Samsung’s foundry business is reportedly in discussions with potential clients for its advanced 2-nanometer chip manufacturing technology. While Qualcomm is a possibility due to a potential 2nm project with Samsung, Tesla is considered a more likely customer, according to Ray Wang, research director of semiconductors, supply chain, and emerging technology at The Futurum Group.

Samsung’s foundry service, which manufactures chips based on designs provided by other companies, is the second-largest global provider, trailing Taiwan Semiconductor Manufacturing Company. In April, Samsung announced its ambition for its foundry business to commence mass production of its next-generation 2-nanometer technology and to secure important orders for this advanced product. Smaller nanometer sizes in semiconductor technology indicate more compact transistor designs, leading to enhanced processing power and efficiency.

Samsung is scheduled to release its earnings on Thursday and anticipates a more than 50% drop in its second-quarter profit. An analyst previously attributed this projected decline to weaker orders for its foundry business and challenges in capturing AI demand for its memory products. The company has lagged behind competitors SK Hynix and Micron in the high-bandwidth memory (HBM) chip market, a crucial component for AI chipsets.

SK Hynix, the current leader in HBM, is the primary supplier of these chips to Nvidia. While Samsung has reportedly been working to obtain certification for its latest HBM chips from Nvidia, a local outlet has suggested that these plans have been postponed until at least September.

China’s AI Chip Race: Can Domestic Alternatives Overtake Nvidia Amid US Restrictions?

Table of Contents

China’s AI Chip Race: Can Domestic Alternatives Overtake Nvidia Amid US Restrictions?

As the U.S. government tightens restrictions on China’s access to advanced semiconductors crucial for artificial intelligence (AI) development, Beijing is doubling down on its efforts to foster a self-reliant domestic AI chip industry. This ambition faces meaningful hurdles, including limitations on acquiring essential chip design and manufacturing technologies.Despite investing tens of billions of dollars, China’s AI chip capabilities still lag behind global leaders like Nvidia, but progress is being made, especially by companies like Huawei. According to SemiAnalysis, Huawei’s Ascend 910C GPU is now only about a year behind Nvidia’s export-restricted H20 chip in performance, signaling a narrowing gap.

The Four Pillars of China’s AI Chip Ecosystem

Building a competitive AI chip ecosystem requires strength across four key areas: chip design, fabrication, equipment, and memory. Hear’s a breakdown of China’s current standing in each:

1. AI Chip Design: Closing the Gap

Nvidia, an american company, dominates the AI chip design landscape with its graphics processing units (gpus). While competitors like AMD exist, Nvidia’s GPUs are the industry benchmark. China’s Huawei, through its HiSilicon division, is emerging as a strong contender. Its Ascend 910B GPU is already in mass production,and the next-generation Ascend 910C is anticipated to further reduce the performance gap with Nvidia. semianalysis estimates the 910C is only a year behind Nvidia’s H20.

Did You know? Nvidia doesn’t manufacture its own chips; it outsources production to foundries like TSMC.

2. AI Chip Fabrication: SMIC’s Progress and Challenges

Taiwan Semiconductor Manufacturing Company (TSMC) is the world’s leading chip foundry, producing the majority of advanced semiconductors. However, TSMC adheres to U.S. export controls, preventing it from taking orders from blacklisted companies like Huawei. This forces Chinese designers to rely on domestic foundries, primarily Semiconductor Manufacturing International Corporation (SMIC).SMIC can produce 7-nanometer chips, but lags behind TSMC’s more advanced 3-nanometer technology. Smaller nanometer sizes generally equate to greater processing power and efficiency. Despite this, SMIC is suspected of producing a 5-nanometer chip for Huawei’s Mate 60 Pro in 2023, demonstrating progress. TechInsights provides in-depth analysis of SMIC’s technological advancements.

