Open models for public institutions
NVIDIA outlined projects aimed at helping Europe's public sector adopt open models. Open models let institutions adapt a model to local languages and services, run it on infrastructure they control and keep sensitive data in controlled environments.
Highlighted projects
- European Commission: Berlin-based deepset is expanding its Haystack Enterprise Platform, the foundation of the Commission's AI@EC platform. The update connects agent development, execution, optimization and governance, adding capabilities to compare costs, evaluate answer quality and trace problems. NVIDIA Triton Inference Server supports information retrieval and ranking services.
- Poland: Deloitte is building agentic systems for healthcare and employment services. The eHealth Centre is developing its Intelligent Services Platform for medical imaging with NVIDIA DGX BasePOD and the MONAI framework. The Ministry of Family, Labour and Social Policy is building an agentic factory to serve more than 400 labor offices, running models including PLLuM, GPT OSS and Inkling for document search, summarization and digital assistants.
- UK: Locai Labs released Nemotron-Personas-UK, a dataset of 1 million synthetic personas reflecting the country's demographic diversity. Uses include healthcare evaluation and bias testing across age, education and rural-urban groups.
- Germany: Materna is introducing Sinfonie, an on-premises platform combining generative AI, voice interaction and 3D avatars. secunet, NVIDIA, Red Hat and deepset are developing a sovereign cloud offering for organizations handling classified information.
Why it matters
The projects show that public services can run AI with local languages and data without depending on outside providers. NVIDIA is also working with SpeakLeash in Poland and the University of Ljubljana in Slovenia on country-specific synthetic persona datasets.
