Sources: soaring appetite for its AI chips and manufacturing constraints have forced Huawei to prioritise AI and slow production for its premium Mate 60 phones
Context & Ripple Effects
Huawei’s manufacturing limits were already affecting products beyond handsets: some Chinese automakers had delayed flagship deliveries amid issues and part shortages tied to Huawei’s MDC 810 computing unit, a sign that constrained compute supply was reaching vehicle programs. This report shows the same allocation pressure moving into Huawei’s premium-phone business.
The later difficulty in scaling Ascend server-chip output suggests that shifting priority to AI did not by itself remove the supply constraint; it instead concentrated Huawei’s scarce manufacturing capacity on a strategic product line.
First-order effects
- Huawei slows Mate 60 output, reducing near-term supply of its premium phones while reserving constrained manufacturing resources for AI chips.
- AI-chip customers become the immediate beneficiaries of Huawei’s allocation decision, while handset channel partners and buyers face tighter availability of the affected model.
Second-order effects
- The trade-off can spill into other Huawei hardware programs: the earlier MDC 810-related delivery delays indicate that constrained computing components can disrupt automakers’ delivery schedules as well as consumer devices.
- Prioritizing AI chips does not guarantee faster fulfillment if fabrication remains the binding constraint; Huawei later struggled to ramp Ascend server-chip production, keeping supply pressure on customers seeking an alternative source of AI compute.
Third-order effects
- If these allocation choices persist, Huawei’s product portfolio will be increasingly shaped by access to manufacturing capacity rather than end-market demand alone, with AI infrastructure taking precedence over some device volumes.
- The episode points to a longer rebuilding cycle for Huawei’s chip business under restrictions: later reports of improving Ascend yields made the chips profitable for the first time, but production scale remains the decisive constraint.
The trend: AI-compute demand is turning semiconductor capacity into a cross-product allocation decision, forcing integrated hardware companies to choose between infrastructure chips and finished devices.