23 Daily — According to 36kr.com, on November 6, 2026, Midea Group unveiled its latest achievement in whole-home intelligence at the 31st Technology Month conference held in Shunde, Foshan. Midea introduced an innovative Agent to Agent (A2A) intelligent dialogue model that overcomes traditional cross-brand appliance interconnection bottlenecks. This model enables autonomous collaboration among hardware devices of different brands and categories, reducing ecosystem cooperation costs and enhancing user experience, thus realizing a new paradigm for cross-brand connectivity in smart homes.
Midea launches A2A intelligent dialogue model to solve cross-brand interconnection challenges
The smart home industry has long been constrained by three systemic issues: data silos among heterogeneous devices, a gap between intelligent interaction and user expectations, and high engineering barriers for system deployment and debugging. Dr. Shang Zhe, General Manager of Midea Whole-home Intelligence, stated that these challenges stem from the industry's lack of a low-cost and widely replicable cross-brand and cross-category collaboration technology standard. Current mainstream solutions relying on cloud-to-cloud point-to-point API integration face exponential increases in coupling costs as ecosystems grow, limiting their scalability for massive device interconnections.
To address this, Midea proposed the Agent to Agent (A2A) intelligent dialogue model, where two AI agents communicate directly. When one agent receives a user request, it conveys it to the partner's agent, which autonomously determines whether the task is feasible and dispatches its devices to complete the action, then returns the result. Sensitive operations require user confirmation. This approach requires only standard communication protocols between AI agents, avoiding one-by-one hardware adaptation costs.
MevoX intelligent agent as the technological foundation with deep reasoning and memory
Supporting Midea’s A2A architecture is its self-developed home intelligent agent MevoX. Unlike typical voice assistants that respond only to explicit commands, MevoX boasts two core capabilities: deep reasoning to infer true user intent from fuzzy or incomplete input by leveraging spatial context and device attributes; and long-term memory that continuously adapts personalized settings based on user habits to avoid repetitive resets.
Leveraging years of smart appliance development and a comprehensive hardware portfolio, MevoX can create a complete cooperation chain from intent understanding through device orchestration in real home environments.
Cross-ecosystem synergy and vehicle-home interconnection applications
Midea has completed A2A integration with Chery iCAR, connecting iCAR models to Midea’s Meiju platform. Vehicle owners can issue natural language commands inside the car to check household consumable levels, with future plans to trigger voice-activated reordering. This surpasses traditional simple car-to-home switch controls, achieving cross-system business-level task collaboration.
Additionally, Midea continues partnerships with leading smartphone brands such as Huawei, vivo, OPPO, Honor, as well as automakers including BYD, NIO, Changan, and Chery, promoting deep person-vehicle-home ecosystem linkage. As the first in the home appliance industry to enable interoperable family intelligent agents via A2A, Midea marks a substantial breakthrough in ecosystem cooperation technology.
Editor's take: A2A technology opens new opportunities for cross-brand smart home innovation and validation methods
Midea’s Agent to Agent intelligent dialogue model breaks the scale limitations of traditional API point-to-point integration, offering a clear technological pathway for cross-brand smart home appliance connectivity. For entrepreneurs, this technology reduces barriers and costs for ecosystem partnerships, fostering more open and diverse smart home solutions.
Verifying this model's feasibility can focus on two aspects: firstly, conduct cross-brand appliance pilots integrating through A2A to assess collaboration efficiency and user experience improvements; secondly, explore integrating with other ecosystems such as vehicles to test cross-lifestyle scenario task collaboration capabilities. Successful validation could help startups differentiate within smart home ecosystems and drive large-scale adoption and deep cooperation of intelligent devices industry-wide.