On October 23, ThunderSoft held the ThunderWorld 2026 Technology Conference and Product Launch, releasing AquaDrive AIOS (DiShui AIOS) 3.0. Also unveiled were an industrial embodied AI solution and applications of the TurboX Autonomous Intelligent IoT Platform on devices such as AI glasses.
In recent years, AI large models have gradually become standard in smart cars, but many products still remain at the voice-interaction level. Even when they can understand complex commands, users often still need to operate manually when navigation, vehicle control, or third-party services are involved.
The newly released AquaDrive AIOS 3.0 upgrades its capabilities around three core functions: autonomous generation, autonomous fusion, and autonomous control.
First, autonomous generation. The new system introduces generative HMI technology that dynamically generates function cards and interaction interfaces based on user intent, real-time scenarios, and task priority, rather than being limited to preset layouts.
Second, autonomous fusion. The new system further integrates lane-level navigation, surrounding environment visualization, 360-degree surround view, and automated parking, reducing interface fragmentation across functions. The relevant technology has been adapted to the mass-production platform of the Chery Tiggo 8 and validated in real urban road conditions.
Third, autonomous control. AquaDrive AIOS 3.0 can autonomously plan and execute complex tasks, maintain interaction continuity through cross-session memory, and convert execution results into editable and revocable interaction cards. Some tasks can be completed on the vehicle side without continuous network dependence.
Beyond making the in-vehicle system smarter, the more important change in AquaDrive AIOS 3.0 is enabling AI to truly participate in the orchestration of vehicle functions and services.
Embodied AI is also a key focus.
In ThunderSoft's view, embodied AI is moving from technology validation to system deployment. In industrial scenarios, robots must not only perform complex actions but also handle environmental changes, continuous operations, and multi-robot coordination; relying solely on more powerful AI models is far from sufficient.
This also explains why ThunderSoft does not treat humanoid robots as its sole development direction.
Currently, ThunderSoft and its subsidiary Xiaowu Intelligent have formed a product matrix covering autonomous transport robots, forklift robots, embodied composite robots, and humanoid robots. Transport and forklift robots are already in mass production, composite robots are in the mass-production ramp-up phase, and humanoid robots are still undergoing scenario validation.
Through a unified AIOS foundation, ThunderSoft is also building a cloud-edge-terminal integrated embodied AI platform that supports unified scheduling, task planning, and continuous learning for robots of different brands and form factors.
According to official disclosure, its robotics business has served more than 100 major customers, with a customer repurchase rate of 60%.
At the current stage, the autonomous intelligence products most accessible to ordinary consumers may still be AI glasses.
At the event, ThunderSoft showcased applications of on-device Agents on terminals such as AI glasses around the TurboX Autonomous Intelligent IoT Platform.
Among them, the AI glasses solution uses the Qualcomm Snapdragon AR1 chip, equipped with a 12-megapixel camera, a five-microphone array, and waveguide display technology, with total weight kept under 35 g.
More importantly, built on the Veda OS system foundation, the AI glasses can use low-power always-on perception to proactively provide reminders, translation, navigation, and other services in specific scenarios without waiting for the user to wake it up each time. For example, during cycling, the glasses can provide voice navigation and also assist in automatically generating vlogs.
To support more devices, ThunderSoft also launched the TurboX Autonomous Intelligence Kit, which covers different hardware such as smart rings, smart cameras, AI glasses, and smart boxes through a unified on-device Agent runtime environment, while introducing security mechanisms such as permission approval and task verification.
From smart cars to industrial robots and then to AI glasses, the products and technologies released by ThunderSoft this time share a consistent underlying logic.
As large-model capabilities become increasingly widespread, the key to the actual AI experience lies in which capabilities a system can invoke, which tasks it can execute, and whether it can ensure the entire process is stable and reliable.
This is also the opportunity ThunderSoft hopes to seize as it moves from traditional operating systems to AIOS in the new wave of AI transformation.
Leikeji AI Hardware Group | Editor: Sanqieryishiyi | Executive Producer: Luo Chao