Chinese technology company Huawei Technologies became the latest casualty of US sanctions against AI advancements of China-native companies. The United States policies now stand in the way of the company accessing advanced AI chip manufacturing equipment to develop a state-of-the-art AI processor.
Huawei’s chips were designed to challenge Nvidia’s market dominance. However, it can’t proceed with its ambitious designs as the US prohibits its chipmaking partners from procuring EUV lithography machines from ASML.
The Huawei situation is just another facet of a competition that goes beyond technology. While this episode is about the intensifying geopolitical tussle for semiconductors and AI, the overarching plot concerns global influence and economic leadership.
Huawei faces mounting challenges in acquiring AI chips
Huawei set out to develop two Ascend chips in reply to Nvidia’s accelerators. However, the company may need to readjust its schedule. It is limited to using 7nm chips, which is considered dated in a semiconductor industry that is moving very fast.
Some have even speculated that it might be stuck with these chips until 2026, which will cast it further adrift of its contemporaries using more advanced 2nm chips.
Huawei’s problems primarily arise from its inability to procure EUV lithography equipment from ASML, a Dutch company that produces most of the world’s lithography equipment. Also, its major chip manufacturing partner, Semiconductor Manufacturing International Corp (SMIC), is having trouble producing enough chips to meet demands.
US sanctions are not specifically targeting Huawei; rather, they are aimed at China. Huawei just happens to be in the crosshairs. The company is heavily invested in R&D, which is integral to China’s success in the global semiconductor market. An inefficient Huawei is a blow to China and its ambition to be a technological competitor to the US.
Chinese companies have limited or no access to critical technology from key suppliers like Nvidia, ASML, and Applied Materials that could help improve the country’s AI shares.
SenseTime, a Hong Kong-based AI company, is struggling to obtain high-end GPUs to power its deep learning models. Baidu, a Chinese company heavily invested in AI and autonomous driving, has resorted to using less advanced chips, which has a knock-on effect on the efficiency and performance of its AI models.
Chinese AI chip startup Cambricon also faces challenges exporting its chips, making it difficult to compete on a global scale.
China’s workaround to technological limitations
In response to the sanctions, China has come up with novel ways to remain relevant in the AI and semiconductor market. Its plan includes a combination of using existing but older ASML lithography machines with a multi-patterning technique. This process requires the machines to perform about four exposures on the silicon wafer, with a margin of error as small as a hundredths of the diameter of a human hair. However, this is not as efficient as using the advanced EUV lithography and is prone to errors and yield losses.
According to a research analyst at Yole Group, “Multi-patterning inherently introduces more process steps, increasing the risk of defects and variability. Additionally, the higher complexity and cost of multi-patterning make it less economically viable for high-volume production of advanced nodes like 5nm.”
The Chinese government is also acquiring semiconductor-related intellectual property, and it’s doubling down on efforts to build domestic production to bolster innovation and reduce reliance on foreign technology.
However, the stakes are high. It took decades for other chip manufacturers to perfect the technology to produce advanced chips. While ASML is gearing up to produce 2nm chips, its Chinese competitors are still stuck on 7nm technology; it would take several years to close that gap, by which time AI chip giants might have moved on to more sophisticated technology.
China will not give up its position in the AI market
The AI market is expected to grow to almost a trillion by 2030, and despite the sanctions, Chinese AI companies have managed to contribute to this growth. For instance, Chinese AI startup Moonshot is currently valued at $2.5 billion, following a billion-dollar round led by Alibaba. Similarly, 01.AI achieved a billion-dollar valuation after its recent round.
Cryptopolitan also reported on potential ways that Chinese companies affected by these sanctions imposed because of geopolitical tensions may achieve their goals regardless.
The incoming Trump administration is expected to be more overt in maintaining the United States’ leadership in emerging technologies, particularly those with strategic military applications. Targeting key players in China’s tech ecosystem is a low-hanging fruit in how the US could slow China’s march toward technological parity.
Land a High-Paying Web3 Job in 90 Days: The Ultimate Roadmap
Source: https://www.cryptopolitan.com/huawei-latest-casualty-us-ai-policies/