AI training compute doubled every 3.4 months after 2012

In its May 16, 2018 analysis “AI and Compute,” OpenAI found that since 2012 the amount of compute used in the largest AI training runs had been doubling every 3.4 months, a total increase of more than 300,000 times. By comparison, Moore’s Law doubled roughly every two years. (OpenAI’s site blocks automated fetching; this figure is corroborated through search and independent summaries, with the canonical URL openai.com/index/ai-and-compute/.)

Sources

Last verified June 6, 2026