GSI Technology, Inc. (GSIT) Q4 2020 Earnings Report, Transcript and Summary

Ladies and gentlemen, thank you for standing by. Welcome to the GSI Technology's Fourth Quarter and Fiscal Year 2020 Results Conference Call. [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 involves risks and uncertainties that could cause actual results to differ materially from those anticipated. These risks and uncertainties are described in the company's Forms 10-Q and 10-K filed with the Securities and Exchange Commission. Many of these risks are currently amplified by and will continue to be amplified by or in the future may be amplified by the COVID-19 global pandemic. Additionally, I have also been asked to advise you that this conference call is being recorded today, May 7, 2020, 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 and Didier Lasserre, Vice President of Sales. I would now like to turn the conference over to Mr. Shu. Please go ahead, sir.

Good afternoon, and thank you for joining the call. Today, we will explain the status of our business in the current environment, provide an update on our APU products and report on our fourth quarter and full year fiscal 2020 financial results. Despite the uncertainty created by the COVID-19 global pandemic, we remain focused on our long-term strategy. The COVID-19 global pandemic has created unique circumstance for all our stakeholders. And we have instituted many preventive measures and are regularly evaluating those measures and others as we continue to better understand the environment that we are and will be operating in. We are confident in GSI Technology's ability to succeed in these difficult times, given our strong financial position. Our team's tremendous experience and knowledge, and our legacy business that provides funding for our new product lines. As of March 31, 2020, we had over $70 million in cash, cash equivalents and short-term and long-term investments with no debt. Outside of Taiwan, there were -- where there are low [indiscernible] home owners increase, the majority of our employees around the world are working from home. Our productivity has remained high and we are moving forward on achieving our product milestones. I'm grateful for all of our employees' commitment to delivering excellence in product, service and support during these challenging times. In the fourth quarter, we completed the benchmarking of our initial APU product, Gemini-I. In early April, we published the benchmark results for Gemini-I in a billion-scale similarity search using the server containing 4 Gemini APU boards running at a rate of 400 megahertz. The results show that Gemini-I performs extremely fast, accurate search on large datasets with a smaller footprint and lower power consumption for a much lower total cost of ownership. Gemini is a high-performance in speed and accuracy with lower total cost of ownership is the significant distinguishment in today's search solutions. Real-time search of large databases is increasingly a critical function based transforming business, scientific and governmental process. The ability to embrace more extensive solutions and provide better performance at a lower cost will enable GSI to end the new market with Gemini-I. There are many differentiators in the Gemini-I's architecture that creates this enhanced performance. The primary differentiator is the memory sensing design that removed the limitation of serial processes. The Gemini APU-based machine trained the database into the memory and also serve as processors. This configuration shows the system to use a fewer CPUs and GPUs in compressed computation speeds from tens-of-minutes to seconds, even microseconds, in numerous critical search applications. Another key differentiator is the [ number of ] ALUs or arithmetic logic unit on the Gemini chip. A GPU typically has thousands of ALUs. And the CPU has about dozen. Gemini-I has 2 million ALU, allowing the APU to break down the query into a much more granular search. Last of the internal memory and the database load into the APU, reduce search times and use less power to perform the computation. Scalability is another critical differentiator of the Gemini APU. For large data bases that cannot fit in one APU chip, we can integrate multiple boards with many chip into the server. We can build servers with multiple Gemini boards to scale with the size of the database by retaining our performance and keeping power consumption low. Gemini-I memory loaded with the entire database allow for extremely fast processing time. In some search educations, current solutions require a large number of extensive GPU and CPUs to deliver performance at scale. The search market is growing in these new solutions, such as the Gemini APU that delivers own performance while keeping the total cost of ownership low. Finally, Gemini can serve the broader search market as a universal hardware solution and address many different, specifically used case sent with flexibility and customizable software. Our software library and algorithm contained a Gemini for unique search applications with a multiple level of custom programming from assembly code up to TensorFlow or PyTorch. There are many different applications for search. We will continue to develop our solution to address more of this application over time. Despite the unique current circumstance for all our stakeholders caused by the COVID-19 global pandemic, we remain focused on executing the multiple opportunity that lies ahead for GSI. That's the last extent of our progress on Gemini-I. The core business will be variable for the near term for all [ Asian ] products continue to be essential for military and defense as well as telecommunications and the networking customers. Doug will comment on our first fiscal quarter outlook in his prepared remarks. Now I will hand the call over to Didier, who will highlight the sales and marketing progress on Gemini-I and review our business segment performance. Please go ahead, Didier.

