Nvidia acquires Hacking Face for $12.9 billion, pledging platform independence and expanded open-model infrastructure

Nvidia announced signing an agreement to acquire the Hacking Face platform in a deal worth $12,930,300,000, a major step aimed at strengthening the infrastructure of the open-model platform and expanding developers' and organizations' worldwide access to AI technologies. The company said the platform will retain its independent brand and its leadership team led by Clem, Julian and Thomas to continue serving the global developer community.

This move follows a decade of platform growth that has made it the core hub for the software and open-model community, today relied on by more than 18 million developers, researchers and content creators to share over 3 million models, 500 thousand datasets and one million applications, while more than 200 thousand companies use it to discover, evaluate, fine-tune and deploy intelligent models in individual and enterprise work environments.

Nvidia pledged to keep the platform open and neutral toward all parties in the tech ecosystem, without requiring developers to use specific compute.The announcement confirmed that the platform will continue supporting various models, frameworks, cloud compute and inference providers, and diverse hardware platforms, and that using Nvidia accelerators will not be a mandatory condition for building or deploying on Hacking Face, ensuring ongoing multi-cloud, multi-accelerator development for different teams.

This acquisition is based on an ongoing commitment to support open-weight models, which enable startups, academic and government institutions to build advanced capabilities and match the model suited to the required task without having to train models from scratch. Nvidia is already the largest contributor of open models and data on the platform, with more than 500 models and over 250 published datasets,The integration of engineering infrastructure aims to raise the platform’s reliability, security testing and the efficiency of evaluation and inference.

For software engineering teams and organizations in the Gulf, Egypt and the Levant, this announcement gains direct operational significance given that regional initiatives and custom Arabic models rely on the platform as a primary hub for hosting and distribution. Maintaining hardware neutrality and multi-cloud diversity gives local institutions full flexibility to deploy their models across various data centers and available cloud infrastructures without fear of vendor lock-in or mandatory compute requirements.

Enhancing the infrastructure for model evaluation and accelerating inference also eases the server management burdens and security-testing costs for startups and digital-transformation teams in the region. This shift provides higher efficiency in selecting the most suitable open model for your operational needs and deploying it locally at the lowest possible inference cost, supporting faster rollout of intelligent applications in critical sectors such as finance, healthcare and government services.