GSI Technology, Inc. (GSIT) on Q2 2024 Results - Earnings Call Transcript
Operator: Ladies and gentlemen, thank you for standing by. Welcome to GSI Technology's Second Quarter Fiscal 2024 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 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, October 26, 2023, 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, and thank you for joining us to review our second quarter fiscal 2024 financial results. I am pleased to share that we have achieved two significant milestones since we reported Q1 2024 earnings. First, was the successful launch of the alpha version of our Copperhead compiler suite, a Python-based tool that can harness the capabilities of the Gemini APU. We anticipate releasing the general release of the Copperhead compiler suite in early 2024. A full production version will be available later in 2024. Currently, we have partners learning, using and developing algorithms with our compiler, including SWAC, a consortium of universities and the companies developing these generation technology for space and academic research group, exploring the future of computing architecture, like MIT and the University of California, Riverside. The second milestone was the completion of the Gemini-II tape-out, which were launched last week. As a result, we are on track to have the chip back in our hands early next calendar year as we expect to begin sampling the device in the second half of 2024. We are targeting Gemini-II partners and customers in low-power data center expansions and enabling data center functions at edge. Example of edge applications could include advanced driver assistance systems and XAG in delivery drone, autonomous robots, unmanned aerial vehicle and satellite. Turning to our financial results for the second quarter. Revenue of $5.7 million was at the high end of our guidance. Also it's worth noting that our second quarter fiscal year 2023 revenue was boosted by inventory builds with several key customers, which presents a challenging year-over-year comparison. In addition to advancing the table of Gemini-II, a significant area of recent focus has centered around our ongoing engagement with the key hyperscaler partner. I'm delighted to report that these discussions are making notable progress. To our constructive dialogue with this leading cloud computing provider, we have gained invaluable insights into the precise design specification required for Gemini-III to align with the early requirement. This collaborated effort has enabled us to charge [indiscernible], while identifying potential partners who can bring the essential financial and engagement -- engineering resource to the table for the successful development, manufacturing and the launch of Gemini-III. The evolution will leverage the conversion of high-bandwidth memory into APU architecture, thereby harness the full potential of in-memory compute advantage. Now I will hand the call over to Didier, who will discuss our business performance further. Please go ahead, Didier.
Didier Lasserre: Thank you, Lee-Lean. Following up on the projects I mentioned last quarter, throughout the second quarter of fiscal 2024, our team continued to pursue opportunities with Gemini-I to advance our customer engagements. Currently, GSI has a fast vector search plug-in available that allows cloud vector search users to seamlessly add APU-accelerated search to their major cloud-provided hosted workloads, with minimal latency from GSI's hosted data centers. This FVS plug-in provides accelerated approximate nearest neighbor search response times, enabling access to a large enterprise service at low power that also meets price points for small- and medium-sized businesses. This will be opening up larger markets for us. Switching to our SAR opportunities. Due to customer feedback, we have decided to launch or low-powered, highly efficient SAR processing as a SaaS offering, along with, of course, also the on-prem version. We are now engaging to bring that service to market. Moreover, we favorably completed benchmarking on a customer's data set. One target we are engaged with a start-up building satellites that can provide computational capabilities on satellites and space has identified the Gemini APU as its preferred provider for computing solutions for space. The big differentiator, in addition to low-power performance, is the product's radiation-tolerant feature. On that note, we will be conducting full radiation tolerant testing on Gemini-I next month. By this, I mean, the full range of tests required for customers considering using Gemini-I in space. One of the applications for that radiation tolerant Gemini-I would be ideally suited for SAR and ATR, which is automatic target recognition, and computing and space to name a few. This summer, we announced that GSI was awarded an SBIR to contract -- I'm sorry, an SBIR contract to perform a feasibility study to adapt Gemini-II to perform computing at the edge in collaboration with the U.S. Air Force and space force. We are currently working on a second SBIR based on the Gemini-II software development, which is very promising and potentially, a similar financial award as the first win. We continue to file more SBIR as they bring two key benefits: number one, a source of revenue, and they also create use cases within the U.S. government for future APU opportunities. Let me switch now to customer and product breakdowns for the second quarter. In the second quarter of fiscal 2024, sales to Nokia were $1.2 million, or 20.3% of net revenues, compared to $1.2 million, or 13.6% of net revenues in the same period a year ago, and $1.9 million, or 33.5% of net revenues in the prior quarter. Military defense sales were 34.8% of second quarter shipments compared to 22.4% of shipments in the comparable year -- I'm sorry, comparable period a year ago and 33.8% of shipments in the prior quarter. SigmaQuad sales were 55.8% of second quarter shipments compared to 58.1% in the second quarter of fiscal 2023 and 58.6% in the prior quarter. I would like now to hand the call over to Doug. Go ahead, Doug.
