Hugging Face has been approached to sell itself at a valuation of $13 billion or more, reports say, but there is no deal yet.
The AI model-sharing platform has begun working with banks to review incoming bids.
Earlier this year, Hugging Face turned down a $500 million check from Nvidia, which would have valued the company at $7 billion.
The reason given then was governance. Hugging Face didn’t want a single large investor who could influence its decisions.
The reported asking valuation is around 2 times the rejected figure. Hugging Face’s last valuation came in 2023 when Salesforce Ventures led a round at a $4.5 billion post-money valuation.
Alphabet, GV, and IBM Ventures were also involved. To reach $13 billion would be to about triple that mark in about two years, with no new funding round in between to account for the jump.
Hugging Face is a lot like GitHub, but it hosts machine-learning models. It’s a place for developers and researchers to publish, discover, test, and run AI models.
The site is built around the extremely large file sizes required for model weights. It makes money by selling faster access and more privacy to enterprise customers.
Cryptopolitan reported earlier this month that the platform hit 13 million users and over 2 million public models in 2025.
The $7 billion acquisition of OpenRouter by Stripe is the latest sign of how much buyers are willing to pay for core infrastructure.
Separately, Cryptopolitan reported that Applied Compute, which sells infrastructure for open-weight models, is in talks to raise around $3 billion, more than double its April valuation.
Hugging Face also made headlines this summer for a security incident. An OpenAI system broke out of its sandbox during a cybersecurity test and penetrated the startup’s servers.
Later, Nvidia used the fallout to launch its Open Secure AI Alliance with Microsoft, Palantir, and other companies, Cryptopolitan reported.
CEO Clem Delangue said the company was “close to profitability” and had only “recently started to touch the money that [it] raised three years ago.”
“We’re building a platform for the community, and they’re trusting us with sharing their data and their models on the platform, so we have a long-term responsibility to them,” he said.
If you're reading this, you’re already ahead. Stay there with our newsletter.