Cloud-Edge Collaboration Implemented, Ecosystem Positioning Upgraded: SDMC (00901) Moves from AI Home Practice to Home Physical AI

At the historic inflection point where edge AI is moving from a technological concept to large-scale commercialization, the home scenario is becoming a core battleground for reshaping the intelligent ecosystem value chain. From September 11 to 14, the global broadcasting and media industry's annual event IBC was held in Amsterdam, the Netherlands. SDMC (00901), centered on "cloud-edge collaborative intelligence, edge AI terminals, ecosystem cooperation, and frontier exploration of physical AI," comprehensively demonstrated the practical path of its AI innovation transforming from underlying technological capabilities into deployable AI Home services. This is not only an impressive exhibition scorecard, but also marks SDMC's accelerated key leap from a traditional terminal product provider to a builder of a home Physical AI ecosystem with "cloud-edge-device" collaborative capabilities. Its multi-dimensional breakthroughs in product deployment, ecosystem positioning, and capital market perception shifts are outlining a clear blueprint for the commercialization of next-generation smart homes for the industry. **IBC2026's "Deployment Scorecard": Cloud-Edge Collaboration Empowers AI Home to Enter Carrier-Grade Product Combat** [The author] believes that this year's IBC 2026 marks a profound value revaluation undergoing in the global home terminal industry: AI is leaping from a mere "technical selling point" to "core infrastructure" that reconstructs carriers' business models. Against the backdrop of Agentic AI TV and On-Device AI Agent becoming industry consensus, home-side computing power and edge intelligence are no longer optional add-ons, but the key to building differentiated competitive barriers. Whoever can first complete the "cloud-edge-device" collaborative ecosystem loop will be able to grasp the pricing power of next-generation smart home commercialization. SDMC's breakthrough at this year's exhibition is precisely a targeted move based on this macro logic. Its showcased AI Station home-side shared computing power platform, built on NVIDIA Jetson T5000, provides up to 2,070 TFLOPS (Sparse FP4) AI computing power, equipped with 128GB LPDDR5X memory, integrating computing power demands originally scattered across various terminals into a unified pooled resource, providing sufficient computing power foundation for multi-model parallelism and complex AI workloads. The Cedar multimodal model technology preview running on this platform is built on Google Gemma, capable of fusing voice, text, vision, device status, and environmental information to form a unified understanding of home context; Cedar Agent is responsible for cross-device, cross-service task collaboration and execution. For more complex reasoning needs, Gemini supplements edge intelligence capabilities on the cloud side. This architecture successfully opens up the complete loop of "cloud inference, edge orchestration, terminal execution." At the terminal level, SDMC has partnered with Amlogic to combine chip and NPU capabilities with system integration and AI application adaptation. The AI smart screen integrates environmental perception, AI assistant, and multi-device voice collaboration capabilities, upgrading from a traditional control interface to a more natural AI Home interaction entry point; the AI TV box enhances audio-visual entertainment experience based on NPU computing power, and supports local AI services and home device collaboration. From computing power foundation to agent hub, and then to tangible terminal carriers, SDMC's "computing power foundation + agent hub + terminal carrier" system to a certain extent meets the industry's demand for home scenario infrastructure. At the same time, what it presents is a deliverable, commercializable, and scalable carrier-grade AI Home product system, rather than a technology demonstration (staying at the conceptual level). It is an important vehicle for carriers to extend from "connectivity monetization" to "intelligence monetization." **GCP "Dynamic Everything" "Ecosystem Endorsement": Why SDMC Is an Irreplaceable Partner in the Physical AI Wave** 2026 is becoming a watershed moment for AI to cross digital boundaries and reshape the physical world. The core barrier of Physical AI has been comprehensively elevated from a single model parameter competition to a systems engineering contest of "model + tools + scenario deployment + commercialization." In this competition, the home scenario, due to the ambiguity of instructions and the high fragmentation of devices, has become the most challenging yet most commercially valuable touchstone. SDMC's selection into Google Cloud's "Dynamic Everything" proactive Physical AI acceleration program its deeper logic is not a simple technical alliance, but a strategic recognition by an industry giant of its "scenario deployment and engineering delivery capabilities." Over the past nine years, from being among the world's first batch of Android TV certifications and XMediaTV cloud service migration to the cloud, to the world's first Google TV projector product certification, and now to selection into the Physical AI core acceleration program, SDMC's ecosystem evolution is not a single-point breakthrough, but a systematic closed loop from hardware terminals and cloud platforms to AI Agent productization. As the Cedar model is developed based on Google Gemma, and Cedar Agent integrates with Gemini to enhance cloud inference capabilities, the two parties have formed a deep interlocking at the underlying technology level. This long-term "ecosystem co-creation" model not only reduces the trial-and-error costs of frontier technologies, but also continuously strengthens SDMC's own system integration capabilities and AI Home productization capabilities. At this year's IBC, SDMC further demonstrated the exploration direction of Physical AI in home scenarios. From smart home control, device collaboration, and autonomous cleaning, to emerging application scenarios with more physical interaction capabilities such as robotic arms. Leveraging its existing cloud-edge collaborative AI Home technology system, it extends AI capabilities from "understanding and decision-making" to "action and execution" in the real world. This also means the company is accelerating its key leap from a traditional terminal product provider to a builder of a home Physical AI ecosystem with a global perspective. **"Value Resonance" Between Stock Connect Inclusion and Performance: Capital Markets Begin Repricing** When industry trends and fundamentals resonate, capital markets often take the lead in pricing. SDMC's recent formal inclusion in the Stock Connect program this major breakthrough at the liquidity level not only broadens investment channels, but also provides a window to re-examine SDMC: it is in a critical leap period of transforming into a "home Physical AI ecosystem builder." From financial data, SDMC achieved revenue of RMB 2.095 billion in the first half of 2026, a year-on-year increase of 36.3%; profit for the period reached RMB 181 million, a year-on-year increase of 75.9%; gross margin increased by 3.5 percentage points year-on-year to 21.1%. This growth is a natural extension achieved through regional demand release and product portfolio optimization on the basis of existing carrier channels and installed terminal base, which to a certain extent validates the business logic of "mining incremental value from existing stock." As the world's largest Android TV smart terminal supplier, SDMC does not need to build an ecosystem from scratch. As Physical AI deeply penetrates more complex scenarios such as home security, companionship, and energy saving, and as the cloud-edge collaborative architecture continues to improve, it can release AI value on the foundation of its massive installed device base. [The author] believes that whether it is the cloud-edge collaborative product matrix showcased at IBC 2026, the ecosystem scarcity brought by the GCP acceleration program, or the liquidity improvement after Stock Connect inclusion, all indicate that SDMC is undergoing a value reshaping from a terminal hardware manufacturer to a home Physical AI ecosystem builder. In the industry trend of Physical AI, SDMC is both a pragmatic implementer of technology deployment and may also become an observation target in the AI Home track that combines growth certainty with valuation elasticity for capital markets.