In order to comply with new regulations on the export of chips to the United States, NVIDIA has been consistently releasing AI chips and graphics cards tailored for the Chinese market.

However, according to sources cited by The Wall Street Journal, since November 2023, major cloud service provider (CSP) in China such as Alibaba and Tencent have been testing samples of NVIDIA’s special chips. These Chinese enterprises have conveyed to NVIDIA that the quantity of chips they plan to order in 2024 will be significantly lower than their initial plans.

According to a report from The Wall Street Journal, in October 2023, the United States announced new regulations preventing NVIDIA from selling advanced AI chips to China. However, NVIDIA swiftly developed a “special edition” chip for China, allowing them to continue selling chips in the Chinese market without violating regulations.

Nevertheless, NVIDIA is facing another challenge: major Chinese CSPs are not actively purchasing the “downgraded” performance versions of the chips.

Chinese enterprises have been testing the highest-performance version, H20, of NVIDIA’s “special edition” AI chips. Some testers have mentioned that this chip enables efficient data transfer among multiple processors, making it a better choice than domestic alternatives for building chip clusters required for processing AI computational workloads.

However, testers also indicate that they need more H20 to compensate for the performance gap compared to previous NVIDIA chips, which increases their costs.

The report indicates that in the short term, the performance advantage of NVIDIA’s “downgraded” chips over domestic Chinese products is diminishing, making Chinese-made chips increasingly attractive to buyers.

Informed sources cited from the report suggest that major players like Alibaba and Tencent are redirecting some advanced semiconductor orders to domestic companies and relying more on internally developed chips. This trend is also observed with the other two major chip buyers, Baidu and ByteDance.

Looking ahead in the long term, Chinese customers are uncertain about NVIDIA’s ability to continue supplying them with chips, as U.S. regulatory authorities have committed to regularly reviewing chip export controls, potentially tightening restrictions on chip performance further.

From the perspective of China’s efforts in the independent development of AI chips, TrendForce previously highlighted in its press release that Chinese CSPs like Baidu and Alibaba are actively investing in autonomous AI chip development.

Baidu developed its first self-researched ASIC AI chip, Kunlunxin, in early 2020, with its second generation scheduled for mass production in 2021 and the third expected to launch in 2024. Post-2023, Baidu aimed to use Huawei’s Ascend 910B acceleration chips and expand the use of Kunlunxin chips for its AI infrastructure.

After Alibaba’s acquisition of CPU IP supplier Zhongtian Micro Systems in April 2018 and the establishment of T-Head Semiconductor in September of the same year, the company began developing its own ASIC AI chips, including the Hanguang 800.

TrendForce reports that T-Head’s initial ASIC chips were co-designed with external companies like GUC. However, after 2023, Alibaba is expected to increasingly leverage its internal resources to enhance the independent design capabilities of its next-gen ASIC chips, primarily for Alibaba Cloud’s AI infrastructure.

According to the data from TrendForce, currently, around 80% of the high-end AI chips used by Chinese cloud computing companies are sourced from NVIDIA. However, in the next five years, this proportion may decrease to 50% to 60%.

Read more

• [News] NVIDIA’s China-exclusive Chip H20 to Begin Production in Q2 by Wistron
• China Continues to Enhance AI Chip Self-Sufficiency, but High-End AI Chip Development Remains Constrained, Says TrendForce

(Photo credit: NVIDIA)

Dutch semiconductor equipment leader ASML Holding N.V. has announced that export licenses for certain equipment have been partially revoked by the Dutch government.

In a press release issued on January 1st, 2024, ASML stated, “A license for the shipment of NXT:2050i and NXT:2100i lithography systems in 2023 has recently been partially revoked by the Dutch government, impacting a small number of customers in China. We do not expect the current revocation of our export license or the latest U.S. export control restrictions to have a material impact on our financial outlook for 2023.”

The press release further stated, “In recent discussions with the US government, ASML has obtained further clarification of the scope and impact of the US export control regulations. The latest US export rules (published October 17, 2023) impose restrictions on certain mid critical DUV immersion lithography systems for a limited number of advanced production facilities.”

Bloomberg reported earlier on January 1^st, 2024, citing unnamed sources, that several weeks before the implementation of export controls on advanced semiconductor equipment in the Netherlands, the U.S. government had requested ASML to cancel the export of certain machines destined for China.

Previously, ASML’s CEO, Peter Wennink, stated that these limitations would exclude the vast majority of Chinese customers in response to the U.S. restrictions. This exclusion is due to the fact that these customers are involved in mature nodes, specifically in the production of semiconductors at 28nm and above.

In addition, last week, the South China Morning Post has cited data, indicating that in November 2023, China had imported critical semiconductor manufacturing lithography equipment from the Netherlands, experiencing a significant surge of 1050% in import value.

