At a Berlin media event, NVIDIA and the European Broadcasting Union announced an expansion of the sovereign AI framework they introduced last year for European public service media. The initiative now includes sovereign cloud providers, open model developers, and systems integrators, with the goal of defining how broadcasters can deploy AI agents while retaining control over data, infrastructure, models, and security. Microsoft will contribute expertise from its Sovereign Cloud, including Azure Local and Foundry Local, while Nebius offers a platform for training and running models on European infrastructure. NVIDIA's open Nemotron models are positioned as a customizable foundation, and Qvest and other integrators will help make the reference architectures practical in real broadcast environments.
Separately, Adobe highlighted new GPU-accelerated capabilities in Frame.io that help creative teams search video faster. The pipeline uses NVIDIA NVDEC for decoding, TensorRT for inference, and CUDA for preprocessing and embedding generation. According to NVIDIA, this results in roughly 5x faster embedding generation, up to 4x higher throughput, and up to 5x lower latency for search processing. A new two-frame-per-second proxy, about 80% smaller than the 360p version it replaces, reduces storage and transfer costs while preserving visual detail needed for search.
For UFA, a German film and TV producer delivering more than 3,500 hours of programming a year, the speed matters when editors need to find a specific reaction or unresolved exchange from footage as old as 15 years. Frame.io's semantic search locates moments by concept, action, or setting, and hybrid search adds notes, transcripts, and approval status. An editor can search for "a character looking concerned," narrow by character name, filter to approved footage, and move directly into review. The work is part of a broader Adobe and NVIDIA collaboration spanning 3D, Firefly Foundry, and digital twin workflows.