The upgrades cover real-time processing and long-term memory to support expansive hardware deployments.

Additional infrastructure enhancements underscore a concerted push to empower developers with end-to-end AI hardware capabilities.

Tuya is rolling out three core AI application ecosystems—AI Home, AI Robot, and AI Energy—to drive AI living, inviting global developers to collaborate for faster AI adoption and real-world deployment.

New capabilities include slide generation, image/video creation and recognition, and a scenario-based intelligence framework powering use cases like Morning Wake-up, Health Management, and Away from Home Protection.

An enhanced AI foundation enables more natural conversations, deeper reasoning, and proactive planning of tasks rather than mere reactive responses.

Tuya’s leadership signals a bold shift to AI-first hardware and software, with AI-driven growth and a belief that almost all hardware is worth rebuilding around AI.

CEO and co-chairman statements emphasize imminent AI-driven growth and widespread AI-enabled hardware applications across the company’s strategy.

Major upgrades to Tuya’s AI infrastructure were announced, including Physical AI Foundation V2.8, Wukong AI 3.0, the T-RTC network, Physical Action Model (PAM), and OmniMem V2.0 long-term memory system, creating a comprehensive developer toolkit.

Market context shows historically mixed but positive AI-tagged reactions to Tuya’s announcements, with the current upgrade aimed at accelerating adoption and catalyzing developer-led implementations.

Tuya adopts a Day One mindset for integrating AI into hardware, inviting global developers and ecosystem partners to help expand AI across homes, robots, and energy systems.

Tuya introduced self-developed AI hardware technologies, including the PVAD model to reduce false activations, and unveiled a new AI hardware toolkit and self-trained models for the AI Home, AI Robot, and AI Energy ecosystems.

The hardware rollout features PVAD and infrastructure upgrades such as Physical AI Foundation V2.8, Wukong AI 3.0, T-RTC, PAM, and OmniMem V2.0 to support three ecosystems.