Hugging Face and Google Cloud Strategic Partnership

Hugging Face and Google Cloud have established a strategic partnership to enable companies and developers to build their own AI using open models and open-source technologies. This collaboration focuses on four primary pillars: open science, open source, cloud infrastructure, and hardware integration.

Integration with Google Cloud Infrastructure

Google Cloud customers can now more easily train and deploy Hugging Face models using Google Kubernetes Engine (GKE) and Vertex AI. This integration allows users to leverage specialized hardware capabilities provided by Google Cloud, including:

  • TPU instances: Purpose-built AI accelerators.
  • A3 VMs: Powered by NVIDIA H100 Tensor Core GPUs.
  • C3 VMs: Powered by Intel Sapphire Rapid CPUs.

Enhancements for Hugging Face Hub Users

The partnership introduces several new experiences for users of the Hugging Face Hub throughout 2024, focusing on deployment and billing efficiency:

  • Inference Endpoints: Models can be deployed for production directly onto Google Cloud.
  • Hugging Face Spaces: AI builders can accelerate their applications using TPUs.
  • Enterprise Hub: Organizations can manage their Enterprise Hub subscription usage and billing through their existing Google Cloud accounts.

Commitment to Open Science and Open Source

The collaboration aims to amplify the accessibility of AI research and innovation. This includes accelerating the integration of the latest AI innovations into Hugging Face open-source libraries, ensuring compatibility regardless of the framework used. The partnership builds upon a history of contributions from Google, such as the Transformer and Vision Transformer architectures, as well as open-source tools like TensorFlow and JAX.

Executive Perspectives

The partnership is driven by a shared vision of making generative AI more accessible to developers.

"Google Cloud and Hugging Face share a vision for making generative AI more accessible and impactful for developers. This partnership ensures that developers on Hugging Face will have access to Google Cloud’s purpose-built, AI platform, Vertex AI, along with our secure infrastructure, which can accelerate the next generation of AI services and applications"†

— Thomas Kurian, CEO of Google Cloud

"With this new partnership, we will make it easy for Hugging Face users and Google Cloud customers to leverage the latest open models together with leading optimized AI infrastructure and tools from Google Cloud including Vertex AI and TPUs to meaningfully advance developers ability to build their own AI models"

— Clement Delangue, CEO of Hugging Face

Sources