The New York Times has turned Epoch AI’s research into a visualization showing the sheer volume of calculations required to train a modern AI model. According to Epoch AI, training compute for frontier language models has grown at roughly 5x per year since 2020, doubling approximately every 5.2 months. This growth stems from three factors: hardware quantity (1.69x per year), training duration (1.53x per year), and hardware performance (1.41x per year). As of September 2026, Epoch AI estimated approximately 20 million AI chips in use globally, a figure that could reach 200 million by the end of 2028.
Source: Read the original article
Disclaimer: this content is for information purposes only and is not financial advice. Cryptocurrencies are highly volatile: you may lose all of your capital. Always do your own research. Legal notice
