AI and compute
OpenAI Blog
Compute used in the largest AI training runs has doubled every 3.4 months since 2012, roughly 7 times faster than Moore's Law's historical 2-year doubling cycle. The total increase over this period exceeds 300,000x, compared to a 7x increase if Moore's Law had applied. This acceleration suggests preparation is needed for AI systems substantially more capable than those available today.
Why it matters
We’re releasing an analysis showing that since 2012, the amount of compute used in the largest AI training runs has been increasing exponentially with a 3.4-month doubling time (by comparison, Moore’s Law had a 2-year doubling period)[^footnote-correction]. Since 2012, this metric has grown by more than 300,000x (a 2-year doubling period would yield only a 7x increase). Improvements in compute have been a key component of AI progress, so as long as this trend continues, it’s worth preparing for the implications of systems far outside today’s capabilities.