Thank you, Lee-Lean. I'm pleased to report that we are actively marketing Gemini. Our marketing team has created sales collateral to educate end users on all the unique performance characteristics and capabilities of Gemini-I. Outreach is now ongoing for facial recognition and drug discovery. We are preparing to commence engaging with military and government customers for signal classification, object detection and cryptography. Due to current COVID-19 shelter-in-place and work-from-home restrictions, we are building remote cloud-based workshops. We plan to use these remote workshops for demonstrations of Gemini-I's functionality and conduct training with customers on our algorithms and software libraries. Let me share a couple of specific examples of Gemini-I's performance in 2 applications we are currently pursuing. I am very excited about Gemini's performance in facial recognition search. Our performance data shows prospective customers that the Gemini-I facial recognition search engine can process databases of up to 200 million faces with a single APU board. In these searches, Gemini-I is delivering highly accurate results while reducing the query times from many minutes to fractions of a second, all with much lower power consumption and edge processing. Another exciting market we are currently developing is drug discovery. You may recall that in the late 2018, we announced a collaboration with the Weizmann Institute in Israel on cheminformatics research. They were an early user of the beta Gemini APU solution. They saw markedly improved productivity with the APU's parallel processing technology with the ability to run multiple molecule queries at speeds that far exceed the CPU search capability. Now researchers at the Weizmann Institute are using newly installed boards containing the Gemini-I APU. Based on the premise that structurally similar molecules exhibit similar biological activities, scientists are researching for the structurally similar compounds to existing drugs that have been successfully -- sorry, that have been successful in treating diseases and finding viruses. With boards containing Gemini-I, Weizmann Institute researchers perform quicker searches using more queries to their database to find molecules with similar properties to their query molecules. The ultimate goal is to decrease the time needed to go from research into clinical trials. Shifting to our sales breakdown for the fourth quarter of fiscal 2020. Sales to Nokia were $2.4 million or 28.3% of net revenues compared to $5.2 million or 41.1% of revenues in the same period a year ago and $2.6 million or 26.3% of net revenues in the prior quarter. Nokia sales have been slower in recent quarters, which we expect to continue through the end of June. The situation is somewhat fluid and is subject to change, but at this time, Nokia's preliminary forecast shows an improvement in sales in the second half of calendar 2020. Military and defense sales were 30.9% of fourth quarter shipments compared to 22.1% a year ago and 18.9% in the prior quarter. SigmaQuad continues to be our best performing product category with sales of 44.7% of fourth quarter shipments compared to 57.4% in the fourth quarter of last year and 62% in the preceding third quarter. I'd now like to hand the call over to Doug. Please go ahead, Doug.

