China has long been preparing for an eventual decoupling with the US and the west. Tech self-sufficiency is the primary motivation behind the Made in China 2025 plan.

The chip war has simply accelerated Beijing’s push in chip independence –

1. semiconductor investment funds by both central and local governments to incentivize domestic innovation and substitution of western technologies

2. mobilize big tech firms to build full chip and AI stacks to ensure their compute demands can be met without US/western tech. The big tech firms making a full throttled push include Huawei, SMIC, Alibaba, Xiaomi, ByteDance, and Baidu

3. break chip embargo through smart engineering such as “stacking and clustering” to improve domestic data center performance without the most advanced semiconductors, for example Huawei’s CloudMatrix 384

4. joint collaboration among academia, research institutes, big business, and the state in leapfrog technologies for next-gen chip development such as third-generation semiconductors (I’ll discuss in detail later), photonic chips, open-sourced RISC-V architecture, etc.

5. build ever greater human capital for chip and AI technologies, including setting up related STEM majors in more universities and funding more PhD graduates

As I wrote before, China represents the largest chip demand globally. It spent more on semiconductor imports ($400 billion plus in 2023) than oil when the country is already history’s largest oil importer. The economic rationale for chip self-sufficiency is evident.

While the US solution to its rare earth bottleneck is a typical financialized response, Beijing’s fix for the chip bottleneck represents an engineering and industrial response on top of increased funding.

When Ren Zhengfei, the founder of Huawei Technologies, was asked about the company’s chip strategy in a recent interview, he admitted Huawei’s Ascend chip still lagged behind the best Nvidia chips “by a generation”.

However, Ren pointed out Huawei was achieving state-of-the-art data center performance by using methods like “stacking and clustering”. Huawei has patented techniques to package chiplets on top of each other to make processors smaller.

Through smart engineering and optimized algorithms, Huawei shows it can reach performance parity for data centers with inferior chips.

Ren emphasized China’s many advantages in developing AI, including “millions of young people studying engineering” and “sufficient electricity generation and transmission grid, and the most developed communications networks in the world”.

I wrote about Huawei’s full tech stack in an earlier article (https://huabinoliver.substack.com/p/huawei-is-transforming-and-building).

In addition to smart engineering to get around chip embargo, China is also focused on a two-pronged strategy to improve its position in the global chip supply chain –

– capacity building across the entire semiconductor value chain with particular emphasis on its downstream leverage as the biggest chip customer and its strength in testing, packaging and mature nodes

– invest in leapfrog technologies, such as the third-generation semiconductors, also known as wide bandgap semiconductors

First, around existing chip technologies, China has built a dominant position in raw materials, assembly and test market, and select legacy logic chip fabrication nodes. Beijing has been rapidly expanding its position in memory chip fabrication.

China is the leading international producer and processor of a wide range of semiconductor-relevant raw materials, including gallium, germanium, magnesium, natural graphite, scandium, tungsten, and the entire range of rare earth elements.

In the mature process market (>22 nm), China is nearing parity with Taiwan with upward of 30% of global market share. China’s share is projected to reach 40% and overtake Taiwan by 2030.

In packaging and testing, China is expected to account for 25% of the global market by 2027. In packaging alone, China leads the world with a 38% market share.

In the memory chip segment, according to South Korean analysis, China’s capacity has already surpassed Samsung and SK Hynix, the global memory chip market leaders. Yangtze Memory Technologies (YMTC) and Changxin Memory Technologies (CXMT) are now the leading players in NAND flash memory and DRAM respectively.

China’s true strength is even more evident downstream. China dominates global production of electronics such as mobile phones and home appliances, and the subsystems and commercial products built on top of them.

As a result, China is one of the largest customers for many top global chip makers such as Intel, Qualcomm, and Nvidia. In 2023, China accounted for 27% of Intel’s revenue, compared to 26% for the US. China represented 46% of Qualcomm’s global revenue while the US accounted for less than 5%.

Beijing is also adept at identifying in-demand growth areas like data centers and artificial intelligence, toward which it can direct its mature node technologies, and do so on a relatively blank competitive canvas, and then scaling to that end.

Beyond improving its competitive position in the existing chip technology arena, China has identified the third-generation semiconductor technology as a potential leapfrog opportunity.

Third-generation semiconductors, also known as wide bandgap semiconductors, refer to materials and integrated circuits made with them such as Silicon Carbide (SiC), Gallium Nitride (GaN), and Indium Phosphide, that have wide energy bandgaps.

Such materials offer superior properties compared to traditional silicon (first-generation) and gallium arsenide (second-generation) and able to handle higher power levels, temperatures, and voltages than silicon semiconductors.

These materials are characterized by high breakdown voltage, high thermal conductivity, high electron saturation velocity, and high radiation resistance, making them ideal for high-power, high-frequency, and high-temperature applications such as EVs, data centers, and clean energy production.

As a result, third-generation semiconductors present multi-billion-dollar addressable markets in the immediate term and are projected to see considerable growth rates in the years ahead.