In an interview with the South China Morning Post, Jan-Peter Kleinhans, Senior Researcher and Head of Technology and Geopolitics Projects at the Berlin-based think tank “Stiftung Neue Verantwortung” (New Responsibility Foundation), mentioned that the impact on sales would not be immediate following the new U.S. restrictions.

Reportedly, this is because ASML has a lead time of approximately 18 months. This implies that the equipment shipped in the fourth quarter of 2023 would have been ordered in the second or third quarter of 2022, and ASML would apply for export licenses at some point thereafter.

Read more

• [News] Surge in Chip Manufacturing Equipment Imports from the Netherlands to China, Soaring Tenfold After U.S. Tightens Restrictions
• [News] EUV as a Strategic Asset in the Most Advanced Processes: Progress in Intel/TSMC/Samsung’s Adoptions

(Photo credit: ASML)

An increasing number of Chinese semiconductor design companies are seeking collaboration with testing and packaging facilities in Malaysia to carry out advanced chip packaging. According to Reuters’ report, this move aims to hedge the risk of potential expanded U.S. restrictions on the Chinese semiconductor industry.

As there is currently only one non-U.S. testing and packaging provider in Malaysia with advanced capabilities, namely ASE Technology Holding Co., a Taiwanese semiconductor packaging and testing firm, industry sources believe that ASE is likely to become the top choice for orders from Chinese enterprises.

Previously, the U.S. has imposed controls on China’s advanced semiconductor manufacturing processes and access to high-performance chips from major companies like NVIDIA. However, advanced packaging has not yet fallen within the restricted scope.

Two anonymous sources reportedly revealed that some of the Chinese businesses are showing interest in advanced chip packaging services. Despite the fact that the chip packaging sector has not yet faced export controls from the U.S., concerns are rising among businesses due to its involvement in sophisticated technology, fearing that it might be targeted for curbs on exports in the future.

Reuters’ report also indicates that due to the relatively affordable investment costs in Malaysia and the availability of experienced workforce and sophisticated equipment, an increasing number of Chinese chip design firms are seeking Malaysian Firms to carry out advanced chip packaging activities, including graphic processing units (GPUs).

Insiders have informed Reuters that the related contracts only involve packaging and do not violate any restrictions imposed by the U.S.. Additionally, they clarified that wafer manufacturing is not included in these contracts.

Two of the sources mentioned that some contracts have already been agreed. However, these insiders prefer not to disclose the names of the involved companies.

Meanwhile, according to a report from Taiwan’s Economic Daily News, when observing the global landscape of advanced packaging, in addition to TSMC, there are integrated device manufacturers (IDMs) like Intel and Samsung, as well as outsourcing semiconductor assembly and test (OSAT) companies like ASE Technology, Amkor, and others that possess advanced packaging capabilities. Among them, only ASE Technology, Amkor, and Intel have production capacity in Malaysia.

Reportedly, industry analysts predict that Chinese companies seeking advanced packaging support in Malaysia, due to geopolitical considerations, are likely to avoid American companies such as Intel and Amkor. Given that ASE is not an American company and can provide high-end packaging services, it is expected to be the preferred choice for Chinese companies.

ASE has previously stated that it will continue to invest in advanced packaging for AI, expecting the performance of advanced packaging to double next year compared to this year.

Read more

• [News] Bypassing U.S. Restrictions? CXMT Showcases GAA Technology for 3nm Chips
• Malaysia: Rising Global Hub for Semiconductor Backend Testing and Packaging in Supply Chain Shift

(Photo credit: ASE Holdings)

The labor dispute sparked by TSMC’s venture into the United States is poised to come to a close. TSMC’s Arizona plant and the local labor union, Arizona Building and Construction Trades Council (AZBTC), announced an agreement yesterday. TSMC will collaborate with the local union to establish a workforce training program and maintain transparency on public safety issues.

However, TSMC also retains some flexibility, aiming to recruit local employees while seeking the option to hire foreign workers with “specialized experience” if deemed necessary.

It is anticipated that this agreement aligns with TSMC’s prior request for the dispatch of 500 professional equipment engineers from Taiwan to the United States. This move, with the successful acquisition of technical construction permits, is expected to facilitate the expedited installation of equipment, ensuring smooth operational processes in the future.

While TSMC’s move to set up a factory in the United States was fueled by favorable policies, it faced challenges as contractors were unfamiliar with U.S. regulations, causing delays in mechanical and electrical integration and cleanroom construction.

To expedite the installation of advanced process equipment for the groundbreaking four-nanometer fabrication process in the U.S., TSMC had to mobilize nearly 500 personnel from Taiwan for intensive installation work before system integration.

This decision sparked a strong backlash from the AZBTC, criticizing TSMC for disrespecting local technicians’ expertise and raising concerns about TSMC’s potential intention to introduce cheaper labor, impacting local job rights.