Thank you, Didier. For the fiscal year ended March 31, 2020, we reported net loss of $10.3 million or $0.45 per diluted share, on net revenues of $43.3 million compared to net income of $163,000 or $0.01 per diluted share, on net revenues of $51.5 million in the fiscal year ended March 31, 2019. Gross margin for fiscal 2020 was 58.5% compared to 61.4% in the prior year. Total operating expenses were $36.1 million in fiscal 2020, an increase of 13.6% from $31.8 million in fiscal 2019. Research and development expenses were $25.2 million, compared to $21.4 million in the prior fiscal year; and selling, general and administrative expenses were $10.9 million compared to $10.5 million in fiscal 2019. The increase in research and development expenses was primarily due to the development of APU, the company's AI processor, including a charge of $2.7 million during the third quarter of fiscal 2020 for purchased intellectual property that is being incorporated into the next generation of Gemini chips. Fiscal 2020 operating loss was $10.8 million compared to an operating loss of $182,000 in the prior year. The increase in operating loss is primarily due to the decrease in revenue and gross profit and higher research and development expenses. Fiscal 2020 net loss included interest and other income of $712,000 and a tax provision of $247,000 compared to $450,000 in interest and other income and a tax provision of $105,000 a year ago. The company reported a net loss of $3.8 million or $0.16 per diluted share, on net revenues of $8.5 million for the fourth quarter of fiscal 2020, compared to a net loss of $102,000 or $0.00 per diluted share, on net revenues of $12.7 million for the fourth quarter of fiscal 2019, and a net loss of $4.6 million or $0.20 per diluted share, on net revenues of $10 million for the third quarter of fiscal 2020. Gross margin was 52.5% compared to 61.3% in the prior-year period and 60.2% in the preceding third quarter. The change in gross margin was primarily due to changes in the mix of products sold and the negative impact of manufacturing overhead on reduced revenues. Total operating expenses in the fourth quarter of fiscal 2020 were $8.4 million, compared to $8.1 million in the fourth quarter of fiscal 2019 and $10.8 million in the prior quarter. Research and development expenses were 2 -- were $5.6 million compared to $5.6 million in the prior-year period and $8.2 million in the prior quarter. The decrease in R&D expenses from the prior quarter reflects a charge of $2.7 million during the third quarter of fiscal 2020 for purchased intellectual property that we are incorporating into the next generation for our Gemini chips. Selling, general and administrative expenses were $2.8 million in the quarter ended March 31, 2020, compared to $2.6 million in the prior year and $2.6 million in the previous quarter. Fourth quarter fiscal 2020 operating loss was $3.9 million, compared to an operating loss of $365,000 in the prior year period and an operating loss of $4.7 million in the prior quarter. Fourth quarter fiscal 2020 net loss included interest and other income of $148,000 and a tax provision of $65,000 compared to $186,000 in interest and other income and a tax benefit of $77,000 for the same period a year ago. In the preceding third quarter, net loss included interest and other income of $207,000 and a tax provision of $84,000. Total fourth quarter pre-tax stock-based compensation expense was $644,000, compared to $580,000 in the comparable quarter a year ago and $629,000 in the prior quarter. At March 31, 2020, we had $66.6 million in cash, cash equivalents and short-term investments and $4.1 million in long-term investments, compared to $61.8 million in cash, cash equivalents and short-term investments and $90 million in long-term investments at March 31, 2019. Working capital was $70.9 million as of March 31, 2020 versus $68.6 million at March 31, 2019 with no debt. Stockholders' equity as of March 31, 2020 was $89.6 million, compared to $93.2 million as of the fiscal year ended March 31, 2019. Fourth quarter results were within the range of guidance provided in our last earnings release for net revenues in the range of $8.5 million to $9.5 million with gross margin of approximately 50% to 52%. Our current expectations for the upcoming quarter are net revenues in a range of $7 million to $8.8 million with gross margin of approximately 48% to 52%. We've expanded our guidance range for the upcoming quarter to reflect the impact of possible changes in customer buying patterns and communication limitations related to shelter in place restrictions that require a significant number of our customer contacts to work from home. The disruption to the marketplace that we are currently experiencing is like anything we've ever had to deal with. While we continue to monitor the business metrics that we have historically used to predict our financial performance, we are uncertain as to whether these metrics will operate consistently with our historical experience. Operator, at this point, we'll open the call to Q&A.

[Operator Instructions] We'll now take our first question from Kurt Caramanidis.

Can you hear me?

Yes Kurt, we can hear you.

Great. Okay. Weizmann, are they getting just some free chips and boards as a collaborator? Are they planning on needing to order any or how is that kind of going? Are they being your guinea pig to be able to sell to other places?

Yes, to both accounts. So we have loaned them in the past and we have given them the upgraded, the new system recently, as we mentioned, and they plan on paying for that system.

