\n<\/p>\n
BEIJING: China\u2019s Huawei expects autonomous agents to account for more \u200cthan 90% of global AI token traffic by 2035, requiring huge increases in computing power alongside new systems to keep independent software secure and controllable.<\/strong><\/p>\n Annual consumption of AI tokens, the units models process when receiving and \u200bgenerating information, will skyrocket over the next decade as agents continuously perceive their surroundings, \u200breason, make decisions and call external tools, the Chinese tech and telecoms \u2060giant said on Wednesday.<\/span><\/p>\n Leading AI researchers and executives in the U.S. have called for slowing \u200bfrontier development after rogue agent incidents and warnings that future systems able to autonomously operate hardware and \u200bacquire resources could become difficult for humans to control.<\/span><\/p>\n Huawei\u2019s \u201cIntelligent World 2035: Turning Vision into Action\u201d report makes no call to slow development. Instead, it assumes agent capabilities and deployment will expand dramatically and lists security and \u200bprivacy protection for autonomous agents among 10 technologies needed to support this.<\/span><\/p>\n The approach reflects a \u200bwider divide over AI pacing. While parts of the U.S. technology industry are debating coordinated limits on how \u200cquickly \u2060frontier capabilities improve, China has generally sought to accelerate AI adoption across its economy, while building technical standards around emerging risks.<\/span><\/p>\n Huawei said moving AI into the physical world would be essential to achieving artificial general intelligence, or AGI, which generally refers to systems able to perform a broad \u200brange of tasks at \u200bor beyond human \u2060levels.<\/span><\/p>\n The Huawei report builds on its earlier forecast that 900 billion AI agents could be active globally \u200bby 2035, roughly 100 times the projected human population.<\/span><\/p>\n Huawei is also \u200bdeploying agents as \u2060part of a security response. It launched an autonomous security-operations platform in April using sensing and enforcement agents to detect and respond to cyber threats, and has published proposals for AI firewalls \u2060aimed at \u200bprotecting enterprise AI systems.<\/span><\/p>\nHUAWEI EXPECTS 900 BILLION AI AGENTS BY 2035<\/span><\/h2>\n