3 Sources
[1]
Huawei Used TSMC, Samsung, SK Hynix Components in Top AI Chips
Huawei Technologies Co. used advanced components from Asia's largest technology firms in at least some of its leading Ascend AI processors, a research firm discovered during teardowns, highlighting China's reliance on foreign hardware as it works to boost domestic production of AI
[2]
What's Really Inside Huawei's 'Domestic' AI Chips? - Samsung Electronics Co (OTC:SSNLF), NVIDIA (NASDAQ:NVDA)
Huawei's latest AI chips, promoted as a milestone in China's push for technological self-reliance, continue to rely on critical components from Taiwan Semiconductor Manufacturing Co. (NYSE:TSM), Samsung (OTC:SSNLF), and SK Hynix, a teardown by TechInsights reveals. The findings highlight China's
[3]
Huawei used Samsung, TSM and SK hynix's parts in its advanced AI chips
Huawei Technologies used advanced components from Asia's largest technology companies in at least some of its advanced Ascend AI processors, Bloomberg News reported, citing TechInsights. TechInsights discovered gear from Taiwan Semiconductor Manufacturing (NYSE:TSM), Samsung Electronics
Share
Copy Link
Teardown analysis of Huawei's Ascend AI processors shows use of components from TSMC, Samsung, and SK Hynix, highlighting China's ongoing dependence on foreign technology despite push for self-reliance.

A recent teardown analysis by research firm TechInsights has revealed that Huawei Technologies Co., China's tech giant, continues to rely on advanced components from major Asian technology firms in its leading Ascend AI processors. This discovery highlights the ongoing challenges China faces in its push for technological self-reliance, particularly in the semiconductor industry
1
.TechInsights found that Huawei's third-generation Ascend 910C chips contain components from Taiwan Semiconductor Manufacturing Co. (TSMC), Samsung Electronics Co., and SK Hynix Inc. Specifically, TSMC produced the dies powering the processors, while Samsung and SK Hynix supplied older-generation high-bandwidth memory (HBM2E) in separate chip samples
2
.This reliance on foreign components comes despite Huawei's promotion of these chips as a milestone in China's technological advancement. The findings underscore the complexities of the global semiconductor supply chain and the challenges faced by Chinese companies in developing fully domestic alternatives.
The situation is further complicated by U.S. export controls imposed in 2020, which have severely restricted Huawei's access to advanced chips and manufacturing equipment. Both Samsung and SK Hynix have emphasized their compliance with these regulations, stating that they halted direct sales to Huawei following the implementation of these controls
2
.To circumvent these restrictions, Huawei has been leveraging its stockpiled inventory. Analysts estimate that the company acquired nearly 3 million TSMC dies through an intermediary, Sophgo, before sanctions severed this channel. These stockpiled components continue to power the 910C chips, which entered mass shipment earlier this year
2
.However, Huawei faces potential supply chain bottlenecks, particularly in high-bandwidth memory. SemiAnalysis projects that the company could face shortages by the end of 2025, as domestic players like Changxin Memory Technologies (CXMT) race to bridge the gap
3
.Related Stories
Despite these challenges, Huawei is pressing ahead with efforts to rival industry leaders like Nvidia in AI performance. At its annual Huawei Connect conference, the company unveiled a three-year plan to cluster large numbers of Ascend processors via a new UnifiedBus interconnect system, claiming superior data transfer speeds compared to Nvidia's upcoming NVLink144
2
.The revelations about Huawei's AI chips underscore the complex interplay between technological ambition, geopolitical tensions, and global supply chains in the semiconductor industry. As China continues its push for technological sovereignty, the ability of companies like Huawei to overcome supply chain vulnerabilities and develop truly domestic alternatives will be crucial in determining the future landscape of the AI chip market.
Summarized by
Navi
19 Oct 2024•Policy and Regulation

27 Oct 2024•Policy and Regulation

19 Nov 2024•Technology

1
Technology

2
Technology

3
Policy and Regulation