Douglas Schirle: Thank you, Didier. We reported a net loss of $4.1 million or $0.16 per diluted share on net revenues of $5.7 million for the second quarter of fiscal 2024 compared to a net loss of $3.2 million or $0.13 per diluted share on net revenues of $9 million for the second quarter of fiscal 2023 and a net loss of $5.1 million or $0.21 per diluted share on net revenues of $5.6 million for the first quarter of fiscal '24. Gross margin was 54.7% compared to 62.6% in the prior year period and 54.9% in the preceding first quarter. The changes in gross margin were primarily due to changes in product mix sold in the three periods. Total operating expenses in the second quarter of fiscal 2024 were $7.2 million compared to $8.8 million in the second quarter of fiscal 2023 and $8.2 million in the prior quarter. Research and development expenses were $4.7 million compared to $6.4 million in the prior year period and $5.2 million in the prior quarter. Selling, general and administrative expenses were $2.5 million in the quarter ended September 30, 2023, compared to $2.4 million in the prior year quarter and $3 million in the previous quarter. Second quarter fiscal 2024 operating loss was $4.1 million compared to $3.2 million in the prior year period and $5.1 million in the prior quarter. Second quarter fiscal 2024 net loss included interest and other income of $71,000 and a tax provision of $33,000 compared to interest and other income net of $14,000 and a tax provision of $37,000 for the same period a year ago. In the preceding first quarter, net loss included interest and other income net of $80,000 and a tax provision of $51,000. Total second quarter pretax stock-based compensation expense was $676,000 compared to $661,000 in the comparable quarter a year ago and $820,000 in the prior quarter. At September 30, 2023, the company had $25.3 million in cash, cash equivalents and short-term investments compared to $30.6 million in cash, cash equivalents and short-term investments March 31, 2023. Working capital was $28.8 million as of September 30, 2023, versus $34.7 million at March 31, 2023, with no debt. Stockholders' equity as of September 30, 2023, was $45.4 million compared to $51.4 million as of the fiscal year ended March 31, 2023. Given the current global economic environment, our current expectations for the upcoming third quarter are net revenues in the range of $5.4 million to $6.2 million, with gross margin of approximately 55% to 57%. Operator, at this point, we would like to open the call to Q&A.
Operator: [Operator Instructions] Our first question comes from Brett Reiss with Janney Montgomery Scott.
Brett Reiss : Until corporate initiatives gain traction, how -- can you just comment on cash burn rates? How long do you think the cash will last without you having to do some sort of financing?
Douglas Schirle: Yes. We've taken a look at that. We typically take a look at it every quarter. And if nothing improves, which we don't expect, we have cash that should last at least a couple of years at this point.
Brett Reiss : And if I may just follow up, what do you think your burn will be for this following year?
Douglas Schirle : This year we'll be around $13 million to $14 million. Yes, we have one kind of extraordinary expense coming up this fiscal year that requires about $2.4 million cash outlay. And that's for the mask set for the Gemini-II product that just taped out that Lee-Lean previously mentioned today. That's somewhat of an irregular occurrence that only happens every few years.
