CHINESE tech giant Huawei on Monday formally unveiled the Tau Scaling Law, a new principle that departs from the geometric scaling logic that has long underpinned chip development, marking the first time China has proposed a guiding principle for the global semiconductor industry.
Huawei expects to launch a new Kirin smartphone chip this autumn that fully adopts logic-folding technology, significantly improving related performance, while high-end chips developed under the Tau Scaling Law are expected to reach transistor density equivalent to the 1.4-nanometer process by 2031.
The announcement was made by He Tingbo, a Huawei board member and president of the company’s semiconductor business department, during a keynote speech titled “New Semiconductor Path in Practice” delivered at the 2026 IEEE International Symposium on Circuits and Systems in Shanghai on Monday.
Based on the law, Huawei has successfully designed and mass-produced 381 chips over the past six years.
The Tau Scaling Law proposes replacing “geometric scaling” with “time scaling.” With the systematic reduction of the time constant, or tau, as its objective, the law seeks to continuously compress signal propagation delays through innovative technologies such as logic folding, thereby steadily increasing transistor density and enabling the sustained evolution of semiconductors and electronic systems.
The Tau Scaling Law establishes a multilevel collaborative optimization framework spanning devices, circuits, chips and systems. By 2031, the transistor density of high-end chips developed under the law is expected to reach a level equivalent to the 1.4-nanometer process, according to Huawei.
Reuters reported on Monday that the target is “significant” because 1.4 nanometers is expected to be close to the global frontier for advanced chipmaking around the end of the decade.
Some international media outlets including Reuters, CNBC and Nikkei have framed the development as part of China’s push for technological self-reliance amid United States sanctions that have made it harder for the country to manufacture cutting-edge chips.
Over the past decades, Huawei developed capabilities across system-on-chip design, optoelectronics and advanced packaging. The portfolio eventually spanned smartphones, artificial intelligence (AI), general-purpose processors, telecommunications, networking and consumer electronics, playing a significant part in Huawei’s 2025 revenue of 880.9 billion yuan ($130 billion).
Speaking about the development of the semiconductor industry, He Tingbo said that, “We believe that openness and collaboration are key to driving ongoing progress in the semiconductor industry. No single company can independently find all the answers along the path of semiconductor evolution.”
“With the Tau Scaling Law, we look forward to working closely with scientists, engineers and industry partners around the world to drive the sustainable development of the semiconductor and electronics industries,” He Tingbo said.
Demand for Ascend chips has risen in China this year, as domestic tech firms seek alternatives to US-based company Nvidia, whose most advanced AI processors are restricted from sale to China, Reuters reported. Last week, Nvidia CEO Jensen Huang told CNBC that the US chipmaker had “conceded” the Chinese market to Huawei.
More than 80 large-language models, including DeepSeek and Qwen, have been adapted and developed based on Ascend, while deployments in sectors such as the internet, finance and health care have demonstrated the reliability of domestic computing power in key fields.
Chinese large-language models have also been adapted to deployed on or used for training and inference with other domestic AI chips and computing platforms, including Hygon DCU, Cambricon, Moore Threads and Alibaba’s T-Head.
On May 19, a spokesman of the National Development and Reform Commission said that core technologies and application demand in the AI sector are both growing rapidly, and that China will always guide domestic large-language models to step up adaptation to homegrown computing chips.
