As edge AI reaches a pivotal threshold where technology concepts are transitioning into large-scale commercial deployment, the household environment is emerging as the central battleground for reshaping the intelligent ecosystem value chain. During the IBC 2026 trade show held in Amsterdam from September 11 to 14, SDMC (00901) spotlighted its practical pathway for converting AI innovations from foundational technical capabilities into deployable AI Home services, with a core focus on edge-cloud collaborative intelligence, edge AI terminals, ecosystem partnerships, and frontier explorations in Physical AI. This presentation not only delivered an impressive exhibition performance but also marked a critical inflection in the company's acceleration from a conventional terminal device provider toward a household Physical AI ecosystem builder equipped with integrated "cloud-edge-device" synergy, with multidimensional breakthroughs in product deployment, ecosystem positioning, and capital market perception outlining a clear blueprint for next-generation smart home commercialization.

The IBC 2026 "Deployment Report": Edge-Cloud Synergy Propels AI Home into Operator-Grade Practical Application This year's IBC 2026 signals that the global home terminal industry is enduring a profound value reassessment: AI is advancing from a mere "technology selling point" to a "core infrastructure" that restructures operator business models. With Agentic AI TV and On-Device AI Agent becoming industry consensus, household-side computing power and edge intelligence are no longer optional enhancements but crucial differentiators for building competitive moats. Those who complete the "cloud-edge-device" collaborative ecosystem loop first will secure pricing power in next-generation smart home commercialization. SDMC's breakthrough at this exhibition represents a precise strategic move grounded in this overarching logic. Its showcased AI Station household-side shared computing platform, built on the NVIDIA Jetson T5000, delivers up to 2,070 TFLOPS (Sparse FP4) of AI computing power, equipped with 128GB of LPDDR5X memory, consolidating previously fragmented terminal-side computing demands into a unified pooled resource that provides ample computational capacity for multi-model parallel processing and complex AI workloads. The Cedar multimodal model technology preview running on this platform, developed on Google's Gemma foundation, integrates voice, text, vision, device status, and environmental data to create a unified understanding of household context, with the Cedar Agent managing cross-device and cross-service task coordination and execution, while Gemini supplements edge intelligence with cloud-side reasoning for more complex inference needs. This architecture successfully establishes a complete closed loop spanning "cloud-side reasoning, edge-side orchestration, and terminal-side execution." On the terminal front, SDMC has partnered with Amlogic to combine chip and NPU capabilities with system integration and AI application adaptation, with AI smart screens integrating environmental sensing, AI assistants, and multi-device voice coordination to evolve from traditional control interfaces into more natural AI Home interaction gateways, while AI TV boxes leverage NPU computing power to enhance audiovisual entertainment experiences and support local AI services alongside home device collaboration. From the computing foundation to the intelligent agent hub and tangible terminal carriers, SDMC's "computing foundation + intelligent agent hub + terminal carrier" framework aligns with industry demands for household infrastructure to a certain degree, presenting a deliverable, commercially viable, and scalable operator-grade AI Home product suite rather than a conceptual technology demonstration, serving as a significant vehicle for operators transitioning from "connectivity monetization" to "intelligence monetization."

GCP "Agentic Physical AI" Ecosystem Endorsement: Why SDMC Is an Irreplaceable Partner in the Physical AI Wave The year 2026 is becoming a watershed moment where AI crosses digital boundaries to reshape the physical world. The core competitive barrier in Physical AI has escalated from singular model parameter competition to a comprehensive systems engineering contest spanning "models, tools, scenario deployment, and commercialization." In this arena, the household domain, characterized by ambiguous commands and highly fragmented devices, stands as both the most challenging and most commercially valuable proving ground. SDMC's inclusion in Google Cloud's "Agentic Physical AI" proactive acceleration program reflects a strategic acknowledgment of its "scenario deployment and engineering delivery capabilities" by an industry titan, rather than a mere technical alliance. Over the past nine years, from securing one of the world's first Android TV certifications, migrating XMediaTV cloud services to the cloud, achieving the first global Google TV projector product certification, to now joining the core Physical AI acceleration program, SDMC's ecosystem evolution is not a singular breakthrough but a systematic closed loop spanning hardware terminals, cloud platforms, and AI Agent productization. With the Cedar model developed on Google's Gemma foundation and the Cedar Agent integrated with Gemini for enhanced cloud-side reasoning, the two parties have achieved deep technological interlocking at the foundational level. This long-term "ecosystem co-creation" model not only reduces trial-and-error costs for frontier technologies but also continuously strengthens SDMC's system integration capabilities and AI Home productization expertise. At this IBC, SDMC further showcased its exploration directions for Physical AI in household settings, ranging from smart home control, device coordination, and autonomous cleaning to emerging applications with more tangible interactive capabilities such as robotic arms. Building on its existing edge-cloud collaborative AI Home technology framework, the company is extending AI capabilities from "understanding and decision-making" to "action and execution" in the real world. This also indicates the company is accelerating its critical transformation from a traditional terminal equipment provider to a globally oriented household Physical AI ecosystem architect.

Stock Connect Inclusion and Performance "Value Resonance": Capital Markets Begin Repricing When industrial trends align with fundamental performance, capital markets typically respond with early pricing signals. SDMC's recent inclusion as a Stock Connect target represents a significant liquidity milestone that not only broadens investment channels but also provides a fresh lens to reassess the company: it is currently in a pivotal transition toward becoming a "household Physical AI ecosystem builder." Financial figures reveal that SDMC achieved revenue of RMB 2.095 billion in the first half of 2026, representing 36.3% year-on-year growth, with period profit reaching RMB 181 million, up 75.9% year-on-year, while gross margin improved by 3.5 percentage points to 21.1%. This growth represents a natural extension driven by regional demand releases and product portfolio optimization built upon existing operator channels and installed terminal base, validating to a certain degree the business logic of "uncovering incremental value from existing stock." As the world's largest Android TV smart terminal provider, SDMC does not need to build its ecosystem from scratch; as Physical AI penetrates deeper into more complex household scenarios such as security, elderly care, and energy conservation, alongside the continuous refinement of its edge-cloud collaborative architecture, the company can unlock AI value across its massive installed device base. This IBC 2026 showcase of the edge-cloud collaborative product matrix, the ecosystem scarcity afforded by the GCP acceleration program, and the liquidity improvements following Stock Connect inclusion collectively indicate that SDMC is undergoing a value transformation from a terminal hardware manufacturer to a household Physical AI ecosystem builder. Within the Physical AI industrial trend, SDMC acts as both a pragmatic implementer of technology deployment and a potential observation target for capital markets in the AI Home arena, offering a combination of growth certainty and valuation flexibility.