Navigating Restrictions: How China’s Semiconductor Industry and Alibaba Are Innovating Amidst Challenges

Introduction

China’s semiconductor industry is reportedly on the verge of achieving basic self-sufficiency, according to the South China Morning Post. This milestone comes amidst ongoing tensions with the United States, which has imposed restrictions on China’s access to advanced semiconductor technology. These developments have spurred China to accelerate its efforts to innovate and become self-reliant in this critical sector.

Yesterday, there was an interesting article in the SCMP about China’s semiconductor industry that got me thinking about the whole situation.

Gerald Yin Zhiyao, chairman and CEO of Shanghai-listed Advanced Micro-Fabrication Equipment China (AMEC), said during a panel discussion last week that China’s semiconductor supply chain can achieve self-sufficiency despite gaps in “quality” and “reliability”, providing fresh evidence that US restrictions may have accelerated China’s chip industry development.

“I had thought we need at least 10 years to find a solution, but with joint efforts from hundreds of companies over the past two years, we can reach basic self-sufficiency by this summer,” he said.

The Impact of U.S. Restrictions

The U.S. has taken several steps to restrict China’s access to advanced semiconductor technology. Starting in 2019, the Trump administration banned U.S. companies from supplying Huawei without export licenses. The Biden administration expanded these restrictions, targeting advanced chips and equipment to hinder China’s military capabilities.

While these measures disrupted China’s access to critical technology, they also spurred increased domestic investment in chip development. Consequently, U.S. firms are losing market share in China, threatening their ability to invest in future technologies and potentially weakening their competitive edge. Moreover, the restrictions highlight the limitations of export controls amidst rapid technological change and compliance gaps with international partners.

Early Signs of Chinese Innovation

Despite these challenges, Chinese companies like Huawei have made significant strides in semiconductor development. In 2023, Huawei launched the Mate 60 series with a 7nm Kirin 9000s 5G chip, marking a significant breakthrough despite being developed with older technology. Huawei has partnered with SMIC, leveraging its 7nm technology and reportedly working on even more advanced chips. Additionally, Huawei’s Ascend AI chips have gained popularity in China, competing with NVIDIA’s A100 due to their efficiency.

Creativity and Efficiency in China’s Approach

Chinese tech companies, like Tencent and Alibaba consistently emphasize, in conversations and earnings calls, their efforts to maximize the efficiency of the chips they have access to. This trend was recently highlighted when a research team from Stanford University plagiarized a model from Tsinghua University (the Chinese MIT if you want so), due to its efficient use of resources. This approach suggests that China is finding innovative ways to circumvent the limitations posed by the lack of the most powerful chips.

Uncertainty in China’s Semiconductor Future

While China has made progress, significant uncertainty remains about its ability to catch up with global leaders in semiconductor technology. Challenges include reliance on older technologies and difficulties in producing the most advanced chips. However, with continued government support and industry collaboration, China has the potential to make further advancements in this critical sector. How this plays out I don’t know, but it’s interesting to watch it from the sidelines.

What Does This Mean for Investors?

The Chinese internet landscape has navigated restrictions from the outset, blocking foreign companies like Google and OpenAI. This environment has fostered the development of independent domestic services, unlike Europe, which lacks a widely-used alternative to Google Search. In China, companies compete primarily with other domestic players, meaning that even if their technology is slightly weaker than foreign counterparts, it doesn’t significantly impact their domestic market position.

For instance, Baidu Search’s inability to rival Google is more about its poor performance than a lack of potential. Other Chinese tech companies, like Tencent, have successfully captured market share from Baidu because, let’s be frank, Baidu search just sucks. Therefore, the limited availability of high-power chips is unlikely to significantly harm these companies domestically, as they are primarily competing with each other.

Example: Alibaba

I’ve been very critical of Alibaba recently, and you can read about it here and here. However, since Daniel Zhang’s departure, things are slowly improving. In an interesting interview with Joe Tsai, the chairman of Alibaba, and JPMorgan, Tsai outlined some of their strategic ideas.

Two key points stood out to me. First, Joe Tsai discussed (see here) the synergies between cloud and AI, emphasizing that anyone wanting to train models must use Alibaba’s cloud services. In the recent quarter, AI-driven cloud computing revenue grew triple digits. Notably, Alibaba is one of the few companies globally, alongside Google, that integrates both a cloud business and a proprietary AI business under one roof. This is unlike Microsoft, which partners with OpenAI, or Amazon, which focuses mainly on cloud computing without a proprietary AI component.

Secondly, Alibaba is hedging its bets in a smart way (see here). Tsai mentioned that AI is too important to rely solely on their own development. Instead, Alibaba leverages its cloud business because smaller tech companies in China cannot afford the chips required to train their models. Alibaba offers computing time in exchange for equity stakes in these startups, allowing them to hold stakes in numerous promising Chinese startups and leveraging their cloud capabilities.

Although I’ve been critical of Alibaba and acknowledged their past mistakes, this development seems underappreciated when compared to Western counterparts. In addition, Alibaba has learned from Tencent to avoid low-margin, highly individualized cloud projects. It appears that their cloud business is finally catching up after a period of stagnation, with double-digit growth predicted for the second half of 2024.