EBU expands sovereign AI framework with NVIDIA & Adobe

The European Broadcasting Union is working with NVIDIA to expand its sovereign framework for European public service media. Adobe has also unveiled new video search work with NVIDIA that is being used by German producer UFA.

The expanded framework adds sovereign cloud providers, open model developers and systems integrators.

The move extends standards introduced last year for European broadcasters. It is intended to define how media organisations can deploy agents while keeping control of data, infrastructure, models and security.

The alliance represents more than 115 member organisations in more than 55 countries and reaches more than 1 billion people. As agents move into newsrooms, production and audience services, broadcasters need a foundation that protects editorial independence, data sovereignty and national resilience.

The expanded initiative will set requirements and reference architectures covering where workloads run, how data is governed, which models are used and how systems are secured throughout their lifecycle.

"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, Chief Technology and Innovation Officer at the EBU.

Microsoft is contributing expertise from Microsoft Sovereign Cloud, including Azure Local and Foundry Local, to support local processing of sensitive editorial material and continuity when external connectivity is unavailable. Nebius is providing a platform designed to let broadcasters train, fine-tune and run their own models on European infrastructure.

Open models are another part of the framework. The organisations involved said this approach allows broadcasters to inspect, evaluate and adapt systems for local languages, archives and editorial policies, with NVIDIA Nemotron models as one possible base for customisation.

Qvest and other systems integrators are also involved. Their role is to turn the reference architectures into systems that can be deployed in broadcast environments.

"Sovereign cloud infrastructure and open models can help Europe's public service media innovate while protecting data, editorial independence and cultural diversity," said Delphine Ernotte, Chief Executive Officer of France Télévisions and President of the EBU.

Video search

Separately, Adobe outlined updates to Frame.io aimed at helping creative teams search video libraries faster. The work uses NVIDIA GPUs for video decoding, AI inference and preprocessing.

According to the companies, the search pipeline uses NVIDIA NVDEC for video decoding, TensorRT for AI inference and CUDA for preprocessing and embedding generation. Adobe is also using NVIDIA NVENC and NVDEC in media ingest to improve video encoding and decoding workflows.

Adobe said the changes deliver about five times faster embedding generation, up to four times higher throughput and up to five times lower latency for search processing. It also introduced a new two-frame-per-second proxy that is about 80% smaller than the 360p version it replaces.

That work is relevant for UFA, one of Germany's larger film and television producers, which delivers more than 3,500 hours of programming each year. Its productions include the daily drama "Gute Zeiten, schlechte Zeiten."

Editors working on recaps or flashbacks may need to locate a specific reaction or exchange in footage that is up to 15 years old. Adobe said Frame.io semantic search lets users find moments within video by concept, action or setting, while hybrid search adds notes, transcripts and approval status.

For example, an editor could search for "a character looking concerned", then narrow the results by character name and approved footage before moving into review. Adobe said the work is part of a wider collaboration with NVIDIA spanning Adobe Substance 3D, Firefly Foundry and digital twin workflows.

The two announcements reflect a wider push across European media and creative industries to apply AI tools in production and archive work while maintaining tighter control over data and operational infrastructure.