Huawei’s New AI Chip Gains Interest from ByteDance and Alibaba

Mar 28, 2026
Huawei’s 950PR AI chip has completed successful customer testing, with ByteDance and Alibaba planning to place orders as the company prepares for mass production later this year.
Huawei’s New AI Chip Gains Interest from ByteDance and Alibaba

Customer testing of Huawei’s new 950PR artificial intelligence chip has concluded successfully, with major Chinese tech firms including ByteDance and Alibaba Group planning to place orders, Reuters reported. The chip, designed to compete with Nvidia’s hardware in China, is expected to enter mass production next month, with around 750,000 units planned for shipment this year.

The 950PR is more compatible with Nvidia’s CUDA software system than Huawei’s previous Ascend 910C, enabling easier migration of AI models. It will be priced at approximately 50,000 yuan ($6,900) per card, with a premium version featuring faster HBM memory selling for around 70,000 yuan. The chip is optimized for inference workloads, which are increasingly in demand as Chinese companies shift toward deploying AI applications.

In a separate announcement, Huawei held its sixth Global Installer Summit in Dongguan, where it introduced Installer Voice Loop — an AI-powered platform designed to collect feedback from partners and enhance product quality and support. The company highlighted the integration of AI and digital power technologies as part of its broader strategy.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Silicon Brief or Daily AI Brief.

Also, consider following us on social media:

AI Chips & Datacenters AI Brief AI Brief (X)
Free newsletter

Silicon Brief

Weekly coverage of AI hardware developments including chips, GPUs, cloud platforms, and data center technology.

AI Token Webinar

Whitepaper

Tensordyne Napier: What If One Rack Could Do the Work of Nine?

Tensordyne

This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.

Read more

You may also like

Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.