Huawei chips are one generation behind U.S. but firm finding workarounds, CEO says

Huawei CEO Ren Zhengfei stated that the company's chips lag U.S. peers by one generation, but Huawei is mitigating this through cluster computing and investing heavily in research, allocating 180 billion yuan annually with a focus on compound chips. Ren downplayed Huawei's capabilities amid U.S. tech restrictions, noting that while single-chip performance trails the U.S., Huawei is using alternative methods to achieve practical results, as evidenced by their AI CloudMatrix 384 system that analysts claim can outperform Nvidia on some metrics, despite U.S. export controls.

Analysis

Huawei Technologies' CEO Ren Zhengfei, in a notable interview with the state-run People's Daily, acknowledged that its chip technology currently lags U.S. peers by one generation, yet emphasized the company's strategic efforts to overcome this through methods like cluster computing and a significant annual R&D investment of 180 billion yuan ($25.07 billion), partly focused on compound chips. Ren's comments, the first from Huawei on its advanced chipmaking since U.S. export controls began in 2019, aim to assuage concerns, stating there's "no need to worry about the chip problem," and coincide with U.S.-China trade talks where tech restrictions are a key topic. While Ren downplayed Huawei's overall prowess, suggesting U.S. exaggeration, the company's Ascend AI chips are making inroads; the AI CloudMatrix 384 system, linking 384 Ascend 910C chips, is reportedly capable of outperforming Nvidia's GB200 NVL72 on some metrics according to analysts like SemiAnalysis. This, combined with U.S. restrictions preventing Nvidia (NVDA) from selling its most advanced chips to China, has enabled Huawei to gain market share from Nvidia, reflected in the negative sentiment (-0.5) for the U.S. chip giant despite a "moderately positive" overall sentiment for Huawei's resilient stance. Huawei's strategy includes dedicating a third of its R&D to theoretical research, aiming for breakthroughs beyond Moore's Law to supplement single-chip performance.

Key Decisions for Investors

• Monitor Huawei's progress in AI chip systems like AI CloudMatrix 384, as its ability to offer competitive, domestically-produced alternatives could reshape the AI hardware landscape within China.
• Re-evaluate exposure to Nvidia (NVDA) considering the heightened competition from Huawei in the Chinese market, particularly as U.S. export restrictions limit Nvidia's ability to sell its most advanced products there, a risk highlighted by the negative per-ticker sentiment.
• Factor in the persistent U.S.-China tech tensions and China's drive for semiconductor self-sufficiency when assessing long-term investments in the global technology sector, as this could create both risks for established international players and opportunities for emerging Chinese firms.