Huawei set to ship 910C AI chips at scale, signaling shift in global AI supply chain

“From a performance standpoint, Nvidia’s new-generation chips — such as the B200 and the upcoming B300 Ultra, based on TSMC’s 4nm process and equipped with advanced HBM3/3E memory — have significantly widened the gap compared to Huawei’s 910C, which is likely built on SMIC’s N+2 7nm process (effectively 14nm) and lacks advanced HBM memory,” said Neil Shah, partner and co-founder at Counterpoint Research.

Shah noted that while Huawei’s chip may theoretically match the older Nvidia A100 or H100 in some tasks, it would require more power and heavy software optimization to handle diverse AI workloads.

“Global adoption of Huawei’s 910C is also hindered by limited developer support, ecosystem maturity, and integration challenges,” said Manish Rawat, semiconductor analyst at Techinsights. “However, it presents a viable alternative to Nvidia’s chips for Chinese enterprises or those affected by geopolitical constraints, especially as US export controls limit access to advanced Nvidia GPUs.”

Implications on enterprise AI adoption

Despite not being the best in the market, the 910C could still prove viable for many enterprise- and hyperscale-AI use cases.

It may take longer to train models compared to US-designed chips, but for many, it’s an acceptable trade-off given current geopolitical and supply chain risks.

“For enterprises – particularly those operating in or sourcing from China – it presents a credible alternative in the face of tightening US export controls, while supporting domestic innovation ecosystems like DeepSeek and reducing dependence on foreign technologies,” said Prabhu Ram, VP of the industry research group at Cybermedia Research.  “Although Nvidia maintains an edge in software maturity and energy efficiency, Huawei’s progress reflects China’s growing strength in competing at the forefront of AI hardware in the emerging AI era.”

If companies operating in or sourcing from China adopt Huawei’s ecosystem, it could significantly influence their procurement strategies, vendor selection, and technology evaluations, driving greater alignment with Chinese technologies.