Okay. Are they a more modest prospect compared to some other larger drug companies? Or how do -- because it seems like right now, that area is as hot as it's ever going to be, and I'm wondering how you're able to get the word out or get some presentations to some of these drug companies?

So we've -- as I mentioned in the discussion earlier, we have the marketing collateral that we've recently put together within the last month. And it consists of a sales brief, which is a 1-page front and back high level explanation, and we've broken it down by market segments. So we have one currently for facial recognition and we have one for the drug discovery. And we put that out to the world. And then as a backup, we have more in-depth presentations and then more in-depth technical write-ups, almost like white papers along with a presentation for each of these market segments. And so we've been doing some of the demos and some of the presentations based off of the initial outreach with the 1-page sales brief.

How do you -- how are you feeling early on about those conversations or presentations?

Good. I mean, certainly, if you seen some of the results that we put out, some of the benchmarking, everything looks really good. I mean, we're certainly early in the process. As I mentioned, we just received these documents within the last few weeks. So it's early in the process, but certainly, the results are what we expected. And so we're communicating that to the marketplace now. So I mean, so far, it's been good feedback.

Okay. And facial recognition, I would agree the results of what you're doing are extremely impressive. So I'm trying to game out how we might start getting some orders of some sort. The facial recognition looks extremely positive. Do you have some type of a partner there? And then how is that sales process? Have some of them already been aware? Or are you starting from scratch? How is that looking?

So we had 2 partners, which we [ sold ] one. One of them was having some financial difficulties, but the other one is still there. They're -- we're doing a POC with them right now based off of their algorithms. And so that's going to take some time. And then beyond them, beside -- beyond the initial partners, we're now, as I mentioned, reaching out to other folks in this space to be able to give them presentations and in more detailed white papers.

Is government part of that?

Government is part of that, yes.

Okay. One other thing and then I'll let if somebody else is in, but either Doug or probably Doug. I get questions, well, I saw so and so selling the stock and I tried to explain, you got -- you guys get paid in options, they come due and many times you're selling some stock to cover the taxes or whatever. Can you just explain -- because sometimes it almost seems like an overhang, people are like, why would they be selling stock if things are looking so good? But explain a little bit about the options, I think, are part of your compensation and then what the mechanics -- I think most of the time, there's really not net selling, it's generally even or acquiring, but just put some color because that seems to be confusing?

Yes. So every employee at GSI Technology receives an option grant when they're hired. And those typically vest over a 4-year period, 25% per year. It's not monthly, it's annually. And that's true for executive officers. Subsequent to that, every year, we have Evergreen Grants, and everyone gets an Evergreen Grant on their anniversary or in the quarter just after the anniversary of the hire date. So those would vest over 4 years for employees and officers when they get theirs, they invest 4 years into the future. It's not a 4-year vest, so it's 100% [ cliff invest ] after 4 years. And all of these options have a 10-year life on them. And there are a number of people with options that are expiring. I think we're on May 10th [indiscernible] started this week, and they're a lot more expiring in August. And a lot of times when you see officers selling shares, in many cases, they sell shares and use the net proceeds of those sales to buy additional shares. So from an officer perspective, most of the officers are acquiring shares, increasing their position in the company, not decreasing. We've also had some Directors who will also get option grants. They have also, over time, been acquiring some shares. So when you see people selling, it's really not because they want to sell, it's because they don't have a choice. They need the money to buy additional shares, which they typically have been using those proceeds to buy additional shares and other grants.

Great. I just needed to have that clarified, I think, just generally speaking, so I appreciate that.

[Operator Instructions] There is a follow-up question from Kurt.

Sorry. I know today is crazy. There's multiple calls. So sometimes there's only 1 or 2 on the call, but I wanted to give somebody a chance if need be. Any update on Rad-Hard, just anything for the good of the order on getting heritage or any outlook out there?

