Huawei Open‑Sources MindSpore AI Framework to Expand Global Developer Community

Huawei announced that its MindSpore AI framework is now open‑source, releasing the code under an Apache‑compatible license to allow global developers to download, modify and contribute to the platform and to foster community‑driven innovation.

MindSpore is designed for end‑to‑end model development, supporting training on Ascend chips and offering tools for inference optimization, aiming to reduce energy use and speed up deployment on edge devices and supports distributed training across multiple nodes.

Developers familiar with PyTorch or TensorFlow may find MindSpore’s API similar, while recent benchmarks show PyTorch holds 85 % research share and a 10 % speed advantage, prompting Huawei to position MindSpore as a low‑power alternative and aims to close the gap.

Industry analysts expect the open‑source move to boost MindSpore adoption in China’s AI ecosystem and could encourage more cross‑border collaborations as companies seek alternatives to US‑dominant frameworks in both cloud and edge environments.