Although TSMC consistently emphasized maintaining good interactions with various unions in Arizona, some union representatives were displeased with TSMC’s practice of dispatching personnel through suppliers. Reportedly, they called on the state and federal governments to pressure TSMC, leading to unexpected delays in the installation of new equipment.

After months of negotiation between TSMC and AZBTC, an agreement was reached, listing agreed-upon priority areas, including union training, communication channels, and on-site personnel allocation. However, TSMC, considering global talent distribution, retained the flexibility to hire foreign construction personnel with specific expertise in certain circumstances.

According to the Greater Phoenix Economic Council(GPEC), the specific terms of the agreement reached between the two parties are as follows:

1. Enhanced Workforce Training and Development

A highly skilled, diverse, and inclusive construction workforce is necessary to meet the timelines of the two fabs. The AZBTC intends to recruit a sufficient number of skilled workers to fulfill manpower requirements of TSMC Arizona’s contractors for the Project. TSMC Arizona will partner with AZBTC on the development of union workforce training programs and curricula. The goal will be to build a construction workforce that can support TSMC Arizona in the near and long term with employment opportunities.。

2. Shared Commitment to Site Safety

TSMC Arizona is deeply committed to workplace safety in the operation of all its facilities. To enhance the partnership, TSMC Arizona will maintain transparency with AZBTC with regards to safety assessments, audits, incident records and improvement plans.

3. Industry Leading, Global Workforce

TSMC Arizona is focused on hiring workers locally in the United States. The AZBTC workforce is highly skilled in constructing microchip manufacturing plants. Circumstances may require TSMC Arizona or its vendors to employ foreign workers with specialized experience.

4. Open and Regular Communication

To ensure TSMC Arizona and AZBTC are fulfilling the spirit of the agreement outlined, and accountable to commitments made, ongoing communication and review via regular forums will be critical. A committee will be formed, consisting of members designated by the affiliated AZBTC unions and members designated by TSMC Arizona and the contractors. These meetings will be held quarterly, one of which will be an annual forecast meeting to project future workforce requirements.

The report notes that TSMC also mentioned in a joint statement that its construction of a fab in Arizona represents the largest single foreign direct investment in the state’s history. This wafer fab is set to be the most advanced semiconductor manufacturing base in the United States, creating thousands of stable and high-paying job opportunities locally.

The ongoing construction of the wafer fab’s two-phase project has already generated thousands of jobs in accordance with the prevailing industry wage standards for members of the AZBTC.

(Photo credit: TSMC)

Read more

• [News] The Alliance among Apple, TSMC, and Amkor May Pose Challenges to Samsung’s Advanced Packaging Strategy
• [News] Apple Expands Amkor Partnership at Arizona Fab with Amkor’s $2B Investment in Advanced Packaging

On October 17, 2023, the U.S. government once again expanded its restrictions on the export of semiconductor devices and products to China. The newly added control conditions now encompass NVIDIA’s L40S, A100, H100H800, as well as general-purpose AI server GPUs tailored for the Chinese market, such as A800 and H800. Additionally, AMD’s MI200 series, MI300 series GPUs, and Intel’s Habana Labs’ Gaudi 2, Gaudi 3 GPUs fall under the regulatory framework.

Recalling the U.S. government’s export restrictions on AI chips issued to IC design firms in September 2022, at that time, only A100, H100, and MI200 series were subjected to control, and the U.S. Department of Commerce granted NVIDIA and AMD a one-year buffer period.

In contrast, the recent regulations not only cover all mainstream AI server GPUs but also eliminate the buffer period for these chip companies. In essence, companies or institutions in countries not permitted for export can only acquire AI server chips with performance potentially inferior to NVIDIA L40S or AMD MI200 series for the next few years.

Furthermore, stricter control thresholds for lithography equipment have led to the inclusion of ASML’s DUV, the 1980Di, in the control list. This equipment is primarily used in the 28 ~ 7nm process. Previously controlled products were focused on the EUV 3000 series for 7nm and below processes and the DUV 2000 series for 16/14 ~ 5nm processes.

This move indicates that the U.S. government’s desire to control semiconductor process technology has officially extended to mature processes of 28nm.

The expanded U.S. controls on AI chips and semiconductor manufacturing devices not only target China but also countries that might collaborate with Chinese institutions and businesses in AI development.

In this scenario, China is left with only two viable options to establish efficient AI computing resources: (1) designing and mass-producing AI server chips itself or (2) utilizing the computing resources of cloud service providers.

As the U.S. is also discussing the potential inclusion of cloud service providers in semiconductor control policies and currently formulating relevant countermeasures, this path remains unreliable for China. Therefore, the only dependable option is to independently design and manufacture AI server chips.

Read more

• [News] Nvidia says US speeded up new export restrictions on AI chips
• [News] Chinese Chip Equipment Rises Amid U.S. Restrictions