Why Huawei is Abandoning 'Moore's Law' to Target 1.4 Nanometers
Huawei is attempting to rewrite the rules of the game in the semiconductor world. If they cannot possess manufacturing equipment, they will simply rewrite the design theory itself—that is the paradigm shift currently taking place.
> A company cornered by sanctions has come to change the very laws of the industry.
Key points of this article
・Redefined performance based on 'signal speed' rather than miniaturization
・Logic exists to target the equivalent of 1.4 nanometers even without TSMC
・Cases exist where containment actually accelerates technological innovation
Huawei announces 'Tau's Law' and declares it will be at the global forefront by 2031
On May 25, 2026, at the international conference 'IEEE ISCAS 2026' held in Shanghai, He Tingbo, Huawei board member and president of the semiconductor business unit, took the stage to announce the 'Tau (τ) Scaling Law.' The goal is to realize a chip with a transistor density equivalent to a circuit line width of
1.4 nanometers
by 2031. 1.4 nanometers is the current world-leading node that TSMC aims to begin mass-producing in 2028, and Huawei has effectively declared that it will stand head-to-head with that level.
Why did Huawei step off the 'make it smaller' race and start the 'make it faster' race?
To understand Huawei's internal decision-making, one must first grasp the 'history of being cornered.' In 2020, access to TSMC was cut off by US export controls. They could no longer outsource high-precision 7nm or 5nm chips, and were forced to abandon the production outsourcing of Kirin chips. The smartphone business was forced to downgrade to 4G, and at one point, it was driven into a state of effective near-death where 'chips could not be obtained.'
What emerged in that situation was a reversal of thinking. If you cannot miniaturize, you should create a theory that improves performance without miniaturization. 'Tau's Law' stops focusing on increasing the number of transistors and sets the shortening of the 'time (τ)' it takes for signals to travel through circuits as the new axis of evolution. If Moore's Law is about 'quantity,' Tau's Law is about 'speed.'
Born from this theory is a technology called 'LogicFolding.' By dividing circuits and stacking them three-dimensionally, the signal transmission path is physically shortened. If the wiring distance is reduced, parasitic capacitance and resistance decrease, and information processing efficiency jumps even with the same manufacturing process. As a result of adopting this technology for the first time in the Kirin chip scheduled for release in the fall of 2026, the transistor density is said to have improved by over 50% from
155 million to 238 million per square millimeter
and energy efficiency improved by 41%, and the maximum clock speed increased by approximately 13%.
On the other hand, it is also considered prudent to be cautious about taking this at face value. Zhang Binlei, an analyst at market research firm Zhenxin Hui, stated, 'While LogicFolding is a groundbreaking advancement, I do not understand it to be simply equivalent to the 3-nanometer class in terms of overall performance and power consumption.' There remain parts where efficiency gains achieved through design cannot completely overcome the limitations caused by the manufacturing process.
Looking at the semiconductor industry as a whole, the limits of Moore's Law have actually been a common concern for the industry. Even TSMC and Apple are trying to avoid this wall with 'chiplet' technology, which connects two chips to operate as a single chip. Huawei's 'Tau's Law' is not a local solution born from the constraints of sanctions, but rather part of a major trend toward 'performance improvement other than miniaturization' that the entire industry is heading toward. The structure is such that the sanctions have, as a result, forced Huawei to land early on the technological direction that the entire industry was supposed to reach in 10 years.
This is how I changed my way of reading the chip war
What I realized keenly during my time as a consultant is that 'resource constraints become the starting point for strategy.' At one manufacturing client, when a raw material procurement route was cut off, the management team initially only thought about 'compensation.' But the team that redesigned it with the constraints as a given ended up fundamentally changing the cost structure. Huawei's movements overlap perfectly with this structure.
What caught my attention in this news was that the announcement was made at an 'international academic conference.' It was presented not as a business announcement, but as a question posed to the engineering community. I interpreted this as a signal that they want it to be 'verified' and that they 'want to refine the theory while receiving criticism.' I believe a company under sanctions uses this approach not because they are confident in their technology, but because they are aware that it won't become the real deal unless it is forged in an open forum.
Huawei's figure of having mass-produced 381 types of chips over the past six years is understated but impossible to overlook. They have a track record of 'mass production,' not just 'theory.' I judge this to be the boundary line between 'bluffing through announcements' and 'genuine development capability.' The moment the Kirin chip is released in the fall of 2026, the true value of this theory will be verified for the first time as a real device online. That point in time will be the first reality check for this story.
A company cornered by sanctions has come to change the 'definition of the game.' That is the essence of what is currently happening in the semiconductor business.
#Huawei #Semiconductor #Business #AI #Online #ChinaTech