Operator: [Operator Instructions] Our next question comes from the line of Orin Hirschman with AIGH Investment Partners.
Orin Hirschman : Can you can you give us a little bit more color on -- is the data center customer interested in Gemini-II for any purpose or only Gemini-III? Have they played with Gemini-I to understand it better? And what is attracting them to the Gemini architecture? And I have one follow-up.
Didier Lasserre : Okay. So let me start with that. So with the hyperscaler, we're engaged with. And we've had when -- we've had ongoing discussions with, and we've since opened up other discussions with other hyperscalers. Certainly, as Lee-Lean mentioned in his earlier comments, the information we gathered from our first engagement was instrumental to be able to sharp our story and our offering for future hyperscale conversations. And so what they're -- interested right now would be a next-generation, specifically, something we'll call Gemini-III for now. Because that one is going to really be geared towards GenAI or the LLM models. With that said, as you know, these hyperscalers are very large companies. And so certainly, Gemini-II is certainly not out of play with them, it just wouldn't be the solution specifically for the GenAI conversations we're having.
Orin Hirschman : And in terms of other hyperscalers and data center customers, it sounds like there's some level of progress. How would you characterize it? Have people played with benchmarks yet for any of them? I think would be an important point to note if it's true if you've gotten that far.
Didier Lasserre : Correct. And so what we've done is we've shown obviously the technology with Gemini-I because silicon is here. Software is here. We can run true benchmarks. We've then obviously run expected benchmarks off of Gemini-II. And so going back, and I just remember what part of your earlier question was, what's interesting about our technology to them? So there's certainly a few areas. Obviously, the performance is important to them. But the low power has certainly grabbed their eye. And also other features like the fact that we're a bit processor. This is very important to a lot of these folks because right now, if they're using a GPU, they're really locked into -- depending on the GPU 8-bit or 16-bit, 32-bit or 64-bit kind of traffic patterns. We're a bit processor. And so we don't care what your traffic looks like. It can be 4 bit. It can be 64 bit. It can be -- make up a number of 1,000 bit, and we can adapt to that on the fly and cycle to cycle. And so that flexibility is also very important to them.
Orin Hirschman : In real life, there's been a lot of talk about a very bit type of super processor in the literature. And obviously, no one's really had one commercially where it's programmable in that fashion. Are you programmable today in that fashion, number one. Number two, is what are some of the applications where that becomes so enticing?
Didier Lasserre : So you're talking about the bit processing?
Orin Hirschman : Yes.
Didier Lasserre : Yes. So there's a lot of research going on right now. And they're finding out that some efficiencies aren't coming in like the expected 16-bit and 32-bit. As I mentioned, 5-bit was one that came up recently. And candidly, I don't remember the exact application or the research on that. But 5 bit was one of them that was important. And so one of the benefits with this bit processing is that as our customers told us, we're future-proof. And what I mean is if there's an application that comes in that needs 5 bit, we're there. If there's one that needs 128-bit, 256 bit, we're there. We don't have to redesign our part to be able to address those needs because of the fact that we have 2 million bit processors on our Gemini-I. And they can work, as I mentioned, as 1 bit or make up a number between 1 million and 2 million. And so we're future-proof from that respect, which is obviously eye opening to them.
Lee-Lean Shu : Yes. Just to comment. Right now, the GenAI, the large language model, is a very big model. It's a very, very memory-intensive application. So they will lead to a big memory. They will lead to very fast memory. So the effort -- one of the effort is they try to reduce the data format, if you have less number of bit to calculate, then you got less memory, you require less memory too. Okay. So we have big processing capability, which you can -- just like Didier mentioned 5 bit, 4 bit, 3 bit, 2 bit, we will be there. And with this is we have more capability to keep up with the software innovation and that's what some of the customer likes about this issue.
