A Key Leap in the Era of Edge AI: The AI Home Strategy and Value Reassessment Logic of SDMC (00901)

When computing power density breaks through the cost threshold, and privacy compliance becomes a necessity in digital life, edge AI is reaching a historic turning point from a technological concept to scaled commercial use. In this wave of global industrial restructuring, home scenarios are becoming one of the key application domains for edge AI due to their rigid demands for real-time capability, privacy, and contextual collaboration. However, the smart home industry has long been trapped in an island of point intelligence, lacking a system-level brain that can integrate multiple protocols, cross-brands, and all scenarios. In this critical window, whoever can build an ecological barrier based on cloud-edge-end collaboration will gain the power to define the next generation of household intelligence. SDMC (00901), leveraging its unique operator channel barriers and early advantages in the Google ecosystem, is making a leap towards becoming a home AI ecosystem builder with its 1+2+X strategy. Recently, SDMC was selected by Hang Seng Indexes Company to be included as a constituent stock of the Hang Seng Composite Index, effective from September 7, 2026, marking a significant step in the company's capital market development. According to the interim results announcement released by SDMC on August 20, 2026, the company achieved revenue of 2.095 billion RMB in the first half of the year, a year-on-year increase of 36.3%; net profit for the period was 181 million RMB, up 75.9% year-on-year; and gross profit margin improved by 3.5 percentage points to 21.1%. Of this, revenue from digital video equipment was 1.734 billion RMB, a year-on-year increase of 40.5%, accounting for 82.7% of total revenue; overseas revenue was 1.947 billion RMB, accounting for 92.9%. As a global leader in AI Home comprehensive solutions, SDMC has continued to increase its investment in R&D during the reporting period, focusing on key areas such as the family AI entity Cedar, edge computing, smart terminals, and underlying storage collaboration, promoting the deep integration of household solutions in the cloud with home edge, establishing a complete technological architecture and commercialization pathway. The commercialization turning point for edge AI has arrived, with home scenarios becoming the core deployment battlefield. Driven by the dual forces of technological iteration and market demand, edge AI is accelerating into a cycle of large-scale commercial use. The formation of this turning point stems from the concentrated maturity of the triplet DRIVE of computing cost, privacy compliance, and real-time response. The improvement of AI chip computing power density and cost reduction has made it economically viable for home terminals to integrate 1-5 TOPS of local computing power, supporting high-frequency tasks such as voice recognition and image analysis to run locally. Due to the specificity of home scenarios, the implementation of purely cloud-based AI applications faces multiple challenges, while edge AI processes household data locally and, in combination with the flagship intelligent experience of cloud AI, achieves an optimal balance of experience and security. Applications like security alarms and fall detection require millisecond-level response, making real-time decision-making capabilities at the edge an indispensable infrastructure. In terms of market space, the high growth potential of the edge AI sector is remarkable. According to data from Fortune Business Insights, the global edge AI market is expected to reach 35.81 billion USD by 2025, and it is projected to climb to 385.89 billion USD (equivalent to 2.6171 trillion RMB) by 2034, with a compound annual growth rate of 29.9% from 2026 to 2034, far exceeding the growth rate of the cloud computing market, as capital and technological resources accelerate towards edge intelligence. The industrial chain has a three-layer structure, and edge AI solutions aimed at smart homes are one of the worthy areas of focus, as they can address the collaborative challenges of "multiple devices, multiple brands, and multiple protocols." Currently, smart homes are transitioning from point intelligence to scenario collaboration, with the data island model controlled by independent apps becoming a major bottleneck to experience upgrades. The next core phase is to establish a family AI hub that possesses a closed loop of perception, decision-making, and execution, which must meet four conditions: localized operation, multi-protocol compatibility, cross-brand interoperability, and AI native functionality. SDMC's AI Home strategy accurately targets this strategic gap by integrating edge AI with intelligent connectivity, reshaping the value chain of the home intelligence ecosystem. The AI Home 1+2+X framework: Building ecological moats through cloud-edge-end collaboration. The strategic enhancement at SDMC is fundamentally based on its global leading position in the Android TV terminal sector, completing the value leap from software and hardware product solutions to comprehensive AI Home solutions. The core support of this transformation lies in two rare assets: the first is its B2B2C channel network, which covers more than 80 countries and regions as of June 30, 2026, serving over 300 operators and retail brand clients, reaching hundreds of millions of household users and establishing the foundation for large-scale commercialization; the second is its deep binding with the Google ecosystem, having laid out its Google TV ecosystem and AI ecosystem in advance, which provides a technological first-mover advantage for its AI strategy. Accordingly, SDMCs AI Home 1+2+X framework is a system-level solution with cloud-edge-end collaborative capabilities. Its core value lies in creating ecological moats through technological integration rather than relying solely on product innovation. 