Chinese tech company Huawei announced that it has developed a new principle called the Tau Scaling Law, which it claims could help close the gap between itself and global semiconductor leaders such as TSMC and Intel. The move comes amid a highly competitive global race to develop advanced chips for artificial intelligence, as countries and companies work to build more powerful and efficient computing systems. At the same time, US export restrictions on advanced semiconductors and manufacturing tools have pushed Chinese firms to speed up domestic innovation and reduce dependence on foreign technology.

Huawei’s semiconductor chief, He Tingbo, said the new approach is designed to overcome the limitations of Moore’s Law and advance chipmaking capabilities further. He also referred to it as “He’s Law” in recognition of his contributions. At a recent IEEE conference in Shanghai, he stated that Huawei’s upcoming Kirin chips will be the first to use this technology and are expected to be released later this year.

Huawei describes the Tau Scaling Law as a framework based on the joint evolution of semiconductor technology and electronic systems. Instead of focusing only on shrinking transistors, it aims to reduce signal delays and improve transistor density through better chip design, architecture, and system integration, thereby enhancing performance and efficiency. The company claims this approach could enable performance levels comparable to 1.4-nanometer-class chips by 2031. The Kirin chip series will be the first to adopt this principle, with applications also planned for Huawei’s AI computing chips.

Moore’s Law, long used as the foundation of semiconductor progress, predicts that transistor counts double roughly every two years. However, as physical limits of miniaturization become more difficult to overcome, the industry is increasingly turning to alternative methods such as advanced packaging, chip stacking, and system-level optimization.

Huawei said the new approach could help it continue improving chip performance despite these limitations and reduce reliance on foreign semiconductor technology under ongoing US restrictions. It also positions the company as a potential domestic alternative to firms like Nvidia in AI computing. The company added that elements of this design philosophy have already been applied across hundreds of chip types used in consumer electronics, networking, and computing over the past several years.

While Huawei projects reaching 1.4-nanometer-class performance by 2031, competitors like TSMC and Intel are targeting similar manufacturing milestones around 2029, suggesting they may remain ahead in timing. The development reflects a broader shift in the semiconductor industry away from purely shrinking transistors toward more advanced design and system-level improvements, an approach already being explored by major AI chipmakers such as Nvidia, Google, and Amazon.

Disclaimer: This image is taken from Shutterstock.