{"version":1,"type":"story","url":"https://digestai.news/story/ai-agents-use-5x-more-tokens-than-humans-per-openrouter-data","json":"https://digestai.news/story/ai-agents-use-5x-more-tokens-than-humans-per-openrouter-data.json","markdown":"https://digestai.news/story/ai-agents-use-5x-more-tokens-than-humans-per-openrouter-data.md","slug":"ai-agents-use-5x-more-tokens-than-humans-per-openrouter-data","headline":"AI agents use 5x more tokens than humans, per OpenRouter data","summary":"AI agents consumed **7.3 trillion tokens** in August, five times more than humans’ **1.4 trillion**, according to data from OpenRouter, a model routing platform. The figures, cited by Futurum Group CEO Daniel Newman and Andreessen Horowitz (a16z), show agents rely heavily on cached prompts—**85%** of their token usage comes from repeating existing data rather than generating new responses. OpenRouter’s internal metrics reveal agents now use **14x more tokens** than humans on its platform, while mixed (agent-human hybrid) usage surged **4.7x** over six months. The trend highlights inefficiencies in AI infrastructure, as cached tokens strain GPU memory (KV cache) and drive demand for high-bandwidth memory (HBM), which is already in short supply. Micron warns of worsening shortages in 2027–2028, and a call center consultancy found **96%** of Claude Code’s agent inputs were re-reading old conversations. Newman predicts the gap will widen further, though the data is limited to OpenRouter’s platform and lacks spending breakdowns.","keyPoints":["AI agents used **7.3T tokens** in August, five times more than humans’ **1.4T**, per OpenRouter data","**85% of agent tokens** come from cached prompts, not new generation, per Andreessen Horowitz","Memory demand for cached tokens strains GPU HBM, with Micron warning of shortages in 2027–2028"],"whyItMatters":"The explosion in cached token usage exposes inefficiencies in AI infrastructure, driving up memory costs and GPU demand. If agent reliance on repetition grows, it could accelerate hardware shortages and inflate cloud bills without proportional value.","category":{"slug":"agents","name":"Agents & Tools","url":"https://digestai.news/category/agents"},"entities":{"companies":["Futurum Group","Andreessen Horowitz","OpenRouter","McKinsey","Micron","Nvidia"],"models":["Claude Code","DeepSeek"],"people":["Daniel Newman","Peter Walker"]},"firstPublishedAt":"2026-10-03T13:10:00Z","updatedAt":"2026-10-03T13:10:00Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"Tom's Hardware","title":"AI agents use 5x more tokens than humans as cached prompts explode, headed for 10x","url":"https://tomshardware.com/tech-industry/artificial-intelligence/futurum-ceo-says-agents-use-ai-5x-more-than-humans-number-will-eventually-hit-10x-but-agents-are-mostly-rereading-what-theyve-already-seen","publishedAt":"2026-10-03T13:10:00Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":{"title":"OpenAI Astra Demand Drives Token Explosion","url":"https://digestai.news/thread/openai-temporarily-pauses-pro-plan-due-to-astra-demand","storyCount":4},"cite":{"text":"Digest AI, \"AI agents use 5x more tokens than humans, per OpenRouter data\", 3 October 2026, https://digestai.news/story/ai-agents-use-5x-more-tokens-than-humans-per-openrouter-data","publisher":"Digest AI","title":"AI agents use 5x more tokens than humans, per OpenRouter data","datePublished":"2026-10-03T13:10:00Z","url":"https://digestai.news/story/ai-agents-use-5x-more-tokens-than-humans-per-openrouter-data"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}