Remark Holdings, Inc.

Good day and welcome to the Remark Holdings Fiscal 2023 Financial Results Conference Call. All participants will be in listen-only mode. Should you need assistance, please signal a conference specialist by pressing the star key followed by zero. After today's presentation, there will be an opportunity to ask questions. To ask a question, you may press star then one on your telephone keypad. To withdraw your question, please press star then two. Please note, today's event is being recorded. I would now like to turn the conference over to Fei Tian. Please go ahead, ma'am.

Thank you, Rocco. Good afternoon, everyone, and welcome to Remark Holding's Fiscal 2023 Financial Results Conference Call. I am Fei Tian, Vice President of Investor Relations for Remark. On the call with me this afternoon is Mr. Kai-Hsing Tao, Remark's Chairman and Chief Executive Officer, and Mr. Todd Brown, Vice President President of Finance. In just a moment, Mr. Tao will provide an update on our businesses, and Mr. Brown will recap our 2023 financial results. Following these remarks, we will open the call to questions. But before I turn the call over to Mr. Tao, I would like to take this opportunity to remind you that some of the statements made today may be forward-looking statements. These statements involve risks, uncertainties, and other factors that could cause actual results to differ materially from those expressed or implied by these forward-looking statements. Any forward-looking statements reflect Remark Holdings' current views, and Remark Holdings expressly disclaims any obligation to update or revise any forward-looking statements after the day's share-off. This disclaimer is only a summary of Remark Holdings' statutory forward-looking statement disclaimer, which is included in full in its filings with the SEC. I will now turn the call over to Remarks Chairman and Chief Executive Officer, Mr. Tao, so he can provide additional information on Remarks businesses and recent developments. Xin, please.

