China's Huawei forecasts billions of agents will dominate AI traffic by 2035

• Huawei expects 900 billion AI agents by 2035 (roughly 100 times the projected human population), requiring massive increases in computing power
• The company lists autonomous-agent security and privacy protection among 10 priority technologies needed to support this expansion
• Huawei views AI as its 'biggest opportunity' and aims to make its computing infrastructure China's equivalent of Nvidia, having already deployed autonomous security platforms

Summary

Key Forecast: Huawei predicts autonomous AI agents will generate over 90% of global AI token traffic by 2035, with approximately 900 billion AI agents active worldwide—roughly 100 times the projected human population.

Main Company: Chinese tech giant Huawei released its "Intelligent World 2035: Turning Vision into Action" report on Wednesday, outlining its vision for dramatic AI expansion over the next decade.

Core Projections:

• AI token consumption (units AI models process) will skyrocket as agents continuously perceive, reason, and make decisions
• Massive increases in computing power will be required
• Security and privacy protection for autonomous agents listed among 10 critical technology priorities

Strategic Positioning: Huawei Chairman Guo Ping identified AI as the company's "biggest opportunity," aiming for its computing infrastructure to become China's equivalent to U.S. chip leader Nvidia. The company launched an autonomous security-operations platform in April and has proposed AI firewalls for enterprise protection.

Market Implications: The forecast highlights a significant U.S.-China divide on AI development philosophy. While U.S. researchers and executives debate slowing frontier AI development due to control concerns, China—represented by Huawei's roadmap—assumes rapid acceleration and widespread AI adoption across its economy, focusing instead on building technical standards around emerging risks.

Technology Focus: Huawei emphasizes moving AI into the physical world as essential for achieving artificial general intelligence (AGI)—systems capable of performing broad tasks at or beyond human levels. The report makes no calls to slow development, contrasting sharply with growing U.S. caution about autonomous systems operating hardware and acquiring resources independently.