While third-generation semiconductors constitute a relatively new field and one with applications to new industries, their properties are not necessarily at the bleeding edge of design and fabrication, affording China a market niche that doesn’t require the cutting technologies it has been denied.

China faces a largely open space, unprotected by a defensive moat of western patents and IPs, to develop its own suite of intellectual properties and proprietary technologies.

And Beijing has prioritized the field. In a May 2023 speech, Xiang Libin, vice minister of China’s Ministry of Science and Technology (MOST), highlighted China’s emphasis on and support for third generation semiconductors: “Third-generation semiconductors represented by silicon carbide and gallium nitride have excellent performance and have huge potential in new energy vehicles, information communication, smart grids, and other fields. The Ministry of Science and Technology has attached great importance to the technological innovation and industrial development of third-generation semiconductors and has given the field long-term continuous support.”

This emphasis and support are reflected in China’s 14th Five-Year Plan period (2021 to 2025), which explicitly elevated wide bandgap semiconductor materials to the level of national strategy, calling to “develop silicon carbide, gallium nitride, and other wide bandgap semiconductors.”

Prioritization in government strategy has translated into competitively oriented actions. The elevation of wide bandgap semiconductors in China’s highest-level blueprint for strategic development kicked off a wave of supporting, operationalizing policies and plans at both central and local government levels.

These plans have outlined financial support measures for companies, market share and technological targets, and industrial initiatives. They have prioritized the entire wide bandgap semiconductor value chain, including applications.

For example, Shanghai’s 2022 “Action Plan to Build a Future Industrial Innovation Highland to Develop and Expand Future Industrial Clusters” described an end-to-end focus from upstream to downstream: “Promote the development of silicon carbide, gallium nitride, and other wide bandgap semiconductor compounds; improve the energy level and mass production scale of crystal preparation technology of wide bandgap semiconductor compounds; actively lay out the wide bandgap semiconductor wafer manufacturing technology; enhance the product design ability of wide bandgap semiconductor chips; and expand the application fields.”

Sorry about the boring language. This is why most people don’t read Chinese government plans, which can be really technical and dense, unlikely the entertainment provided by clown Trump.

Another example is the Shenzhen Action Plan to Cultivate and Develop Semiconductor and Integrated Circuit Industry Clusters (2022-2025) describes a gallium nitride and silicon carbide project intended to “seize the commanding heights of the industry and enhance product market dominance and voice.”

Beijing’s Zhongguancun Science Park has declared its intention to accelerate the construction of Zhongguancun Shunyi Park as a third-generation semiconductor industry cluster with global influence, “with an emphasis on silicon carbide, gallium nitride, gallium oxide, and diamond.”

China’s industry discourse parallels national and local policy.

Zhang Rujing, the founder of state-owned chip foundry Semiconductor Manufacturing International Corp (SMIC)., has called third-generation semiconductors an area in which China can “overtake the west on the straight.”

And Yu Chengdong, CEO of Huawei’s consumer business, said that China hopes to “achieve leadership in a new era” of third-generation semiconductors. He noted, “the gap between third-generation semiconductors at home and abroad is not as obvious as that of first- and second-generation semiconductors. Domestic manufacturers can catch up with foreign manufacturers and complete domestic substitution.”

China’s emphasis on third-generation semiconductors has already yielded domestic champions in the field. Take, for example, the photonic integrated circuit module company Zhongji Innolight (Innolight).

Unlike companies like SMIC and Huawei, Zhongji is far from a well known name in the West. But, building on third generation semiconductor technology, and indium phosphide in particular, it is the world’s leading provider of optical module solutions, small hardware that helps to network data centers and transmit high-throughput data flows that propel cutting-edge artificial intelligence applications.

Zhongji Innolight is also the only manufacturer in China that mass-produces and supplies 100-gigabit data center optical modules. It has become a key provider for global big tech companies developing data centers, including AI hyper-scalers such as AWS, Oracle, Alibaba, and Tencent.

The third-generation semiconductor ecosystem also demonstrates how Beijing can leverage its underappreciated dominance within the semiconductor supply chain.

Gallium nitride is a key material in third-generation semiconductors. And China is the dominant global gallium producer, accounting for approximately 98% of the world’s supply. China enjoys similar monopoly in many other critical minerals as discussed earlier.

In short, third-generation chip technology does not rely on the most cutting-edge of ever-shrinking transistor size. The utility of third-generation semiconductors for processing aligns neatly with the demands of contemporary critical, and growing, applications, like data centers and electric vehicles.

And China has been quietly developing a decisive upper hand in the sector for years, leveraging its materials advantages, its manufacturing capacity, and an emphasis on cultivating research and commercial champions.

Across the board, the third-generation semiconductor case underscores the futility of a US semiconductor industrial policy that does not account for China’s resilience and innovation.

In summary, as China and US continue to intensify their geoeconomic and geopolitical competition across multiple technological domains, each has to adapt and counteract the chokehold moves by the other.

China’s strategy is to focus on market economics, engineering innovations, industrial scale, and investment in human capital to overcome US challenges. The end goal is to use the tech embargos its adversaries have imposed as a catalyst to achieve tech self-reliance and sovereignty.

Vir: Hua Bin