OpenAI Reports AI Computing Power Growing 300,000 Times Faster Than Moore's Law

OpenAI has released new analysis revealing an explosive growth in computing power used for AI training that far outpaces traditional technology trends.

According to the organization's findings, the amount of compute used in the largest AI training runs has increased by more than 300,000 times since 2012. This represents an exponential growth rate with a doubling time of just 3.4 months—dramatically faster than Moore's Law, which predicted computing power would double approximately every two years.

To put this in perspective: if AI compute had followed Moore's Law over the same period, we would have seen only a 7-fold increase instead of the 300,000-fold jump actually observed.

OpenAI emphasizes that improvements in computing power have been a critical driver of AI progress. The organization suggests that if this exponential trend continues, we should prepare