en.Wedoany.com Reported - On July 9, Xiaomi announced that its open-source IoT operating system openvela has been repositioned as an "AI hardware intelligent base," and its lightweight application framework is being fully opened to developers for the first time. The installed base of openvela ecosystem devices has exceeded 180 million, growing 80% in less than a year, covering smartwatches, wristbands, earphones, smart speakers, AI glasses, and other types of smart hardware.
The core role of openvela is to provide a unified software base for fragmented IoT devices. IoT hardware categories are complex, with significant differences in chips, screens, sensors, communication modules, power consumption conditions, and interaction methods. If each type of device requires separate system and application framework adaptation, development efficiency would be consumed by extensive underlying compatibility work. Built on the NuttX kernel, openvela provides service frameworks for graphics, connectivity, application management, and AI, enabling low-power devices, screen-equipped devices, wearables, home devices, and new AI hardware to share a lightweight system capability.
The installed base exceeding 180 million indicates that openvela has moved from an open-source project to a large-scale device base stage. For an operating system, device scale not only represents shipment volume but also signifies that system stability, power consumption performance, driver adaptation, connectivity protocols, application frameworks, and development toolchains have been validated through extensive real-world scenarios. According to previous information from the OpenAtom Foundation, Xiaomi Vela technology has achieved large-scale application in categories such as watches, wristbands, earphones, smart speakers, and AI glasses, and has become the lightweight key technology base for Xiaomi HyperOS.
The opening of the lightweight application framework is a significant move for openvela towards the developer ecosystem. The lightweight application framework is designed for resource-constrained IoT devices, allowing developers to create interactive interfaces, device functions, and AI applications without rewriting extensive underlying code. For hardware such as smart wearables, earphones, smart homes, children's devices, health monitoring devices, and small robots, the lightweight application framework can lower the barrier to application development, enabling more devices to have the capability to update, expand, and run lightweight applications.
After upgrading to an AI hardware intelligent base, openvela's focus has shifted from "connecting devices" to "enabling devices with on-device intelligent capabilities." Its technical system includes a lightweight inference framework, unified AI interfaces, end-cloud collaboration capabilities, and debugging and performance analysis tools. It can support devices in running some AI capabilities locally, as well as calling edge and cloud models to complete more complex tasks. For AI hardware, the system base needs to simultaneously handle sensor input, low-power operation, network communication, model invocation, privacy protection, and application interaction, and cannot rely solely on a single large model interface to deliver the product experience.
At the end of 2024, Xiaomi launched Xiaomi Vela as the open-source openvela to the industry, using the Apache 2.0 license on GitHub and Gitee. After open-sourcing, openvela is no longer just a system capability within the Xiaomi ecosystem but an AIoT operating system project that can be used by chip manufacturers, module manufacturers, hardware teams, and application developers.
openvela's current focus is on the 180 million device scale, the AI hardware intelligent base positioning, the opening of the lightweight application framework, on-device AI capabilities, IoT connectivity protocols, and developer toolchains. For smart hardware companies, openvela provides an integrated software infrastructure covering the underlying system, device connectivity, application operation, and AI capability invocation.
