Huawei Targets Advanced Chip Design Amid Sanctions

Huawei Technologies aims to produce industry-leading semiconductors using new technology within five years. This initiative underscores efforts by Beijing to counter United States sanctions, which have hindered China's ability to build cutting-edge chips.

Credit: USNPLASH

The Chinese technology company announced its high-end chips will feature transistor density equivalent to 1.4-nanometre processes by 2031. This target was revealed during a semiconductor symposium in Shanghai.

China's most advanced proven chipmaking capability is generally seen at about 7 nanometres. A 1.4-nanometre process is projected to be near the global forefront for advanced chipmaking toward the end of the decade.

Achieving this level through conventional manufacturing alone is viewed as unlikely for China. Washington has restricted the country's access to advanced lithography tools and other key semiconductor technologies.

Taiwan Semiconductor Manufacturing Company (TSMC), the world's largest producer of the most advanced chips, currently uses 2-nanometre manufacturing technology. TSMC plans to introduce a 1.4-nanometre process for mass production in 2028.

Huawei unveiled a new principle for improving chips, called the Tau Scaling Law. The company noted the industry can no longer rely solely on shrinking transistors for computing breakthroughs, a pattern known as Moore's Law.

Transistors have become so small that their dimensions are measured in only a few atoms. The Tau Scaling Law instead focuses on reducing the time signals and data take to move through chips and computing systems.

The global chip industry is increasingly investing in post-Moore's Law solutions, including advanced packaging and chiplets. This search has become particularly urgent for China.

US export controls have restricted Chinese companies' access to the most advanced chipmaking tools. This particularly affects equipment needed to make chips at leading-edge process nodes.

Alternative routes to higher performance are central to Beijing's goal of building a world-leading, self-sufficient semiconductor industry.

He Hui, director of semiconductor research at Omdia, said Huawei is proposing a shift from traditional node-driven scaling to system-level efficiency scaling.

Instead of depending solely on smaller transistors, the company is focusing on shortening interconnect, lowering latency, and improving data movement inside the chip. He described this as a credible way to extract more performance when leading-edge lithography is constrained.

The stakes for Huawei's chip breakthroughs are doubly high. Frontier technologies have become an increasingly important pillar of future economic development and geopolitical leverage for China.

Huawei's Ascend chip series is central to powering Chinese artificial intelligence (AI) models. These include DeepSeek's latest flagship model V4, released last month.

Kirin smartphone chips from Huawei, scheduled for launch this year, will be the first to use a Tau Scaling architecture called LogicFolding. The company said LogicFolding would shorten wiring inside chips and considerably improve performance.

LogicFolding will also be applied to Ascend chips by 2030. It will further be used in large AI clusters, which consist of hundreds or thousands of chips that power data centres.

Huawei's chip division has designed and mass-produced 381 chips over the past six years. These chips, based on the Tau Scaling Law, are used in industries including smartphones and AI computing.

Huawei was placed on a US trade blacklist in 2019. This action cut the company off from many US-origin technologies, including chips and software. It also restricted Huawei's ability to rely on global contract chipmakers.

After the restrictions were imposed, Huawei entered what it described as an "extreme survival mode." A secret backup chip project, led by He Tingbo, president of Huawei's semiconductor business and director of its Scientist Committee, became central to its survival strategy.

The company mounted a surprise comeback in 2023 with the launch of its 5G-capable Mate 60 series smartphones. These devices were powered by a system-on-chip produced by China's biggest contract chipmaker, Semiconductor Manufacturing International Corporation (SMIC), using 7-nanometre technology.

SMIC shares rose 7.6% on Monday after Huawei's announcement of its LogicFolding architecture. SMIC has also recently invested in post-Moore's Law pathways, establishing an advanced packaging research institute in Shanghai in January.

Demand for Ascend chips has risen in China this year. Domestic technology firms seek alternatives to US company Nvidia, whose most advanced AI processors are restricted from sale to China.

Nvidia CEO Jensen Huang said earlier this month that the company had "largely conceded" China's AI chip market to Huawei.

Analysts acknowledge progress but state China remains behind global leaders in the most advanced process technology.

Brady Wang, associate director at Counterpoint Research, noted that cost, power, heat, and system integration remain major challenges, especially for Cloud AI servers.

Wang added that China may narrow the gap with global leaders in the short term, but a technology gap with the most advanced nodes will still remain.

Huawei's chip head, He, acknowledged that its latest approach still faces major hurdles. These include the need for new chip-design tools suited to Tau Scaling and the challenge of preventing overheating, from mobile chips to large AI data centres.

He stated that, given all the various constraints, good solutions have been found. He believes Huawei's solutions for mobile computing and AI computing will be competitive in the coming 10 years.

---

• Huawei aims for 1.4-nanometre equivalent chip density by 2031, a significant advance amid US sanctions.

• The company introduced the Tau Scaling Law, a new principle focusing on cutting data movement time within chips, moving beyond traditional transistor shrinking.

• Huawei's LogicFolding architecture, based on Tau Scaling, will feature in upcoming Kirin smartphone chips and be applied to Ascend AI chips by 2030.

Source: REUTERS