3. Advanced Chip Equipment: the Lithography Bottleneck

SMIC’s ability to meet Huawei’s GPU demands is constrained by export controls, particularly from the Netherlands.ASML, a Dutch company, is the world’s leading supplier of advanced chipmaking equipment, specifically lithography machines. These machines are essential for transferring intricate patterns onto silicon wafers. The Netherlands, complying with U.S. regulations, blocks the sale of ASML’s most advanced extreme ultraviolet (EUV) lithography machines to China. Jeff Koch, an analyst at SemiAnalysis, notes that EUV lithography is the most significant barrier to China’s advanced chip production. ASML’s website offers details on their lithography technology.

4. AI Memory Components: Catching Up in HBM

high Bandwidth Memory (HBM) is a critical component for AI applications, working alongside GPUs to store data. South Korea’s SK Hynix, Samsung, and U.S.-based Micron are the leading HBM manufacturers.South Korea also adheres to U.S. chip restrictions,limiting HBM sales to China. In response, Chinese memory chip maker ChangXin Memory Technologies (CXMT) is developing HBM in partnership with Tongfu Microelectronics. However, analysts estimate that CXMT is still several years behind the global leaders in HBM technology. SK Hynix is a major player in the HBM market.

china’s AI Chip ambitions: A Summary

Area	Global Leader	China’s Status	Key Players
Chip Design	Nvidia (US)	Closing the gap	Huawei (HiSilicon), Enflame Technology, Biren Technology
Chip Fabrication	TSMC (Taiwan)	Lagging behind	SMIC
Chip Equipment	ASML (Netherlands)	Major bottleneck	SiCarrier Technologies
Memory (HBM)	SK Hynix (South Korea)	Early stages of development	CXMT, Wuhan Xinxin Semiconductor Manufacturing

Pro Tip: Monitor industry publications and analyst reports for the latest developments in China’s AI chip industry.

Evergreen Insights: The Broader Context

China’s push for domestic AI chip production is part of a broader strategy to achieve technological self-sufficiency and reduce reliance on foreign suppliers. This strategy is driven by concerns over national security and the potential for disruptions to supply chains. The U.S. export controls are intended to slow China’s technological advancement,but they also incentivize domestic innovation. The long-term impact of these policies on the global semiconductor industry remains to be seen.

Frequently Asked Questions About China’s AI Chip Industry

Why is China investing heavily in domestic AI chip development?: China aims to achieve technological self-sufficiency and reduce reliance on foreign suppliers due to national security concerns and potential supply chain disruptions.
How do U.S. export restrictions affect China’s AI chip industry?: U.S. export restrictions limit China’s access to advanced chip design and manufacturing technologies, hindering its progress but also incentivizing domestic innovation.
What are the main challenges facing China’s AI chip manufacturers?: The main challenges include lagging behind in chip fabrication technology, limited access to advanced chipmaking equipment, and catching up in the development of high-bandwidth memory (HBM).
Which Chinese companies are leading the charge in AI chip development?: Huawei (HiSilicon) is a leading chip designer, SMIC is the primary chip foundry, and CXMT is developing HBM memory chips.
How far behind is China compared to global leaders in AI chip technology?: while progress is being made,China still lags behind global leaders like Nvidia and TSMC in several key areas,particularly in chip fabrication and HBM memory.
What role does lithography play in AI chip manufacturing, and why is it a bottleneck for China?: Lithography is essential for transferring intricate patterns onto silicon wafers, and China’s access to advanced EUV lithography machines is restricted, limiting its ability to produce cutting-edge chips.

What are your thoughts on China’s ability to overcome these challenges? How will this impact the global tech landscape?

Disclaimer: This article provides general information and should not be considered as professional financial or investment advice.

Share your insights and join the conversation! Subscribe to our newsletter for the latest updates on global technology trends.

SK Hynix Inc

Samsung Secures $16.5B Chip Supply Deal with Potential Qualcomm, Tesla Orders

China Chips: Supply Chain Race vs US Curbs