay. So a lot of what we are seeing in China right now has to do with mature process technology versus memory. And we have a very strong position in those fabs. Most of those fabs are owned by customers who we've done business with in the past. That said, we actually are penetrating with Purion products. They had bought our legacy tools in the past. And we expect that business to remain strong through at least this year and into next year. Starting in 2018 as you referenced, Craig, there are a number of projects that are on the drawing boards right now. Some, they've actually broken ground and are -- have -- there's construction and there are buildings. We expect that to start to ramp up in 2018. We think it will be a mix of DRAM and NAND. Always a question about where is the technology going to come from, but I think a number of those players are actually working through that. And we should hear -- there should be some announcements at some point over the second half of the year that will clarify some of that. We do very well in the memory market. Many of those customers use some of our existing memory customers as reference sights. It's very well known where Axcelis has its strength and which product. And so as that market picks up, we fully expect to get our fair share of that business in China in memory and really add that to what we're already doing on the mature process technology side of the business.

Excellent color. The next 2 are really related to the target model. They may be for Doug or Kevin. First, at the high end, the progression from $450 million to $460 million in revenue, up to $550 million. What accounts for that new higher range of revenue? Is it the view that there's greater share to be had out there for Axcelis? Or that the market is bigger? Or is there assumption that there may be new product areas that the company could pursue?

It's primarily share. So last year, we gained 14 new customers for Purion. In the first quarter alone we gained 7 new ones. So we feel we've got good opportunity to continue to gain the share. And so the 4 -- what was the old model at $460 million, which really assumed around a 35% market share. We think we can get higher than that. And hence, we're showing the model that shows what it would look like when we get there.

Excellent. And then, the second question, and this one's a bit trickier because the model parameter's down and the middle of the income statement go from dollars to percentages. But it looks like the old OpEx assumption was around 20%. The new one looks like it's closer to mid-20s. Does that reflect some of the investments that the company was talking about on this call? Or what accounts for the change in the operating expense assumption model to model?

It is the investments we are talking about, Craig. We're adding some infrastructure to support China. We're adding additional R&D funding on these product extensions that develop some new offerings. And there's some other things with the variable comp side that's changed. But it really is all the things we highlighted before driving this a little bit higher. And when you get out to the $550 million model, I think, if you look at the range, it's actually a very reasonable range to be in. And right now, I guess, the other thing I'd add to there are -- there's a lot of evals going on right now. We're kind of treading along new customers, but with that comes a cost. So the good news is we're fanning out, but there is a lot of evals out there, and there is several more planned for this year. And those are going to spill into '18 for sure. So it's probably one area where we might have been a little bit more optimistic when we've put out that model together a few years ago about how much eval expense we would have at this point in the revenue curve. But the reality of it is, we've got to do this to grow the business.

That's a nice problem to have.

Yes.

The other thing, Craig -- Craig, I just want to add, sort of a conclusion on that is, is we're in this unprecedented market where there is just great opportunity across all the segments, across all the geographies, and if there is a time to invest and to spend to make sure we can grow our market share, now is the time to do it. And that's one of the other reasons why we're able to add that other column on to the target business model.

The last question on that topic then, from where we are today, which is 27.5% of sales of the model that you have longer term would be 200 basis points below that. Does the business need to invest significantly so that OpEx as a percent of sales would rise materially before we get to a lower level down the road? Or do you think you can manage growth in your operating expense profile so that we're moving more steadily on a downward trajectory towards that longer-term target that you have?

Yes. Well, I may certainly think this year the Q2 number, $26 million, is a peak and full year will be probably closer to an average of $25 million a quarter. So it's moving to the $450 million range. I think a lot of things are in place, and a lot of things working on right now are pretty much there. So some of that becomes more of the variable piece of the equation, rather than heavy investment in any particular thing.

And this concludes the Q&A portion of the call. I will now turn the call back to Mary Puma, who will make a few closing remarks.

Well, we're pleased with our first quarter performance and excited about the prospects of a strong year for Axcelis and the industry. As always, I want to thank you for your continued support and hope to see you in the coming months.

We will be attending several conferences, including the 18th Annual B. Riley Institutional Investor Conference in Santa Monica, California on May 24; the 14th Annual Craig-Hallum Institutional Investor Conference on May 31 in Minneapolis, Minnesota; and the 2017 Stifel Technology Internet and Media Conference on June 6 in San Francisco, California. Thank you, again.

This concludes the presentation. Thank you for your participation in today's conference. You may now disconnect. Good day.