BEIJING – Chinese artificial intelligence startup DeepSeek has delayed the release of a new AI model after encountering difficulties in training it with domestically produced Huawei chips, the Financial Times reported on Thursday.
The development highlights the significant hurdles China’s technology sector faces in its ambitious push to achieve self-sufficiency and replace advanced U.S. components.
According to the report, the setback for one of China’s leading AI companies underscores the performance gap that still exists between domestic hardware and the industry-leading chips supplied by U.S. firm Nvidia.
In response to stringent U.S. export controls aimed at curbing China’s access to high-performance semiconductors, Chinese tech firms have been under pressure to pivot to local alternatives. Huawei’s Ascend series of AI chips has emerged as the country’s most prominent contender to Nvidia’s dominant processors, which are the global standard for training large and complex AI models.
The reported failure by DeepSeek to effectively utilize the Huawei hardware for its latest project suggests that China’s homegrown technology and its supporting software ecosystem may not yet be mature enough to handle the most demanding AI development tasks. This can pose a significant challenge for Chinese companies striving to compete at the forefront of the global AI race.
The incident serves as a practical example of the real-world implications of the ongoing U.S.-China tech rivalry, demonstrating that Beijing’s goal of building a fully independent, high-performance technology supply chain remains a complex and ongoing challenge.
