GSI Technology Inc. (NASDAQ:GSIT)

Q1 2018 Earnings Conference Call

July 27, 2017 4:30 PM ET

Executives

Lee-Lean Shu Chairman, President and Chief Executive Officer

Douglas Schirle Chief Financial Officer

Didier Lasserre Vice President-Sales

Analysts

Kurt Caramanidis Carl M. Hennig, Incorporated

Jeff Bernstein Cowen

Operator

Ladies and gentlemen, thank you for standing by. Welcome to the GSI Technology's First Quarter Fiscal 2018 Results Conference. [Operator Instructions]

Before we begin today's call, the company has requested that I read the following Safe Harbor statement. The matters discussed in this conference call may include forward-looking statements regarding future events and the future performance of GSI Technology that involve risks and uncertainties that could cause actual results to differ materially from those anticipated. These risks and uncertainties are described in the company's Form 10-K filed with the Securities and Exchange Commission.

Additionally, I have also been asked to advise you that this conference call is being recorded today, July 27, 2017, at the request of GSI Technology.

Hosting the call today is Lee-Lean Shu, the Company's Chairman, President and Chief Executive Officer. With him are Douglas Schirle, Chief Financial Officer; and Didier Lasserre, Vice President of Sales.

I would now like to turn the conference over to Mr. Shu. Please go ahead, sir.

Lee-Lean Shu

Good afternoon, everyone, and thank you for joining us. Today, we reported first quarter net revenue of $10.7 million and a gross margin of 52.4%. Both were below the range of guidance that we provided earlier in the quarter, in part due to a mix of lower-margin product deliveries to one of our largest customers and additional inventory reserve for assets and obsolete products that were greater than our normal quarterly average. We expect a return to a higher-margin mix of sales to these customers in our second fiscal quarter.

Overall, during the first quarter we continued to see slowness in our primary telecommunication and networking markets along with continued weak sales in Asia.

However, beyond our traditional market of high-speed SRAM and low-latency DRAM, we are seeing high interest in our newest market segment, our extremely high-performance SigmaQuad radiation-hardened SRAM products targeted at aerospace and the defense applications. We are also receiving inquiries of our patent in-place associative computing technology and the intellectual property that is under development, which focuses on large emerging markets such as big data applications, computer vision and cybersecurity.

I would like to touch on our associative technology for a moment to provide some background for some of our newer stakeholders. We obtained this technology through our acquisition of MikaMonu in November 2015. This technology is for application in evolving new markets such as big data including machine learning and big combination of neural networks, computer vision and the cybersecurity. We currently hold 15 U.S. patents and a number of pending patents for our in-place associative computing technology. The technology changes the concept of computing from serial data processing where data is moved back and forth from the processor to the memory to parallel data processing, computation and the search directly in the main processing array.

While still in development, we are currently marketing new product sets based upon this technology to a variety of users and we believe we will have an effective base of potential customers once we bring this new product to market.

We remain on schedule to complete the design of our initial product by the end of calendar 2017. This technology will readily improve the computation and the response times they have currently and will have continue to be required in a variety of big data application. We are very excited about our business and the quarters ahead.

Douglas Schirle

We reported a net loss of $1.5 million or $0.07 per diluted share on net revenues of $10.7 million for the first quarter of fiscal 2018 compared to net income of $260,000 or $0.01 per diluted share on net revenues of $12.9 million in the first quarter of fiscal 2017 and a net loss of $1.3 million or $0.07 per diluted share on net revenues of $10.4 million in the fourth quarter of fiscal 2017.

Gross margin was 52.4% compared to 51.9% in the prior year period and 56.4% from the preceding fourth quarter. First quarter fiscal 2018 operating loss was $1.5 million compared to operating loss of $1.5 million in the prior quarter and operating income of $389,000 a year ago.

Total operating expenses in the first quarter of fiscal 2018 were $7.1 million compared to $6.3 million in the first quarter of fiscal 2017 and $7.4 million in the preceding fourth quarter.

Research and development expenses were $4.3 million compared to $3.5 million in the prior year period and $4.2 million in the preceding quarter. The increase in R&D expenses was related to the development of our associated processing unit.

Selling, general and administrative expenses were unchanged at $2.8 million in both the quarter ended June 30, 2017, and the prior year quarter ended June 30, 2016, and down sequentially from $3.2 million in the preceding quarter. Total first quarter pretax stock-based compensation expense was $478,000 compared to $518,000 in the prior quarter and $443,000 in the comparable quarter a year ago.