Orin Hirschman : And just going back to benchmarking, have you done benchmarking for some of the data center and hyperscaler customers?
Didier Lasserre : So candidly, most of the benchmarking we've done have been based essentially off of POCs, where customers have come in and said, "This is what I'm using. Here's my data set. What do your numbers look like?" And so we've done real life case POCs. We've done that for one of our Israeli customers that we talked about and then also one of our recent SAR customers. As far as broad market benchmarking, we haven't done a lot, but we know that's critical and that's something that's on our agenda to start doing. It's just been a resource issue for us to be able to make that happen. But we understand it's important, and that's what we're next on the list.
Orin Hirschman : And I see. Two more questions, if I may, and then I'll let other people ask. So in terms of proof of concept using customer data, have any of the hyperscalers given you data sets to actually show them what you can and can't do with it?
Didier Lasserre : No. These are other applications. These have been SAR applications and other applications. The hyperscalers know. Because as I mentioned, the discussions early on have been revolving around this next-generation device. And so we've given them -- we have given the benchmarking what we anticipate we could do, absolutely, but they're not true benchmarks, they're calculated benchmarking.
Orin Hirschman : Okay. And last question, just on the you mentioned as you're using the Process-as-a-Service for the people's data, and you had some early tests going back a while ago and that, that needed to be revamped or some additional APIs. Where are you up to with any of the potential customers offering it as a plug-in or anything like that?
Didier Lasserre : Yes. So there's two areas that we're focusing the SaaS on, fast vector search and SAR. And as I mentioned in my comments earlier, the SAR was after discussions with one of the folks that we're talking to in that area. So we have an on-prem solution, which for some of their customers, is needed. And then obviously, one of the guys we're talking to you wants to put this capability on the satellite. So that would be a physical sale as well. But right now, most of their production they're doing is via SaaS over AWS using GPU instances. And so that's when we -- after discussing with them, that's something that we looked at. And so that obviously makes sense for us to do. So we're opening up the SAR for that. The fast vector search we've been putting in place, we're still doing some benchmarking internally for some of the factor search and databases, guys like v8. And so that's an ongoing project for us.
Orin Hirschman : In terms of trying to get that fast factor searches a plug-in so that customers could actually play with it through AWS or something like that. Any progress there?
Didier Lasserre : There is -- I'm trying to understand what I can see there. So we've had some discussions recently with a very large data center company to basically integrate easier into their system right now. And so those discussions have been happening over the last couple of weeks. And then certainly, once those get worked out, then we'll be able to have a smoother offering right now. Because as you know, right now, it's -- we're basically -- they're offloading the searches off to our data center that we have put into a facility called CoreSite. And so we're trying to streamline that offering.
Operator: [Operator Instructions] Our next question comes from the line Luke Benett, an investor.
Unidentified Analyst: To zoom out from parsing applications and the prospects for various partnerships, could we assume that when Gemini-I hits the scene and you all are able to offer that to be tooled by various users, will that outperform a large swath of the current high-performance computer offerings that are on the market, including the leaders of NVIDIA, AMD, Samsung, even Google, custom chips. Is that safe to say?
Didier Lasserre : So -- okay, so I'm trying to follow your question here. So first of all, we do have Gemini-I already. So I'm assuming you were referring to Gemini-II.
Orin Hirschman : Gemini-II, correct. Yes.
Didier Lasserre : Yes. And so Gemini-II, and then you also mentioned a custom chip. So Gemini-II is not a custom chip. So Gemini-II is going to be a standard offering that we're going to have. And so yes, obviously, we've done some calculated benchmarkings, which have looked very favorable. And once we have the chip in hand, as Lee-Lean said, we'll have it in hand early next year, and it's probably going to take at least the spend to really get something we can do benchmarking. So it will be around summertime before we can do true benchmarking. But yes, we anticipate them to be very favorable over what we have and what's in the market space. Now -- but I do want to add to the comment, you said custom chip. So one of the things -- and we've mentioned this in the past, that we're also looking to do is obviously offer IP sales as well, because we do understand we have a unique technology. But we also understand that certainly, we'll start with the hyperscalers. They do a lot of their own custom silicon. And so one of the conversations that we're having with the hyperscalers besides talking about Gemini-III for the GenAI is we're also talking about IP sales as well for any custom basis that they're doing on their own. So you asked a lot of questions in there. I'm hoping I hit all the topics.
