Skip to content
Hot eventDeveloping

SID AI: LLM Output Matches Indexed Public Web Every 15 Days

1 reports1 reporting sourcesUpdated 1 h ago

Start here

AI digest

SID AI's research team estimates that large language models will generate about 15 quadrillion output tokens in 2026, enough to match the roughly 602 trillion tokens on the indexed public web every 15 days. The team says it triangulated three methods to reach a total 2026 token volume of 200–230 quadrillion, with output tokens making up about 6.5%. The report puts accumulated human digital text at around 8.7 quadrillion tokens, of which 0.6 quadrillion sits on the indexed public web. On that basis, SID AI argues LLM output passed the indexed public web by early February 2026 and the full stock of private human digital text by late summer. It projects that, if growth continues, the 15-day interval could shrink to 1–2 days by 2028.

Generated from reports · updated 50 min ago

Report timeline

Follow the reports to see the event from different sides.

November 18
  1. Hacker News · AI
    SID AI estimates LLMs now write an indexed public web's worth of output tokens every 15 days

    SID AI estimates that LLMs will generate about 15Q output tokens in 2026, enough to match the roughly 602T tokens on the indexed public web every 15 days. The report triangulates three methods to put total 2026 token volume at 200-230Q, with output tokens at about 6.5%, and estimates accumulated human digital text at around 8.7Q tokens, of which 0.6Q sits on the indexed public web. It argues LLM output surpassed the indexed public web by early February 2026 and the full stock of private human digital text by late summer, and projects the 15-day interval could fall to 1-2 days by 2028 if growth continues.

Heat for this event

Not enough continuous observations yet to draw a trend.