Huawei releases 505B-parameter openPangu model trained on Ascend chips
The 512K-context release tests whether Huawei can turn its chip, model and inference stack into a credible alternative to Nvidia-centered AI.
By Ryan Merket · Published
Primary source: X - Poe Zhao
Why it matters
Huawei is using an open-weight model to validate a vertically integrated AI stack built around Ascend instead of Nvidia GPUs. Adoption will depend on tooling, reproducibility and a license that bars use in the EU.
Huawei has released openPangu 2.0 Pro, a Mixture-of-Experts language model that Huawei says was trained entirely on its Ascend AI processors, putting weights from one of the largest models built on a non-Nvidia compute stack into developers' hands.…
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.