Huawei eyes Latin America deployment of Ascend AI chips, cloud chief confirms

Huawei is studying whether to make its newest Ascend AI chips available through its cloud and AI services in Latin America, according to Mark Chen, president of Huawei Cloud Latin America. Speaking at the Rio Web Summit, Chen confirmed the exploration but declined to provide a timeline, noting the Ascend 950 family is still in early domestic deployment. The first chip in the series, the 950PR, launched commercially in China in March, and major Chinese AI firms including Alibaba, ByteDance, and DeepSeek are already using it.

Why it matters: This move would project Chinese-designed AI silicon into a region long dominated by US suppliers, advancing the hyperscaler-distribution moat pattern—where infrastructure reach becomes a strategic barrier. Huawei is leveraging its cloud platform as the distribution layer for its own chips, creating a vertically integrated alternative to Nvidia in markets with weaker US export enforcement. Latin America now becomes a contested zone in the broader geopolitical substrate between US and Chinese AI infrastructure.

The ground-level view: Beijing’s self-reliance strategy has a hardware component—the Ascend line—and its cloud operations in Latin America have grown enough that a real deployment is plausible. Huawei’s decision to keep the 950 generation domestic-first suggests supply constraints remain, but the regional push signals long-term intent to build a non-Nvidia compute corridor into the Americas. For enterprises and governments in the region, this could mean genuine alternative access to frontier AI compute, if Huawei executes.