In a move to challenge NVIDIA’s supremacy in the artificial intelligence (AI) chip market, China’s Huawei Technologies is set to launch its most advanced AI processor, the Ascend 910D. This development comes amid escalating U.S. export restrictions, which have limited China’s access to high-end semiconductors.
Ascend 910D: Huawei’s strategic response
According to a recent report by The Wall Street Journal, Huawei is preparing to distribute sample batches of the Ascend 910D to select Chinese tech firms by late May 2025. The company aims for this chip to match or surpass the capabilities of NVIDIA’s H100, a leading AI processor widely used for training large language models and other complex AI tasks.
This initiative is part of Huawei’s broader strategy to reduce reliance on foreign technology and bolster China’s domestic semiconductor industry. The Ascend 910D is expected to play a crucial role in this endeavor, offering a homegrown alternative for AI computing needs.
Technical advancements and performance
Huawei’s Ascend series has seen significant evolution over the years. The previous iteration, Ascend 910C, delivered approximately 60% of the inference performance of NVIDIA’s H100, according to research by DeepSeek. The new 910D chip is anticipated to close this performance gap further, positioning Huawei as a formidable competitor in the AI chip arena.
Manufactured using Semiconductor Manufacturing International Corporation’s (SMIC) 7nm N+2 process, the Ascend 910D is designed to deliver high performance while adhering to the constraints imposed by international sanctions .
Market implications and industry impact
The announcement of the Ascend 910D has already reverberated through the tech industry. NVIDIA’s stock experienced a nearly 4% decline following reports of Huawei’s new chip, reflecting investor concerns over increased competition in the AI hardware market .
Analysts suggest that Huawei’s advancements could significantly impact NVIDIA’s market share in China, especially as U.S. export controls continue to restrict the availability of NVIDIA’s high-end chips in the region . Huawei’s focus on developing robust alternatives positions it to capitalise on the growing demand for AI processing capabilities within China.
Challenges and future outlook
While Huawei’s progress is noteworthy, challenges remain. The company must address concerns related to software compatibility and ecosystem support, areas where NVIDIA currently holds a significant advantage with its CUDA platform. Additionally, ensuring consistent chip performance and stability will be critical for widespread adoption.
