Huawei Technologies’ founder and CEO, Ren Zhengfei, has made a rare public disclosure about the company’s chipmaking capabilities, admitting that its semiconductors are one generation behind those of U.S. competitors. Despite this, Ren says Huawei is finding alternative ways to boost performance using innovative techniques like cluster computing and compound chips, even under the constraints of ongoing U.S. sanctions.
Ren’s remarks, published in China’s People’s Daily, represent Huawei’s first public commentary on its progress in chip design since the imposition of sweeping U.S. export controls in 2019. These comments come amid renewed U.S.-China trade talks in London, where tech restrictions are expected to be a central issue.
While Huawei’s chips still lag behind the most advanced designs from U.S. companies like Nvidia, Ren highlighted how the company is compensating for hardware limitations through software innovation and computing architecture strategies.
“Our single chip is still behind the U.S. by a generation,” Ren acknowledged. “We use mathematics to supplement physics, non-Moore’s law to supplement Moore’s law, and cluster computing to supplement single chips.”
These approaches, he says, are producing practical results despite technological constraints. Moore’s Law, which refers to the exponential increase in chip performance over time through miniaturization, has faced physical limitations in recent years. Ren suggests Huawei is innovating beyond this traditional framework.
Cluster Computing: A Key Workaround
Ren placed particular emphasis on cluster computing, a process in which multiple processors are linked together to operate as a single unit. This method is a critical workaround, especially for Huawei, which lacks access to cutting-edge lithography machines and advanced manufacturing equipment that its U.S. and Taiwanese rivals use.
This technique has allowed Huawei to launch powerful systems such as “AI CloudMatrix 384”, which links 384 Ascend 910C chips into a single AI model training platform. The system has reportedly achieved performance benchmarks that surpass Nvidia’s GB200 NVL72 system on specific metrics, according to semiAnalysis founder Dylan Patel.
These developments are not just technological but geopolitical, as they signal China’s growing self-sufficiency in high-performance computing despite export bans from the West.
The Rise of Ascend: A Competitor to Nvidia in China
Huawei’s Ascend series of AI chips is now seen as a leading alternative to Nvidia’s chips within China. Nvidia, the global leader in AI hardware, has been prohibited from selling its most advanced GPUs to Chinese firms under U.S. restrictions. This has created a vacuum in China’s AI chip market one that Huawei is actively filling.
While Nvidia’s chips remain more powerful overall, Huawei’s restricted competitors have lost significant Chinese market share as domestic firms, incentivized by the government, turn to Huawei’s AI hardware.
In April, the launch of AI CloudMatrix 384 was widely covered by tech analysts, who praised its performance and potential to support large-scale AI model training within Chinese firms. According to Patel, Huawei now possesses AI systems that can rival and sometimes beat Nvidia’s offerings, positioning the company as a serious contender in the AI hardware race at least within domestic borders.
U.S. Restrictions a Catalyst, Not a Crippling Blow
Ren downplayed concerns that U.S. sanctions would stall Huawei’s growth, stating there is “no need to worry about the chip problem.” Instead, he painted the export controls as a challenge that forced the company to innovate faster and more creatively.
Huawei invests roughly 180 billion yuan (approx. $25 billion) annually in research and development. About one-third of that investment goes into theoretical research, with the remainder used for product innovation and engineering.
“Without theory, there will be no breakthroughs, and we will not catch up with the United States,” Ren emphasized.
This philosophical focus on theory signals that Huawei is playing the long game, betting that long-term investment in foundational research will yield future breakthroughs that can close the gap with Western rivals.
Ren also indicated Huawei is seeing promise in compound chips semiconductors made from a combination of elements rather than traditional silicon. These materials, such as gallium nitride (GaN) or silicon carbide (SiC), offer better energy efficiency and performance at high frequencies, making them well-suited for AI, 5G, and high-performance computing applications.
If Huawei can make breakthroughs in compound chip technology, it could leapfrog traditional performance constraints and sidestep barriers associated with silicon-based chip production, where it still lags behind industry leaders like TSMC and Intel.
Despite highlighting Huawei’s achievements, Ren struck a modest tone:
“The United States has exaggerated Huawei’s achievements. Huawei is not that great. We have to work hard to reach their evaluation.”
His candid admission underlines the complexity of China’s chip race. Huawei is not yet on par with American giants in raw capability, but its resourceful use of alternative computing models, domestic market dominance, and steadfast R&D investments make it a formidable challenger in a constrained environment.
As U.S.-China tensions continue and export controls tighten, Huawei’s path will serve as a case study in innovation under pressure a company forced to engineer its way out of isolation and find strength through necessity.
