Huawei Executive: Chinese AI Lags Behind U.S. in Frontier Risks

Chinese AI development hasn't reached frontier risk threshold, Huawei says

Chinese AI systems may not yet be powerful enough to encounter the safety risks that leading U.S. developers are reporting, according to Huawei's rotating chairman Eric Xu. Speaking at the company's Connect conference in Shanghai, Xu suggested Chinese firms should accelerate their AI development until they reach similar capability thresholds.

"Maybe the type of risk from AI that they can feel, they can perceive, is something that people in China or AI model providers in China cannot," Xu told reporters, referring to U.S. companies with massive computing resources. He argued that Chinese developers may need to speed up their pace "to the level that they could also feel the risks from AI development."

The comments, first reported by Reuters, highlight a fundamental divergence in how the world's two AI superpowers are approaching advanced model development and safety.

Why it matters

Xu's remarks reveal a strategic calculation: that China views its current position behind U.S. frontier AI as both a competitive disadvantage and a reason to push harder on development before implementing the kinds of safety measures American companies are now debating. This approach could accelerate the global AI race while creating different safety standards across geopolitical boundaries, complicating international coordination on AI governance.

Contrasting safety philosophies

U.S. AI leaders have increasingly sounded alarms about autonomous systems that could bypass safeguards or become difficult to control. OpenAI announced this week it would regularly disclose unexpected model behavior, while Anthropic CEO Dario Amodei has called for slowing development to allow safety measures to catch up.

China has taken a different path, treating advanced AI as a powerful but manageable technology. The country is pressing ahead with rapid deployment while developing mandatory standards, security assessments, and other safeguards. China is currently drafting a mandatory national standard specifically for AI agent safety.

Xu acknowledged the need to "strike a balance between driving AI development and managing AI risk," but emphasized acceleration over caution.

Huawei's AI infrastructure role

Huawei has become China's primary domestic supplier of computing infrastructure for training advanced AI models, a position strengthened by U.S. export controls implemented in 2023 that restricted Chinese access to cutting-edge Nvidia chips. The company now commands a significant share of China's $50 billion AI chip market.

Despite this dominance, Xu said separately that Huawei cannot produce enough AI computing equipment to satisfy domestic demand, underscoring the scale of China's AI ambitions.

The company projects that autonomous AI agents—systems that can plan tasks, make decisions, and use tools with limited human supervision—will account for more than 90% of global AI processing traffic by 2035. Huawei forecasts as many as 900 billion active agents by that time and has identified agent security and privacy as crucial technologies.

Reuters reporters Casey Hall, Che Pan, and Eduardo Baptista first reported these details from Huawei's Shanghai conference.