Highlights:

  • Hugging Face operates a platform bearing the same name that developers utilize to exchange open-source AI models.
  • According to Hugging Face, users can upload their training dataset to AutoTrain, which will then autonomously identify the most fitting open-source AI model for their project.

Hugging Face Inc., a startup running a platform for hosting open-source AI projects, has just secured USD 235 million in funding.

Prominent tech industry players like Amazon.com Inc., Google LLC, IBM Corp., and Salesforce Inc. were among the contributors to the Series D funding round. Intel Corp., Nvidia Corp., Qualcomm Inc., Advanced Micro Devices Inc., and Sound Ventures also contributed to the funding round.

Recently, reports indicated that Hugging Face was considering raising new capital. The latest raise of the startup was reportedly led by Salesforce, as revealed by The Information. Following the new investment, Hugging Face’s valuation has surged to USD 4.5 billion, more than double its value from the previous funding round in May.

Hugging Face operates a platform bearing the same name that developers utilize to exchange open-source AI models. Its interface resembles that of GitHub. The Hugging Face AI model page is structured into three primary sections: the first provides a model description, the second grants access to its code, and the third encompasses a user forum.

Developers depend on Hugging Face not only for sharing AI models but also for exchanging associated technical resources. As per the company, its platform hosts numerous training datasets, enabling companies to construct new neural networks. Additionally, it provides access to what it calls “Spaces,” which are sample applications showcasing practical applications of the AI models available on its platform. In total, Hugging Face accommodates over 120,000 AI models and 20,000 datasets.

The company makes money by offering a paid version of its Enterprise Hub platform. This offering enables companies to establish an internal Hugging Face environment tailored for their software teams. Within this environment, developers can seamlessly share AI models and associated assets with colleagues while gaining access to files from the public version.

In addition to Enterprise Hub, the company provides two other commercial products to streamline AI model development.

The initial offering is named AutoTrain. According to Hugging Face, users can upload their training dataset to AutoTrain, which will then autonomously identify the most fitting open-source AI model for their project. Once a neural network is chosen, the service trains it using the user-provided dataset.

Developers can upload the AI models they generate using AutoTrain to a Hugging Face service called Inference Endpoints. Inference Endpoints offers a managed cloud environment tailored for hosting neural networks, alleviating users from the usual manual setup tasks associated with AI inference infrastructure.

Clement Delangue, the Co-founder and CEO, informed a famous media house that the startup intends to allocate its latest funding for “more open source AI and platform building.” To drive these initiatives forward, the startup plans to bolster its existing 170-person workforce.