iCreatAI News
Skip to content
Trending storyWatching

SID Study Estimates LLM Output Token Volume Versus Human Text

1 reports1 sources2 days agoUpdated

Understand this story

AI overview

SID AI released a study estimating that in 2026, large language models worldwide processed roughly 200–230Q tokens, of which about 6.5% were output tokens, totaling about 15Q—equivalent to producing the text volume of an indexable public internet every 15 days. The study compares LLM output tokens with the volume of human-generated digital text. SID AI says its findings found broad evidence that a viral claim on X was directionally correct, and even conservative.

AI overview based on source reports · Updated 2 days ago

Coverage timeline

Follow the coverage to explore different perspectives.

Nov 18
  1. Hacker News AI
    AI is writing a public web every 15 days

    The paper estimates that LLMs will process roughly 200–230 quadrillion tokens in 2026, about 6.5% of them output tokens (~15Q), and that 2026 output tokens alone surpassed the ~0.6Q tokens on the indexed public web by early February, implying an indexed-public-web's worth of output every 15 days. It puts the accumulated stock of human-generated digital text at about 8.7Q tokens (0.6Q indexed public web, 8.1Q private), which LLM output surpassed by late summer, and treats China's reported 500T tokens per day as technically plausible but hard to verify.

Activity for this story

Current comparable activity
3
Peak comparable activity
10Nov 19 00:00
Change in the past 24 hours
-49%

Trends compare the same fully observed participants, which may cover fewer participants than current activity. Move the pointer or tap the chart to inspect hourly activity; use the left and right arrow keys with a keyboard.