Recently, Chinese tech giant Huawei made a high-profile announcement of the “Tao Law” in Shanghai, claiming to bypass the limitations of lithography machines and achieve an equivalent 1.4-nanometer process level by 2031. However, experts believe that this is just an alternative route under technological blockade, not a real breakthrough, reflecting Huawei’s carefully designed strategic narrative.
The news from Huawei has also been met with caution in the capital market. According to Southern Finance, driven by the announcement of the “Tao Law” by Huawei, the A-share Sci-Tech Innovation 50 Index surged by 5.88% on May 25th; over 300 stocks in the semiconductor sector saw increases, with Huawei Technologies Co., Ltd. rising by 10.45% and SMIC by over 4%.
However, on May 26th, the A-share market saw a sudden turn in direction, with 4,077 stocks across the market falling. In the market, the semiconductor and computational hardware industries all declined, with lithography machines, CPUs, and memory storage leading the drop, while AI applications showed significant decreases.
According to official media reports like China News, on May 25th, Huawei’s Board Director and President of the Semiconductor Business Department, He Tingbo, officially announced the “Tao Law” at the ISCAS 2026 seminar in Shanghai. He claimed that this was China’s first new guiding principle for semiconductor industry development at a global level, constituting a complete new theoretical framework on how to continuously improve chip performance.
He Tingbo mentioned that the upcoming “Kirin 2026” mobile chip to be released in autumn this year will use logic folding technology for the first time, increasing transistor density by 53.5%.
This Huawei executive also projected that by 2031, high-end chips based on the “Tao Law” would reach a transistor density equivalent to a 1.4-nanometer process. This implies that through system-level time optimization, Huawei will achieve integrated density and computational capabilities equivalent to the 1.4nm process level.
Chinese state media extensively publicized the above news, stating that it was China’s first time proposing new guiding principles for industry development in the global semiconductor field, finding a new path for humanity amidst various technological blockades.
In an interview with Epoch Times, Shen Mingshi, a researcher at the Taiwan Institute for National Defense and Security Studies, stated that Huawei’s emphasis on “transistor density” and “system folding technology” leans more towards design optimization rather than a true breakthrough in underlying process capabilities.
“The concepts of ‘design breakthrough’ and ‘process breakthrough’ are entirely different,” he said. “A genuine process breakthrough is built upon mastering basic technological capabilities and advancing step by step; however, if one merely adopts someone else’s design approach or bypasses certain process restrictions, there will still be a fundamental difference from truly advanced processes.”
Xie Peixue, Deputy Researcher at the Cyber Security and Decision Simulation Research Department of the Taiwan Institute for National Defense and Security Studies, told Epoch Times that Huawei’s recent concepts such as “logic folding” and “Tao Law” do not represent genuine process breakthroughs but are alternative paths forced to develop under U.S. technological blockades.
Regarding Huawei’s claim of “bypassing lithography machine restrictions”, Shen Mingshi believes that this underestimates the importance of lithography machines.
He stated that China currently cannot acquire ASML’s EUV equipment from the Netherlands and has to seek alternative solutions. “However, the core basis of truly achieving a 1.4-nanometer manufacturing process still relies on advanced lithography technology.”
Xie Peixue pointed out that the global semiconductor industry has long been aware that the traditional Moore’s Law is approaching its limit. Therefore, international giants like TSMC’s SOIC and AMD’s Chiplet are also developing 3D packaging and heterogeneous integration technologies.
“However, the difference lies in the fact that other international giants consider packaging technology as a supplement to advanced processes, while Huawei, due to the lack of EUV equipment, is forced to focus on packaging and architectural design.”
He said that TSMC’s true 2-nanometer and future 1.4-nanometer technologies are fundamentally based on EUV (extreme ultraviolet) technology, allowing for more transistors to be packed into a two-dimensional plane, improving performance and reducing power consumption simultaneously.
He added that Huawei is not pursuing this path. “Its so-called ‘process breakthrough’ is not a genuine process advancement from a physical perspective.”
“Huawei’s logic folding more resembles converting a one-story house into a high-rise building, changing the original two-dimensional chip design into a three-dimensional stacked structure, then improving computational efficiency by shortening internal wiring and reducing signal transmission delays,” he said.
Both experts pointed out that heat dissipation and yield issues will be the most challenging real obstacles for Huawei on this path.
Xie Peixue stated, “3D packaging and advanced stacking require extremely high alignment accuracy, and if any layer has flaws, the entire chip may be rendered useless.”
In response to Huawei’s recent high-profile announcement, both experts categorized it as a carefully designed discourse power struggle.
Shen Mingshi pointed out that if a company truly possesses advanced process capabilities, there is no need to introduce new terms. He believes that the current approach seems more like shaping a strategic technological narrative by promoting new concepts and terms to instill confidence in the domestic Chinese market, supply chain, and investors.
However, he emphasized that if the relevant technological system ends up being limited to the domestic Chinese market, then China’s semiconductor industry still lacks real competitiveness in the international high-end market.
Analyzing Huawei’s claim of reaching the “equivalent 1.4-nanometer process level by 2031,” Xie Peixue identified the true strategic intent behind it. He stated that Huawei’s high-profile promotion is a carefully calculated discourse power struggle.
“Since TSMC plans to mass-produce 1.4-nanometer chips by 2028 and Huawei aims for 2031, essentially, they are signaling to the Chinese market and supply chain that even under U.S. sanctions, the gap between us and the cutting-edge technology of the world is only about three years, and we have our own technological roadmap.”
He believes this reflects a reality – China’s semiconductor industry has gradually realized that, without access to ASML’s advanced EUV equipment, the traditional geometric shrinking route has almost become a dead end.
Amid the heated discussion around the “Tao Law”, attention has once again been drawn to Huawei’s technical realities in recent years.
According to Bloomberg, technology consulting company TechInsights discovered that after dismantling Huawei’s Mate 70 Pro Plus in 2024, its built-in Kirin 9020 chip was not thoroughly redesigned but rather incrementally improved on the basis of its predecessor, the Kirin 9010. The Kirin 9020 still uses the same 7-nanometer technology as last year, produced by SMIC, but with modifications to the circuit layout.
Bloomberg quoted analysts saying that Huawei’s current chip technology still lags behind what TSMC introduced five years ago. Compared to TSMC’s 7-nanometer chip in 2019, the Kirin 9020 is slower, consumes more power, and has lower yields.
In the field of AI large models, Huawei’s technology has also sparked controversy. After the release of Huawei’s AI large model “Pangu,” the research team HonestAGI published a technical analysis report on the software code hosting platform GitHub. It accused Huawei’s “Pangu” large model of having a structure highly similar to Alibaba’s Tongyi Qwen-2.5 14B model, with a correlation coefficient of 0.927, far exceeding the industry standard of 0.7.
This study revealed that this similarity indicates that Huawei’s model may have been “upcycled,” rather than trained from scratch.
Meanwhile, an employee claiming to be from Huawei’s Noah’s Ark Laboratory anonymously posted on GitHub, alleging that Huawei’s Pangu large model had been layered multiple times before being based on Alibaba’s Qianwen and DeepSeek, among other competing AI products.