So we finished the qualification. So we are now on the 288 megabit, SQ-II+ family, which is the lead family we started with. We have finished qualification, so we now have -- and not only did we finish our internal qualification, but our partner finished their V certification. So they are QMLV certified now. So we are now able to offer QMLV equivalent devices. So we grabbed all the data sheets. Now we're still obviously going through the -- like you said, we're getting the heritage and going asking some of the folks. So it's not clear whether our first orders will be Rad-Hard or Rad-Tolerant because we do have some nice wins in the Rad-Tolerant. We're just waiting for the funding and the orders to come from that. As far as -- we were talking about in the past about hardening also our SyncBurst NBT family along with APU. APU, as we've discussed in the past, we've done some preliminary testing back in November and the results were very promising. We were hoping to get back into the facility to do more, be in time actually in April, but obviously, with everything going on with the COVID-19, that's on hold right now. So adding new families right now is on hold until we can get back into the facility. But at this point, the parts that are qualified and we do have testing on, we're continuing to promote and it's just a matter of the timing of the funding and when those will come to fruition.

Okay. Great. And you did some training on APU, was it January before all this stuff hit?

It was last November.

Oh, I thought you didn't have anything at the office in your training facility there this year. I was thinking something was in January? With the people that you've contacted before, do you see any chance of indications of interest, some type of orders through the summer, too hard to tell. What's kind of your gut on that?

Yes, it's too difficult to predict. And I'll tell you what, I was having a conversation sometime back when somebody asked me, so when do you expect an order as I was talking about our Rad-Tolerant opportunity. And I said, I expect the order last third quarter because that's when it was supposed to come in. And so for me to make a prediction...

I'm sorry, I'm on to -- I am on APU. I apologize. I switched gears. On APU -- I meant on APU, I thought -- yes, sorry about that. So I thought you had a training for APU in January at the headquarters or some kind of training thing. So my question is more about APU. Do you see getting any indication of interest or possible POs of some sort this summer as some of these people have maybe been tracking along the way? Or what's kind of your feeling about that, I apologize I meant APU, not Rad-Tolerant...

I'm sorry. Yes. So the APU, that is correct. The training was done in the first quarter, it was -- I can't remember if it was at January or beginning of February. And -- but you're correct. We went out and did the training, and it happened to be for a government entity who actually purchased a board from us back in December. And so that -- those conversations are still occurring. We've also done something additional for that -- for the APU in that market spaces. We've become a member of something called SHREC, it's S-H-R-E-C, and so we're part of that member and so what we have is we have some active PhDs and also other people within the SHREC organization or Action Star organization, but group that we'll be doing some work with us on adding additional resources for the APU, specifically in areas where they can write additional libraries or algorithms or they can make changes to the FPGA that is currently on our APU board to be able to utilize more of the resources within the FPGA that's on the board. And so that's happening now. And those are all geared towards the military government type of applications. So that's a recent within the last quarter update.

So do you think this summer, it's possible to get some initial orders? Or how are you thinking about that?

Hard to predict, Kurt. I mean, I could say yes and it doesn't happen. It's just really hard to predict when it's going to happen, but certainly, as we mentioned in the conversation at the beginning, between signal classification, between object detection, cryptography, there's a massive amount of interest for this type of technology in that government space.

Okay. And congratulations, I mean, I was unsure how the performance was going to look, but to somewhat of a lay person, I mean, certainly, it looks saleable. It just depends on getting around the environment we're in. But certainly, the performance metrics look very impressive. And one question, Lee-Lean was talking about overall cost. So you were talking about the price of these. So at the price you were thinking of the APU boards, it still would be cost-effective when you're looking at speed, power, performance footprint or do you -- or would you have to have a lower price than what you're originally thinking?

No. There's no reason for us to have any lower prices in any of our initial assessments.

Okay. Great guys. It's been a long road, and now we've got a pandemic. So work through that and look forward to how this all pans out.

There are no further questions in the queue at this time.

Thank you all for joining us. We look forward to speaking with you again when we deliver our first quarter fiscal 2021 results. Thank you.

Ladies and gentlemen, this concludes today's conference call. Thank you for your participation. You may now disconnect.