Huawei Technologies says it has found a new pathway to narrow the roughly five-year gap with industry leader TSMC, potentially advancing its chipmaking without the cutting-edge lithography equipment China cannot access.
The LogicFolding bet
Speaking at a chip conference, Huawei semiconductor chief He Tingbo said the company will begin making 1.4-nanometer chips by 2031 using its own "LogicFolding" technology, a process that boosts chip performance by increasing transistor count and optimizing data-transmission speed. By comparison, TSMC has said it will begin mass production of the same class of chip in 2028. He added that Kirin mobile chips launching this fall will be the first to adopt the LogicFolding architecture. "This year we have prepared a surprise for the whole industry," she told the audience. "Not saturation, not continuation, but a big leap ahead."
Defying the EUV consensus
The crux of the claim is that Huawei believes it can advance significantly without ASML's extreme ultraviolet (EUV) lithography machines, widely seen as essential for mass-producing chips at 5nm and below. Huawei has previously filed patents pointing to self-aligned quadruple patterning, a technique that etches wafers multiple times to raise transistor density. If it can manufacture 1.4nm chips in volume, it would challenge the prevailing view that EUV is indispensable for the most advanced semiconductors that power frontier AI.
A new scaling theory
He said the LogicFolding architecture rests on Huawei's own "Tau Scaling Law," which the company is positioning as an alternative to Moore's Law and has dubbed "Her's Law." Huawei said it has designed and made 381 chips over six years based on the principle. Markets reacted quickly: Shanghai's Star 50 Index hit a record after the announcement, while foundry partner SMIC rose more than 18 percent and Hua Hong Semiconductor surged by its daily limit.
Context and caveats
The disclosure extends China's semiconductor self-sufficiency drive following years of U.S.-led export controls, and follows a Huawei roadmap to fill the gap left by restricted Nvidia AI accelerators. Still, it remains unclear whether Huawei can reach the cutting edge by experimenting with a non-mainstream path. The push parallels other regional efforts to build domestic AI silicon, including Alibaba's in-house AI chip and consolidation such as recent semiconductor acquisitions.
Reporting based on coverage from Bloomberg via Fortune and the South China Morning Post.