Thank you for joining the call. Our most important achievement in 2023 was setting the groundwork for 2024 and beyond. The primary highlight was the signing of our business partnership with Microsoft. Up until this point, our sales efforts have been mostly direct. However, that has changed, and that is due to the value proposition that the Remark AI platform has brought to the table. It took almost 18 months of due diligence on our platform, and we are now honored and excited to be able to leverage with Microsoft the thousands of sales teams to market our solutions to their enterprise customers. This collaboration allows Remark to rapidly broaden its sales and marketing for the Microsoft Azure Marketplace, providing institutional credibility for prospective IT departments approval, as well as the rapidly scaling Remark AI's inference and training models. The five year agreement is expected to drive $240 million of AI business for Remark by 2029. But with all the hoopla on AI, I want to clarify on how we are positioned. Right now, AI is all in the press, but most typically it's about generative AI or large language models. At Remark, we have been doing generative AI since 2017. However, we felt back then, as we do now, that it's a tough business proposition. Cool technology does not mean people will buy your product in a profitable manner. The open secret, as they say, is that in generative AI, everyone is copying each other and therefore will result in a disastrous war. At Remark, while generative AI is a feature we provide, our strength and focus is on computer vision and the building of large vision models, which in turn leads to our core platform, of multimodal models as a service. In 2023, we launched our first fast AI platforms. First large vision model business application. The large vision model smart construction. Fast AI platform is a few shot training framework based on our large vision model. A large computer vision and semantic multimodal model. The release version is its $34 billion parameter model with the proprietary dynamic prototype fusion network. It is designed to address limited training samples in complex environmental problems in various industries, and it's achieved extremely high performance with Microsoft COCO dataset for complicated object detection and segmentation. The large vision model, otherwise known as LVM, smart construction product, has learned federal, state, and industrial-level safety, quality, and environmental protection guidelines and SOPs, as well as OSHA laws. It can autonomously perform safety, quality, and environmental check-through video streams from fixed security surveillance cameras and wearables, i.e., 360-degree cameras, hardhead-mounted GoPro cameras. It is specialized for worker safety, hazard risk management, meaning fire, collision, and traffic, equipment operation safety, environmental safety, construction progress estimation, and quality checks. It is even powerful enough to allow for new and updated SOP and rules to take effect as soon as they're updated. It provides construction business intelligence by real-time video analytics and autonomous work locks. Currently, the product is being customer tested in one of the largest construction projects in the world, located in Saudi Arabia. Starting in Q4 2023, working with our cloud partners, our smart safety platform, SSP, is being upgraded to a cloud-based global SaaS platform with guaranteed scalability and mission-critical processing capabilities. The SSP covers public safety applications like access control, crowd safety management, fire and smoke real-time alerts, weapons detection, air evasion, fighting, and unattended object detection. Enhanced with our LVM, our SSP is being used with various levels of public safety agencies in the United States and the United Kingdom. But what is multimodal? Multimodal models can process a wide variety of inputs, including text, images, and audio as prompts and converts those prompts into various outputs, not just the source time. I'll give you two examples of what Remark AI does. We recently announced our large multimodal model, an AI-powered aviation safety platform called ASP. Remark AI developed ASP to safely reduce Inspection times associated with aircraft engine maintenance, as well as provide anti-collision tools for aircraft towing and other functionalities, all while improving the collection and reporting of essential performance data in a multimodal format. Given the recent safety inspection failures and supply chain disruptions experienced by Boeing and its airline customers, there has never been a better time to utilize AI to improve safety performance. The ASP uses a large multi-modal model to provide several functionalities, including in three essential components, an engine expansion kit, an aircraft towing anti-collision system, and a large multi-modal model for business intelligence and reporting. The portable aero engine inspection kit detects damage to aircraft engines, including damage such as combustion chamber erosion, tear detection, blade curl, ablation, dimples or pits, lap wear, and misalignment. Upon recognizing such damage, the AIK evaluates it for severity and automatically generates an inspection report with detailed images, thereby reducing staff workload and inspection times from days to approximately 30 minutes, allowing for real-time pre-flight inspections before takeoff. The Towing Anti-Collision System, TACS, provides aircraft towing operators with functionalities that look for, warn of, and thereby potentially prevent accidents. The system helps those towing airplanes to avoid obstacles by utilizing the cockpit string to provide a real-time, three-dimensional view of the aircraft's travel path, a dynamic environmental status, and potential collision warnings. The functionalities include collision prediction, speed monitoring, and detection of abnormal driving, and an incorporated slide arm to provide comprehensive coverage under all weather conditions. The Aviation Industry Business Intelligence Reporter is a powerful AI system that analyzes massive business data, files, photos, videos, and other files, and then provides valuable business intelligence. The system allows for data processing, predictive analysis, semantic search, and statistical analysis. and it provides intelligent recommendations that allow for process improvement in many aspects of aviation operations. We've already launched a similar platform like this, but for trains, subways, and railways, and we are soon to launch our Remark AI multimodal safety platform for marine and sea safety as we have significant potential customers in the Middle East and South America. Our pipeline continues to be very strong, and is only getting stronger. With our recent win at the Clark County School District, which is top five largest in the United States, it has set us up well for other large school districts across the U.S. In addition, currently, as we are targeting the U.S. domestic security market, there are 100 million domestic security cameras and 200 million global cameras, excluding China, that do not provide intelligence or AI real-time analytics. Assuming a base case of $1,200 annual stream per camera to empower it with AI analytics, this represents a $360 billion total addressable market. With that, we believe we are well positioned and in the final stages to win a large security contract with over 100 migrant centers. Walking you through the numbers, to show how quickly this business can sail in just this industry alone. We generally sell six solutions, which we charge $100 per solution. One, violence detection. Two, facial recognition. Three, weapons detection. Four, slip, trip, and fall. Five, fire and smoke detection. And six, intrusion detection. So if you multiply $100 times six, that's $600 a month times 12. which equates to $7,200 a year. $7,200 a year times 100 cameras, which is the typical amount of cameras located in each migrant center, equates to $722,000 a year. And with 100 migrant centers, that equates to $72 million a year. This is an exciting day here. As the revenue is all software, it does not include our harder revenue, which still generates approximately 50% gross margin. I hope this is a strong example of what we see for Remark AI in 2024 and beyond.