GSI Technology, Inc.

Ladies and gentlemen, thank you for standing by. Welcome to GSI Technologies' third quarter fiscal 2022 results conference call. At this time, all participants are in a listen-only mode. Later, we will conduct a question and answer session. At that time, we will provide instructions for those interested in entering the queue for the Q&A. 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, January 27, 2022, at the request of GSI Technology. Hosting the call today is Li-Lin Xu, the company's chairman, president, and chief executive officer. With him are Douglas Shirley, chief financial officer, and Didier Rousserre, vice president of sales. I would now like to turn the conference over to Mr. Xu. Please go ahead, sir.

Good afternoon, and thank you for joining us to review our physical third quarter 2022 financial results. Our third quarter revenue grew year over year by 19% due to increased order from our Western customer base, including network and the telecom, military and the defense, as well as automotive, medical, and test and measurement sectors. The increase in orders from other customers offset the lower orders from Nokia in the quarter. was margin input by 800 basis points year-over-year, reflecting our ability to manage our real-time supply chain challenges and increase costs. Our net loss narrowed year-over-year by 12%, and we ended the quarter with $48.1 million cash, cash equivalents, and short-term investments. We continue to prioritize the allocation of our capital on R&D for the APU, which includes the software and API development for Gemini 1 and the table for Gemini 2. The software development team continue to expand our library of algorithms and API for our target application in search. In February, we plan to release a compiler stack NDDL will provide an overview of this in his sections. At the end of the calendar 2021, we announced that GSI was among the leaders in the bidding scale of the Approximate Leaders Label Search Challenge. We were encouraged by this outcome. The results showed that the APU technology and the software performed on par with prominent industry leaders in AI, or with much broader resource than GSI. Contests teach us a lot about what we need to further develop and build, and how we can improve our performance. Contests like Big ANN and the Market Challenge are also effective ways to raise GSI's profile in our industry and with potential customers. This contest spotlights all teams' capabilities and increases awareness of all APU's unique value propositions. Last year, our first place win in the MAFE Challenge, resulting in our introduction to IAI slash ELTA, who engaged with GSI to develop a soft image processing acceleration system using the APU and our software. Yalta is funding the R&D to build a SaaS system based on APU technology. Yalta will grant an exclusive perpetual license of the fast SaaS algorithm to GSI to allow GSI to sell a SaaS system, including the algorithm, to other customers. Assuming that the POC is successful, Yalta will deploy the system in volume. The GSI team continue to build a pipeline of customer engagement and increase our POC activity for the APU. In addition to ELTA's subproject, we are in discussions with several organizations to engage in POC for a variety of user cases for search applications we are targeting. Our current POC engagements include indexing medical image and text, encrypted communication, and natural language processing. The prime contractor is interested in testing for military and defense applications, and we have a government agency considering the APU for object detection and the facial recognition. We anticipate that a portion of this POC will move to design wings and eventually to production. A few years back, once we had the Gemini One chip available for testing, we made a strategic decision to pursue search instead of the training and inference markets, where there were already large, well-funded competitors. The APU performed significantly better in search applications, and we decided to target our results to develop that market. This has proven to be a good decision, as we are now seeing that search may ultimately be a large end market. We recognize that the APU can work very efficiently with a trained database to rapidly search and deliver results with low latency, low power consumption, and scalable configuration. Accuracy, a smaller footprint, lowering the total cost of ownership, Having space-ready chips and reducing power consumption are key differentiators for the APU. We remain committed to our strategy and dedicating our capital to deliver the best technology for the rapidly growing search market. Supply chain constraints continue to impact our order fulfillment. While there has been some improvement, the situation remains fluid. We have passed on some of the price increase and have factored in the extended lead times, but we do not expect significant relief from this concern before the end of year 2022. We have increased our operation focus on managing the supply chain to fulfill orders that we have in hand for the upcoming quarters. That will provide our outlook for the fourth quarter in the prepared prepared comments. This year will be a crucial year to proving all the feasibility of our groundbreaking APU technology. Our next important milestone is the release of the compiler stack in February. This release shows a salary customer adoption of our groundbreaking technology that can boost lower enterprise AI capacity and bring that capability to standalone servers at age. We are in the early stage engaging with customers who recognize the APU potential for their applications. The compiler stack should help us grow our pipeline. It is an exciting time for GSI, and our team is focused to achieving our goals for current customer engagement while continuing to build our pipeline to new opportunities. Now I will hand the call over to Didier, who will discuss our business performance further. Please go ahead, Didi.

