A domestic Nvidia rival is what China desires, but it's proving to be a difficult task.

A domestic Nvidia rival is what China desires, but it's proving to be a difficult task.
A domestic Nvidia rival is what China desires, but it's proving to be a difficult task.
  • Beijing is intensifying its efforts to develop a competitive alternative to Nvidia's AI chips as part of its ongoing efforts to reduce dependence on American technology.
  • A number of obstacles hinder China's progress from U.S. export limitations that impede domestic semiconductor production to a shortage of technological expertise.
  • Analysts identified companies including Huawei as the key rivals to Nvidia in China.

Beijing is intensifying its efforts to develop a competitive alternative to Nvidia's AI chips as part of its ongoing efforts to reduce dependence on American technology.

Beijing's efforts have been hindered by the challenges posed by U.S. sanctions and China's dominance in space, at least in the short term, according to analysts.

Nvidia's growth has been fueled by cloud computing giants purchasing its server products, which incorporate its graphics processing units (GPUs). These chips allow companies, such as OpenAI, to train their large AI models on vast amounts of data.

Applications like chatbots and other emerging AI applications rely on these AI models.

Since 2022, the U.S. government has imposed restrictions on the export of Nvidia's advanced chips to China, with the restrictions becoming stricter last year.

Semiconductors are crucial to China's goal of becoming a top AI power.

Some of China's top competitors challenging Nvidia were identified by CNBC, including technology giants Huawei and and startups such as Biren Technology and Enflame.

The overarching view is that they are lagging behind Nvidia at this point.

According to Wei Sun, a senior analyst at Counterpoint Research, these companies have made significant advancements in creating AI chips designed for specific purposes (ASICs).

"Nevertheless, facing competition from Nvidia remains significant obstacles in terms of technological disparities, particularly in the area of general-purpose GPUs. It is unlikely that we will be able to catch up with Nvidia in the short term."

China's key challenges

One of the challenges faced by Chinese firms, as highlighted by Sun, is a "shortage of technological expertise".

The biggest obstacles to China's aspirations are the U.S. sanctions and their consequences.

American technology restrictions have been imposed on some of China's leading Nvidia competitors, who have been placed on the U.S. Entity List. At the same time, key AI-related semiconductors and machinery have been restricted from being exported to China due to U.S. curbs.

Many Chinese GPU players design chips but rely on a manufacturing company to produce them. For a while, this would have been TSMC. However, U.S. restrictions mean that many of these firms cannot access the chips made by TSMC.

SMIC, China's largest chipmaker, is falling behind TSMC due to Washington's restrictions on its access to a crucial piece of machinery from Dutch firm ASML, which is necessary for manufacturing advanced chips.

SMIC's capacity is being utilized by Huawei's development of advanced chips for smartphones and AI, as stated by Paul Triolo, a partner at Albright Stonebridge.

The main challenge will be SMIC, which must allocate limited resources for its advanced node production between Huawei, which currently dominates, GPU startups, and many Chinese design firms that may be cut off from TSMC, as Triolo explained to CNBC.

Nvidia is more than just GPUs

Nvidia's success can be attributed to both its advanced semiconductors and its CUDA software platform, which enables developers to create applications for its hardware. This has resulted in the creation of an ecosystem around Nvidia's products that others may struggle to replicate.

Triolo stated that the key is not only about the hardware but also encompasses the entire ecosystem, developer tools, and the ability to adapt and advance the ecosystem as technology progresses.

Huawei leading the pack

Huawei was recognized as a top leader in China by Triolo due to its Ascend series of data center processors.

The Ascend 910C, the next generation of chip from the firm, is expected to be as powerful as Nvidia's H100 product, according to a report in August.

Nvidia named Huawei as a competitor in various fields such as chips, AI software, and networking products in their annual report this year.

U.S. export controls on Chinese firms could 'get even worse' if Trump is re-elected: Analyst

Huawei has several advantages in the software and developer community, according to Triolo. However, like other companies in the industry, it faces competition from Nvidia.

Triolo stated that the GPU software support ecosystem is heavily centered on Nvidia and to a lesser extent AMD, and Huawei faces significant challenges in both manufacturing enough advanced GPUs, such as the Ascend 910C, and improving the hardware's performance, due to U.S. export controls that limit SMIC's ability to produce advanced semiconductors.

Chip IPOs ahead?

Over the past two years, Nvidia competitors in China have faced significant challenges, with Biren Technology conducting layoffs in 2022 and Moore Threads following suit in 2023, both citing U.S. sanctions as the reason.

Despite the challenges, startups continue to seek funding to achieve their objectives. According to Bloomberg, Enflame and Biren are among those looking to go public to secure financial backing.

Triolo stated that Biren and other GPU startups have experienced industry personnel from Nvidia, AMD, and other leading western semiconductor companies, but they face the additional challenge of not having the same financial depth as Huawei.

"Both Biren and Enflame are looking to issue initial public offerings (IPOs) in Hong Kong in order to obtain funds for further hiring and growth."

by Arjun Kharpal

Technology