Clem, cofounder and ceo of Hugging Face here.
Wanted to share our new collaboration with Google Cloud. Every day, over 1,500 terabytes of open models and datasets are downloaded and uploaded between Hugging Face and Google cloud by millions of AI builders. We suspect it generates over a billion dollars of cloud spend annually already.
So we’re excited to announce today a new partnership to:
- reduce Hugging Face model & dataset upload and download times through Vertex AI and Google Kubernetes Engine thanks to a new gateway for Hugging Face repositories that will cache directly on Google Cloud
- offer native support for TPUs on all open models sourced through Hugging Face
- provide a safer experience through Google Cloud’s built-in security capabilities.
Ultimately, our intuition is that the majority of cloud spend will be AI related and based on open-source (rather than proprietary APIs) as all technology builders will become AI builders and we're trying to make this easier.
Hugging Face just released HuggingChat Omni, that selects the best model for every prompt automatically amongst 115 open-source models including gpt-oss, deepseek, qwen, kimi, smolLM, gemma, aya and many more already!
Omni uses a policy-based approach to model selection thanks to a model router called Arch-Router from Katanemo.
Introducing Gaia2, the follow-up to the agentic benchmark GAIA, allowing analysis of considerably more complex behaviors. Gaia2 is released with the open Meta Agents Research Environments (ARE) framework to run, debug and evaluate agents. ARE simulates complex real world-like conditions and can be customized to further study agents behaviors. Gaia2 dataset is released under CC by 4.0 license, and ARE under MIT license.
Xet might be one of the most important AI technology that nobody is talking about!
Under the hood, it now powers 5M Xet-enabled AI models & datasets on HF which see hundreds of terabytes of uploads and downloads every single day.
What makes it super powerful is that it massively speeds up & reduces costs of data transfer thanks to methods like content-defined chunking (CDC). Instead of treating a file as an indivisible unit, CDC breaks files down into variable-sized chunks, using the data to define boundaries.
That's what allows Hugging Face to offer a platform for 10 million AI builders in open-source at a fraction of the cost.
Hugging Face just released a VS Code extension to run Qwen3 Next, Kimi K2, gpt-oss, Aya, GLM 4.5, Deepseek 3.1, Hermes 4 and all the open-source models directly into VSCode & Copilot chat.
Open weights means models you can truly own, so they’ll never get nerfed or taken away from you!
So we’re excited to announce today a new partnership to: - reduce Hugging Face model & dataset upload and download times through Vertex AI and Google Kubernetes Engine thanks to a new gateway for Hugging Face repositories that will cache directly on Google Cloud - offer native support for TPUs on all open models sourced through Hugging Face - provide a safer experience through Google Cloud’s built-in security capabilities. Ultimately, our intuition is that the majority of cloud spend will be AI related and based on open-source (rather than proprietary APIs) as all technology builders will become AI builders and we're trying to make this easier.
Questions, comments, feedback welcome!