Thank you, Lilien. As Lilien stated, we have new customer engagements in multiple applications ranging from e-commerce to military and defense for various use cases. In our approach to the search market, we are pursuing opportunities that align with the applications we feel the APU is best suited. Right now, our team is seeding the field, and we fully expect that some of this will bear fruit. Our approach to supporting these customer engagements depends on the end users' needs. For some, we loan them a board to demo, others have access to our Sunnyvale data center to run their models, and then some have purchased some boards. Each of these engagements target unique market applications, and the ultimate size of each opportunity may vary. Since many are in the early stages of development, it is challenging to measure. With that said, they are all of sufficient size to merit allocating our resources. Today we are working with universities with unique applications, with broad market opportunities, and we also have prime contractors that could eventually turn into large orders and a government agency evaluating for the APU for border security. There's also an e-commerce company looking at the APU for natural language processing. You can see why we're excited about these engagements, as they could potentially develop into customers and growth markets. Our objectives are to make the POCs as successful as possible for each entity, expecting a portion to move to design wins and eventually to production. As part of the GSI developing a strategy for the AI and HPC acceleration markets, we will soon be releasing an APU compiler stack. This release will deliver the environment for software and AI developers to write full applications using C for our LITA-E and LITA-S PCIe boards. The release will include example applications and training materials. Customers that need a specific use case algorithm can edit our large collection of vector library functions and write their own or write their own APU fragments in Python using our low-level Bellix stack product. Achieving the APU compiler stack milestone should accelerate customer adoption and is one of the most important steps in front of us to move the APU to the next level. As I said earlier, the GSI team has been busy creating APIs for search workloads like image indexing, object recognition, natural language processing, as well as HPC workloads in the enterprise and real-time applications like SAR, ATR, and signal classification at the edge. We have documented that the APU technology can offer enterprise AI capacity in the edge I'm sorry, in an edge standalone server in support of that migration. The APU compiler stack is a powerful tool for researchers and developers that promises to significantly expand the market applications for our non von Neumann processor. To help build a pipeline, GSI will introduce a series of training seminars to support the customization of our APIs for specific applications. The APU technology is available in a PCIe board for enterprise and edge server applications, capable of searching tens of millions of vectors per board. The platform is highly scalable, allowing us to connect unlimited boards installed in servers. The hardware capability can also function as a cloud-based software as a service for open search and elastic search acceleration. Once we launch the compiler stack, we plan to provide more detail in a press release. Let me switch now to the customer and product breakdown for the third quarter. In the third quarter of fiscal 2022, sales to Nokia were 1.9 million or 24% of revenues compared to 2.8 million or 42% of revenues in the same period a year ago and 1.9 or 23.8% of net revenues in the prior quarter. Military defense sales were 27.1% of third quarter shipments compared to 26.0% of shipments in the comparable period a year ago and 27.4% of shipments in the prior quarter. Sigma quad sales were 40.5% of third quarter shipments compared to 62% in the third quarter of fiscal 2021 and 52.4% in the prior quarter. In closing, as Lien mentioned, we have ongoing challenges with supply chain constraints with substrate lead times out more than a year. Overall, the impact of increased wafer prices, assembly labor costs, rising substrate prices, and expediting charges has led us to raise prices to all of our customers in December. I'd now like to hand the call over to Doug. Doug, go ahead, please.