First quarter fiscal 2018 net loss included interest and other income of $98,000 and a cash provision of $81,000 compared to $142,000 of interest and other income and a tax provision of $271,000 a year ago. In the preceding quarter, net loss included other included interest and other income of $183,000 and a cash provision of $2000.

In the first quarter of fiscal 2018, sales to Nokia were $4.5 million or 42.3% of net revenues compared to $4.2 million or 40% of net revenues in the prior quarter and $5.4 million or 41.9% of net revenues in the same period a year ago.

First quarter direct and indirect sales to Cisco Systems were $852,000 or 8% of net revenues compared to $745,000 or 7.2% of net revenues in the prior quarter and $1.5 million or 11.7% of net revenues in the same period a year ago.

Military/defense sales were 24.5% of shipments compared to 24.3% of shipments in the prior quarter and 12.5% of shipments in the comparable period a year ago.

SigmaQuad sales were 51.1% of shipments compared to 53% in the prior quarter and 55.7% in the first quarter of fiscal 2017.

At June 30, 2017, we had $49.1 million in cash, cash equivalents and short-term investments; $14.1 million in long-term investments, $57.6 million in working capital; no debt; and stockholders' equity of $86.6 million.

Looking forward to the second quarter of fiscal 2018, we currently expect net revenues to be in the range of $10 million to $11 million. We expect gross margin of approximately 53% to 55% in the second quarter.

Operator, at this point, we'll open the call to Q&A.

Question-and-Answer Session

Operator

[Operator Instructions] We'll go first to Kurt Caramanidis with Carl M. Hennig, Incorporated.

Kurt Caramanidis

We're about 20 months in, I think, now on the APU and amazingly you're right kind of on the original schedule you had. What, if anything, could delay a year-end, finishing the design by year-end?

Lee-Lean Shu

Right now, we, I think, we are still on schedule. Well, the designs are very complicated. Unless we have some hiccup in the final verification, but actually I don't really see that. But so actually I think we're going to stay on schedule.

Kurt Caramanidis

Okay, great. And then can you talk about there was a speed advantage back when you launched. Can you just update on where you think your speed advantage is today in that recommender systems, data mining search, that kind of thing?

Didier Lasserre

Kurt, this is Didier. It's a combination of performance in power, as we've talked about in the past [Audio Gap] I'm sorry, can you hear us?

Kurt Caramanidis

Yes, I can now.

Didier Lasserre

There's some feedback here. So it's a combination of performance and power where 500 times have that 500 times advantage over the current solution for those applications you're talking about, the recommender systems and SmartSearch as well.

Kurt Caramanidis

And then what do you think even if it's industry numbers or something, the market size for the APU is, if you have anything like that? I know it seems quite large but...

Didier Lasserre

Yes, we don't have anything at this point that we're ready to share. We're still putting it together. What we're finding in the more contact we have with the potential customers and the more type of application we're finding, I mean this whole AI machine learning market is a very young, growing, developing market.

And as you know, as we've talked about the past, when we originally looked at this market, we're looking at the training of the systems. And as we've gotten into it, we realize and we have much more an advantage and there's less of a good solution out there for the recommender systems, and for Search.

Recently, we've also found out that we think that we're going to be a very good fit for natural language processing. I mean, if you look at the dialogue that you plan on having with computers, the computers need to be able to look back at history of your current stations to understand what the context is they're having. And a lot of that done is done real-time memory, which is what we have.

As Lee-Lean mentioned earlier, we're not going off chip to do the memory, we're doing the processing in-memory. So we have an advantage there, which is a new market. And there's also areas where, if you look at existing markets for training, for example, if you train the machine, it takes an amazing amount of time especially with the amount of data that's given for the big data guys, and so what's happening is some of these folks, if there is extra pieces of data that come in, they don't want to have to retrain the system.

And so that's where we fit in as well. We can come in and do the additional training much, much faster or what we've also seen, is talking to some customers, is that they'll use some of their existing systems and solutions to train the system up to a certain percentage, maybe 60%, and then from then on out they use our solution going forward because it's much, much faster. So for us to talk about market size right now is just premature.

Kurt Caramanidis

So it seems to be expanding as you continue to talk to people. And then could you kind of lay out the process after design completion then it goes off to be made and kind of just a little track there?

