The development marks a notable shift for Huawei, whose earlier Ascend chips struggled to achieve widespread adoption among private-sector firms. The company’s latest processor has performed strongly in customer testing, sources said, signalling improved market confidence in its AI hardware capabilities.

A key factor behind the increased interest is enhanced compatibility with software ecosystems historically dominated by Nvidia. This has helped address longstanding developer concerns around integration and usability, while also improving performance in real-world deployments.

The chip is understood to be optimised for inference workloads, reflecting a broader shift across the AI sector from model training to deployment at scale. This aligns with growing enterprise demand for infrastructure capable of supporting production-grade AI services, rather than purely experimental or research-focused environments.

Huawei is targeting shipments of approximately 750,000 units in 2026, with production expected to ramp up and broader deliveries anticipated in the second half of the year, according to sources.

The reported orders come against the backdrop of tightening US export restrictions on advanced AI chips, which have limited access to leading-edge hardware from US suppliers. This has created an opening for domestic alternatives, as Chinese technology firms seek greater control over critical infrastructure components.

If confirmed, commitments from ByteDance and Alibaba would represent a significant endorsement of Huawei’s AI strategy and highlight accelerating momentum behind China’s push for semiconductor self-sufficiency, particularly as AI-driven workloads continue to scale across cloud and digital infrastructure environments.

RELATED STORIES

Telkom and Huawei: Building the foundation for future-ready networks in the AI era

AI chip demand continues to strain big tech supply chains