Hugging Face estimates nearly $1B in training compute shared on its Hub
Article timeline
The articles in this story. Times are ET.
merve@mervenoyannXOpen-source AI's economic value is unmatched, per Merve NoyanAIMerve Noyan of Hugging Face says the economic value created by open-source AI is unmatched. The post is a short statement with no figures of its own, and it links to a quoted post by Thomas Wolf estimating that labs including Meta, NVIDIA, Alibaba, and DeepSeek have shared close to $1B of training compute on the Hugging Face Hub.
Thomas Wolf@Thom_WolfXHugging Face Hub models represent nearly $1B in shared training computeAIThomas Wolf, Hugging Face's CEO, estimates that Meta, NVIDIA, Alibaba, DeepSeek, Moonshot, Mistral and about 30 other labs have shared close to $1B of training compute on the Hub. He says that figure is a floor, with about half from official compute numbers and the rest educated guesses, and that it excludes roughly 900k community fine-tunes. He puts the end-2023 total under $40M.