1 refers to the household AI entity Cedar, which serves as the self-developed brain of the family. As of June 2026, Cedar has further integrated the Gemini 3.5 large model, continuously enhancing capabilities such as visual perception, proactive service, identity recognition, and personalized memory. Its hybrid cloud-edge deployment architecture allows some capabilities to be delegated to AI smart TV boxes and other terminals, supporting network self-maintenance and local offline AI customer service. During the reporting period, Cedar completed its R&D and is currently in customer testing and commercialization stages. The strategic significance of Cedar is in its evolution from an interactive entry point to the decision-making hub of home scenarios. 2 represents the two foundational aspects of intelligent connectivity and edge AI. The intelligent connectivity layer is centered around the Wi-Fi 8 AI optical fiber gateway, a product that integrates tri-band Wi-Fi 8 and 10G optical fiber access while also incorporating an AI NPU, enabling intelligent traffic scheduling, Wi-Fi Sensing wireless perception, and local AI inference collaboration. On the software level, the AI optical fiber gateway, combined with the remote management and application lifecycle management capabilities of prplOS, can support operators in large-scale device management and continuous functional upgrades, turning it into a programmable edge computing node. It is worth highlighting that in addition to the gateway product, SDMC's AI smart TV box is also a significant carrier of the family edge computing system. SDMC was the first to position the AI smart TV box equipped with computing power as the central hub of the operator's AI Home, further integrating the AI entity with household network devices, showing significant phased results at MWC 2026. In the future, the company will continue to develop edge AI devices such as household computing centers, enhancing the family edge computing system. It can be said that SDMC's AI smart TV box has evolved from a traditional streaming media playback terminal to a household edge computing node that integrates local computing power, runs the Cedar AI entity, and collaborates with devices throughout the home. X refers to the continuously expanding terminal ecological matrix, which has grown from streaming media boxes to include smart projectors, cameras, household service Siasun Robot & Automation, and smart glasses. These terminals deeply collaborate with Cedar and central household devices to achieve scenario-based interactivity. When cameras or other sensing devices detect abnormal events, the AI Home system can make judgments based on household context and coordinate compatible devices to complete alerts and responses, thereby forming a closed loop from perception to decision-making to execution. From leading hardware boxes to operational household AI hubs: the logic of open ecological connectors and value reassessment. More importantly, SDMC's product innovation path fundamentally completes the value reshaping from functional terminals to an AI hub. The Wi-Fi 8 AI optical fiber gateway has evolved into the core node of household edge computing, featuring built-in computing power (integrating AI NPU, decision latency in the millisecond range), perceptual enhancement (Wi-Fi Sensing for non-intrusive monitoring), and programmability (supporting dynamic application deployment based on prplOS), transforming the gateway from a network connection device into a sustainable value-adding operational platform. In parallel, the AI smart TV box, leveraging its extensive household coverage and high-frequency user interaction characteristics, has become a carrier of the Cedar AI entityable to serve as both an independent AI customer service and scenario automation engine, while also forming computational collaboration with the gateway to achieve dynamic cooperation and load balancing for edge-side tasks. Cedar is extending from a single device capability to a cross-device AI service layer, providing value-added services such as network self-maintenance, local offline AI customer service, and scenario automation, thereby shifting the operational maintenance model from human-driven to AI-driven. SDMCs strategic insight lies in not pursuing a closed ecosystem but becoming a connector of mainstream ecosystems through a dual-entry strategy. On one hand, by being compatible with the Matter protocol, Cedar can control cross-brand Matter-certified devices, bridging different ecosystems and smart hardware from various brands to achieve inter-device connectivity and collaborative interaction; on the other hand, by deeply integrating into Googles business ecosystem and capability ecosystem, it provides customers with last mile delivery and offers users unique scenario definitions. This dual-entry strategy enables SDMC to occupy an irreplaceable ecological position in home scenarios. In conclusion, the reason SDMC's strategic landing possesses high rarity lies primarily in its comprehensive assembly of three core elements that are challenging to obtain simultaneously in the home edge AI sector: first, a B2B2C channel barrier that serves over 300 operators and retail brand customers, establishing a moat for scaled monetization; second, comprehensive cooperation with the Google ecosystem, securing a core ecological position for user scenarios; third, the 1+2+X system-level architecture capability, achieving a qualitative transformation from simply stacked products to whole-house intelligent collaboration. At this critical threshold where edge AI transitions from concept to implementation, SDMC is transforming household scenarios into operational, value-added digital assets through the Cedar AI entity and edge computing foundation. Understanding this shift from selling terminals to selling scenarios is the underlying logic for evaluating its investment value and reassessment potential post-listing.