Huawei unveils Tau Scaling Law as workaround for US chip sanctions, targets 1.4nm chips by 2031

Reviewed byNidhi Govil

18 Sources

Share

Huawei's chip division president Tingbo He announced a new semiconductor design approach called Tau Scaling Law that focuses on signal speed rather than transistor miniaturization. The company claims this method will deliver chips with 1.4nm-equivalent performance by 2031, potentially closing the gap with Western chipmakers despite US export controls blocking access to advanced lithography equipment.

Huawei Proposes New Path for Chip Development

Huawei has introduced a fundamental shift in semiconductor design that could reshape how China navigates US chip sanctions. At the IEEE International Symposium on Circuits and Systems in Shanghai, Tingbo He, president of Huawei's chip-design subsidiary HiSilicon, unveiled the Tau Scaling Law—a new principle that prioritizes signal speed over transistor miniaturization

1

. Known in China as the "chip queen," He announced that Huawei has spent six years developing this approach and has already designed and mass-produced 381 chips based on the principle

2

.

Source: SiliconANGLE

Source: SiliconANGLE

The announcement represents a direct challenge to decades of industry orthodoxy. Rather than following Moore's Law—which dictates progress through doubling transistors every two years—Huawei's new chip design focuses on semiconductor optimization at the system level

4

. "Six years ago geometric scaling plateaued for us," He explained, referring to lithographic miniaturization. "We soon realized semiconductor evolution is more than geometric scaling"

1

.

LogicFolding Chip Architecture Promises Dramatic Gains

At the core of Huawei's strategy lies LogicFolding, an architecture that physically folds and stacks logic circuits into a dual-layer framework. By drastically shortening internal wiring to eliminate signal delay, the resulting hardware achieves a 55% increase in transistor density and a 41% boost in power efficiency

2

. The company's upcoming Kirin processors for the flagship Huawei Mate 90 series will be the first commercial chips to feature this architecture, launching in autumn 2026

5

.

Source: FoneArena

Source: FoneArena

"For both AI training and inference, the win is not just in shortening compute time. It is in shortening the time that data spends moving, between chips and inside a chip," He stated

1

. This emphasis on system-level performance and signal speed represents a workaround for US chip sanctions that have blocked Huawei from accessing extreme ultraviolet lithography machines from ASML

4

.

Targeting 1.4nm-Class Chips to Close Technology Gap

Huawei projects it can design high-end chips with transistor density equivalent to a 1.4-nanometer process by 2031

5

. This timeline would significantly reduce China's chipmaking lag, as TSMC is expected to introduce 1.4nm-class chips in 2028

1

. The company aims to scale LogicFolding architecture to its Ascend AI processors and high-capacity data center clusters by 2030, providing local alternatives to restricted Nvidia hardware

2

.

Source: Reuters

Source: Reuters

US export controls currently prohibit Huawei from working with Taiwan Semiconductor Manufacturing Company, forcing reliance on China's SMIC, which uses older generation lithography machines

1

. By some estimates, China is more than five years behind the leading edge in conventional chipmaking

1

.

Implications for China's AI Development and Market Competition

The breakthrough carries significant implications for China's AI development. Following the announcement, shares for China's largest contract chipmaker, SMIC, surged by 7.6%

2

. The development represents a major symbolic and practical win for Beijing's push toward technological self-sufficiency amid EUV restrictions and concerns about over-reliance on foreign semiconductor players

2

.

However, Huawei faces growing domestic competition. IDC data shows that while Huawei shipped 812,000 AI accelerators in China during 2025, Alibaba claimed second place with around 265,000 units

3

. Nvidia still holds 55% market share with 2.2 million AI accelerators shipped to China in 2025

3

. ByteDance and Tencent also have resources to mount serious challenges in the Chinese AI chip market

3

.

Skepticism and the Path Forward to Overcome Limitations of Moore's Law

Not everyone is convinced by Huawei's claims. Lennart Heim, an independent semiconductor and AI policy analyst, suggests Huawei's strategy indicates the company is running into limits on performance gains from shrinking and densifying chips alone

1

. When US markets opened following the announcement, Nvidia, AMD, and Intel shares rose, suggesting competitors remain unfazed

3

.

Critical questions remain about whether transistor density translates directly to performance, and how issues like overheating will be addressed

3

. Huawei has not published independent performance data to support its projections

4

. He promised the company would prove viability with a new chip "before winter 2026," calling it "not saturation, not continuation, but a big leap ahead"

1

. The first test arrives in autumn when Kirin processors reach buyers, while the longer validation runs to 2031

4

.

Today's Top Stories

© 2026 TheOutpost.AI All rights reserved