Huawei Predicts AI Agents Will Generate 90% of Token Traffic by 2035

Huawei has released projections showing autonomous AI agents will account for more than 90% of global AI token traffic by 2035, a forecast that underscores sharply different approaches to AI development between China and the United States.

The Chinese telecommunications and technology company expects roughly 900 billion AI agents to be active globally by 2035—approximately 100 times the projected human population at that time. These agents will continuously perceive their environment, reason through problems, make decisions, and interact with external tools, driving massive increases in token consumption, according to details first reported by Reuters.

Computing Infrastructure at Scale

The projected surge in autonomous agent activity will require substantial expansion of computing power alongside new security frameworks to keep independent software systems controllable. Huawei has identified security and privacy protection for autonomous agents as one of 10 critical technologies needed to support this expansion.

Huawei chairman Guo Ping told employees the company views AI as its "biggest opportunity" and aims to position its computing infrastructure as China's equivalent to Nvidia in the United States.

Contrasting Development Philosophies

Huawei's "Intelligent World 2035: Turning Vision into Action" report makes no recommendation to slow AI development. Instead, it assumes agent capabilities and deployment will expand dramatically over the next decade.

This stance contrasts with growing calls from leading AI researchers and executives in the United States to slow frontier development. These concerns have intensified following incidents involving rogue agents and warnings that future systems capable of autonomously operating hardware and acquiring resources could become difficult for humans to control.

The divergence reflects broader differences in AI policy. While segments of the US technology industry debate coordinated limits on how quickly frontier capabilities should advance, China has generally pushed to accelerate AI adoption across its economy while building technical standards to address emerging risks.

Physical World Integration

Huawei emphasized that moving AI into the physical world will be essential for achieving artificial general intelligence—systems capable of performing a broad range of tasks at or beyond human levels.

The company is already deploying agents in security applications. In April, Huawei launched an autonomous security-operations platform using sensing and enforcement agents to detect and respond to cyber threats. The company has also published proposals for AI firewalls designed to protect enterprise AI systems.

Why it matters

Huawei's projections reveal how China's largest technology companies are planning for unrestricted AI scaling at a moment when Western counterparts increasingly question whether rapid capability gains should continue unchecked. The 900 billion agent forecast—if realized—would represent a fundamental shift in computing architecture, with machines generating the vast majority of AI workload rather than serving direct human requests. For enterprise leaders, this signals both the infrastructure investments China expects to make and the security challenges that will accompany billions of autonomous systems operating with minimal human oversight.

The forecast and development roadmap were detailed in a report released by Huawei on Wednesday, with additional context provided in chairman remarks released Tuesday, according to Reuters correspondent Eduardo Baptista.