Lee-Lean Shu

Okay, that is praise but once we get the silicon or we get packaged chip, coming back from that, firstly, we're going to print to what we call the controller module, okay, the controller module, which you can park into the PCI slot in the server. It's pretty much like the quad carrier model so you can, the customer can park into the existing system. So based on that, we're going to [indiscernible] and so the customer can just take it and then that's going to their application and the development of their software. Of course, we have to develop the library to go along with the module for customer to use. That's the current trend. Hopefully, we can quickly develop the design win based on that.

Douglas Schirle

I think Kurt also must know the timing. Once we finish the design, what are the steps, when do we get the first wafer stack and finished parts included.

Lee-Lean Shu

Yes, we expect to get the packed units along the March, April timeframe next year. And hopefully, by the second half of the year, we can have the module already fully tested and the module ready, so we can deliver to the customer.

Kurt Caramanidis

Okay, great. Finally, on the rad-hard, it sounds like you're getting more interest as time is going on there, too? Or am I reading that wrong on your commentary?

Didier Lasserre

No, that's correct. What we've done when we originally spoke about entering this rad-hard market, we were targeting one family, which was the 288-megabit SQ-II+. What we found is there's interest in other technologies as well, so we're quickly introducing also a 144-megabit and, I'm sorry, a 72-megabit NBT, which is No Bus Turnaround, and SyncBurst options as well. So they're in a different package, which should be a quicker turnaround.

And so we're going to be hopefully be able to expand that market just by adding quicker some additional families. We spoke in the past that we would bring on additional families over time, but with the interest we've seen, we've decided to accelerate the new product introduction for rad-hard.

Kurt Caramanidis

Okay, great thanks guys, very exciting timeframe here.

Didier Lasserre

Thanks Kurt.

Operator

We'll hear next from Jeff Bernstein with Cowen.

Jeff Bernstein

Hi guys, yeah, a couple of questions. Talk a little bit about Nokia. They have, I guess, a new 7750 router that's coming out in the second half of the year that, I guess, they're kind of talking about as sort of a new platform, a cloud, a giant cloud router. It looks like it's still based on the same course, but a lot more of them as the current product. So can we assume you guys are still in there as the SRAM content go up in that thing?

Didier Lasserre

So generically, I can't answer that question because if you look at the original design in the 7750, it has their FP3 processor. And with the FP3 processor, they used our 72-megabit SQ-III. Now what you're talking about it could be the midlife kicker or it could be the next-generation, it's unclear. The mid-life kicker will continue to use the FP3, which will use both 72-megabit and also 144-megabit SQ-IIIs, which, of course, we will be supporting those two.

Now the next-generation part they're talking about would use a completely different processor from them, which is an FP4 and it's not right now what the memory is for that. So it's hard for me to answer your question generically, but the answer is there is a mid-life kicker we know about we'll be in and in the next-generation it's not clear yet.

Jeff Bernstein

Got you, okay. And I'm sorry, you said on the mid-life kicker, you would be supplying 72-meg and then additionally what else?

Didier Lasserre

144 megabit so that's not additionally, it's one or the other. So depending on the option, they will either double the density of the current solution we support them with or they'll use the same density.

Jeff Bernstein

I see. Okay. And then on the APUs, I think you guys announced Taiwan Semi 28-nanometer process for that. Can you just talk about that's obviously not their most bleeding edge. Can you just talk about the puts and takes on that?

Lee-Lean Shu

Yes, we feel that performance-wise we don't need the state-of-the-art part of technology. I think just by the front gate, on either we could be just like what we already talked about, the 500 time of performance advantages. So we are not relying on the semiconductor processing technology to win. We are going to rely on the opportunities and the architecture to win, okay. Of course, instead of a 16-nanometer or 10-nanometer, you use 20-nanometer is a very mature processor and old IP is always available. So it's much quicker to bring to the market. That's our philosophy.

Operator

[Operator Instructions] We'll go next to George Gaspar, a private investor.

Unidentified Analyst

Thank you. Good afternoon. My question is related to government deliveries. If I recall, in the past quarter, you indicated that you were getting closer to some high-value premium delivery activity to the government installations or projects. Can you highlight how you're doing on that and what we can expect that to generate revenue going forward?

Didier Lasserre

Are you talking about the rad-hard product line, George?

Unidentified Analyst

Yes.

Originally posted here:

GSI Technology's (GSIT) CEO Lee-Lean Shu on Q1 2018 Results - Earnings Call Transcript - Seeking Alpha