Douglas Schirle : Yes, that does kind of answer the two implications there. That Gemini-2 will be a compelling competitor and depending on how you're able to benchmark or how much tooling and rendering it requires to via IP or via that chip directly, but that it will. Yes, essentially, at 16-nanometers, the competing toe-to-toe or you mean clearly out competing in a lot of these 5-nanometer leading high-performance chips.
Didier Lasserre : If you recall, Gemini-I, the benchmarking we've done in the past have been very positive against what's out there today. Now with Gemini-II, as we discussed in the past, it has 10x of performance and 8x the L1 memory on chip. So it's certainly going to be something that's going to be very compelling for the market.
Operator: Our next question comes from George Gasper, an investor.
Unidentified Analyst: Could you go into a little additional information for us on the military and defense applications and the interconnect that the company has had in Israel over the years, in fact, initially the whole Gemini approach came out of there? And considering what's going on in space, is there broader applications for the advancements that you're making going into space? And is what's happening with Israel right now in terms of this situation in the Middle East. Is there -- are you seeing anything happen that could create a more military-related effort?
Didier Lasserre : Okay. So I'm going to try and get all those questions now. So let's start with the military. So yes, there are a lot of different possible use cases with the military, which is also why we've had some successes recently with these SBIR that we've been filing. So some of the interest that we've seen have -- from like we'll call the mill arrow kind of area is certainly SAR for sure. Object detection, is another automatic target recognition, another. We've also been for imaging as well for change detection. Just -- and also one of the customers that we're engaged with, in fact, they have two of our servers on loan that they're doing demonstrations with is they want to start putting data centers in space. And so that's another -- just a standard data center application in space. And so that's another one. As far as I'm not sure what the Israel site. So as you say, this original technology was an acquisition we made, and the company was out of Israel. But I'm not sure what the question was revolving around that acquisition.
Unidentified Analyst: Well, basically, just as a reference, I mentioned it. But considering what's going on in the Middle East right now and the connections that the company has had with Israel, I would think that there would be an attempt to try to get some applications moving forward to -- give Israel a more recognition in how they're trying to monitor things. Is that…?
Didier Lasserre : Yes. So we've had some discussions, obviously, with fat that we've talked about in the past and other entities there on several applications. I won't go to exact detail, but there's certainly -- I'm sure we'll -- there will -- certain areas that we've talked about with them will, most likely get accelerated now because of obviously what's happening now. But right now, we're not having those discussions right now. There's more pressing issues in Israel right now.
Unidentified Analyst: I see. Okay. Well, it looks like you're about to make some generally movements forward with Gemini-II going to III. Hopefully, this really starts to turn the company around in terms of its revenue volume and would expect that by the end of this fiscal year. There's some real momentum going forward. Can you comment on it?
Didier Lasserre : Yes. There's certainly momentum. And as time goes on, we hope to announce more traction. But the end of the fiscal year, any revenue that comes will be coming from Gemini-I. So as we've discussed, Gemini-II, we're not going to be even sampling it until summer of next year. And so that falls into fiscal '25.
Operator: As there are no further questions, I would now hand the conference over to the management for their closing remarks.
Lee-Lean Shu : Thank you all for joining us. We look forward to speaking with you again when we report our third quarter fiscal 2024 results. Thanks.
Operator: Thank you. The conference of GSI Technology has now concluded. Thank you for your participation. You may now disconnect your lines.
