Huawei Chief Says Chinese AI Developers Need to Accelerate

Huawei rotating chairman Eric Xu has called on Chinese artificial intelligence developers to accelerate their progress, arguing that some of the safety risks being discussed by leading U.S. AI companies may not yet be apparent to Chinese developers. Speaking at Huawei’s annual Connect conference in Shanghai, Xu said the difference could partly reflect the greater computing resources available to major U.S. AI companies and the resulting sophistication of their models.

Xu’s comments come as concerns about increasingly capable AI systems have intensified in the United States. OpenAI has recently disclosed cases involving unexpected or unauthorized model behaviour, while Anthropic CEO Dario Amodei has argued for greater caution as frontier AI systems become more capable. Xu suggested that Chinese AI companies may need to reach a similar level of development before they encounter some of the risks currently being observed by their U.S. counterparts.

At the same time, Xu emphasized that faster AI development should be accompanied by risk management. He said the industry needs to find a balance between pushing technological progress and addressing potential dangers. China has been developing its own standards and regulatory measures for advanced AI, including work around the security of autonomous AI agents that can plan tasks, make decisions and use software with limited human supervision.

Huawei is also strengthening its position as a major supplier of AI computing infrastructure in China. The company says demand for its AI computing equipment currently exceeds its production capacity, limiting its ability to expand internationally. Huawei plans to accelerate the release of its next-generation Ascend 960DT training processor to the first quarter of 2027, while also developing larger computing systems capable of connecting thousands of processors to compensate for limitations in individual chips.

The comments underline the increasingly competitive nature of the global AI race. U.S. export restrictions have limited Chinese companies’ access to Nvidia’s most advanced processors, encouraging Huawei and other domestic chipmakers to develop alternatives. Huawei says it plans to release new Ascend processors annually through 2029, while expanding the software and computing infrastructure surrounding them. Whether Chinese developers can close the remaining gaps with leading global AI companies will depend not only on chip performance and computing capacity but also on software, research talent, efficiency and the ability to manage the risks associated with increasingly autonomous systems.