2 min read
Add as a preferred source on Google

NVIDIA and EBU Expand AI Framework for European Broadcasters

The work adds cloud, model and integration partners. A separate Adobe project reports faster Frame.io processing and a smaller search proxy.

A graphics card sits under glass beside a video drive. / TokenPost.ai
A graphics card sits under glass beside a video drive. / TokenPost.ai

The European Broadcasting Union (EBU) and NVIDIA announced Oct. 19 that their AI framework for public media is gaining cloud, model and integration partners to give broadcasters more control over AI deployments.

The work is intended to produce requirements and reference architectures for where broadcasters run AI workloads, how they govern data, which models they select and how they secure systems across the lifecycle. Cloud providers will map the standards to deployment options as the framework develops.

Microsoft is contributing expertise from Microsoft Sovereign Cloud, including Azure Local and Foundry Local. Nebius provides a platform for training, fine-tuning and running models on European infrastructure. NVIDIA Nemotron open models and Qvest integration work are also part of the initiative.

The effort extends a partnership the EBU and NVIDIA announced in June 2025 to develop sovereign AI and cloud technologies for public service media. The earlier collaboration included the EBU’s Dynamic Media Facility and Media eXchange Layer architectures among its foundations. “These standards give EBU members a shared foundation to strengthen resilience, retain control and deliver accountable advanced services that sustain audience trust,” said Antonio Arcidiacono, the EBU’s chief technology and innovation officer.

Separately, Adobe is using NVIDIA hardware to decode video, with TensorRT and CUDA supporting preprocessing and search-embedding generation in Frame.io. In early results from selected workloads, embedding generation ran about five times faster, while 95th-percentile processing latency improved fourfold to fivefold. A new search proxy running at two frames per second was about 80% smaller than the 360p version it replaces.

Frame.io can find footage by concept, action or setting, then combine results with notes, transcripts and approval status. The EBU framework focuses on governance and deployment choices, while the Frame.io project applies AI tools to processing and searching video archives.

German producer UFA delivers more than 3,500 hours of programming annually, including the daily drama “Gute Zeiten, schlechte Zeiten.”

Simon Yoon

Reporter

Simon Yoon reports on blockchain technology for TokenPost. Send corrections or tips to info@tokenpost.com.

Loading…