News
NVIDIA detailed public-sector deployments in Europe using open models for sovereign AI, including projects in the EU, Poland, Germany, France, Greece and the UK.
NVIDIA said open models are being used in a series of European public-sector projects aimed at keeping sensitive data and infrastructure under government control while supporting local-language AI tools. The company’s Oct. 22, 2026 blog post describes deployments and planned systems across the European Commission, Poland, Germany, France, Greece and the U.K., with several efforts centered on sovereign or on-premises infrastructure.
Evidence and context
In the European Commission example, Berlin-based deepset is expanding its Haystack Enterprise Platform as the foundation of AI@EC, the Commission’s shared platform for building and running AI applications on commission-managed infrastructure, according to NVIDIA. The update adds tools to compare operating costs, evaluate answer quality and trace problems, while NVIDIA Triton Inference Server supports shared services used to find and rank relevant information. NVIDIA said the platform builds on reusable components, role-based access, audit trails and guardrails already used for applications such as document summarization, knowledge management and multilingual policy analysis.
Poland’s eHealth Centre is building an Intelligent Services Platform for medical imaging with NVIDIA DGX BasePOD and the MONAI framework, NVIDIA said. The system is intended to help physicians analyze X-rays, CT scans and MRI images, detect abnormalities and prioritize urgent cases. In a separate Polish project, Deloitte is supporting the Ministry of Family, Labour and Social Policy on an agentic factory designed to serve more than 400 labor offices. NVIDIA said the system will run models including PLLuM, GPT OSS and Inkling for document search, summarization and digital assistants. NVIDIA also said it is working with SpeakLeash in Poland and the University of Ljubljana in Slovenia on country-specific additions to its synthetic personas program, with localized datasets planned to follow.
In the U.K., London-based Locai Labs released Nemotron-Personas-UK, a dataset of 1 million synthetic personas reflecting demographic and geographic diversity, according to NVIDIA. The company said the dataset is intended for uses such as healthcare evaluation and bias testing across age groups, education levels and rural and urban populations.
Germany features two separate efforts in NVIDIA’s post. Materna is introducing Sinfonie, an on-premises platform combining generative AI, voice interaction and 3D avatars, with NVIDIA Audio2Face used for animation. NVIDIA said the system is based on a project for a German state administration and is meant to let employees access institutional knowledge through spoken questions while keeping data on their own infrastructure. Separately, secunet, NVIDIA, Red Hat and deepset are developing a sovereign cloud offering built on secunet’s SINA Cloud foundation, which the company said includes security components authorized for use with classified information. The planned stack combines NVIDIA accelerated computing and Nemotron open models, Red Hat OpenShift AI, deepset’s Haystack AI agent orchestration and infrastructure operated by secunet in German data centers.
France’s CANUT selected LINAGORA and Docaposte to provide a sovereign collaborative workspace built on Twake.ai, NVIDIA said. The framework is available to more than 2,500 eligible organizations, including municipalities, schools and healthcare institutions, without separate tenders. The workspace includes email, document sharing and collaboration tools, with NVIDIA technology supporting speech transcription and infrastructure hosted in France. Région Occitanie is the preferred partner for deployment scaling.
In Greece, KIEFER is developing an agentic factory on NVIDIA DGX SuperPOD with DGX GB200, paired with Sophea open weight models and Greek-language tools for knowledge retrieval, speech and digital assistants, according to NVIDIA. The company said the system is designed to make institutional information easier to find and use for Greek-speaking public-sector teams.
NVIDIA also highlighted its Open Models Codefest, which it said brings together startups, software providers and labs to turn open models into locally controlled solutions for public-sector challenges such as medical reasoning, AI tutoring and model governance. The top five teams named in the post were Pleias, Zehnlab.ai, Disseqt, Algoritmica GmbH and ABA Technology. NVIDIA said Deutsche Telekom and T-Systems provided compute infrastructure for the teams to train and deploy large language models.
The post follows NVIDIA’s Oct. 21, 2026 blog item on sovereign intelligence across industries and its Oct. 15, 2026 update on the Nemotron family. Background material supplied with the source says Nemotron provides downloadable model weights, datasets and training recipes for customization on customers’ own infrastructure, and that NVIDIA supports deployment through tools including vLLM, SGLang, Ollama, llama.cpp and NIM microservices. The same background notes that Nemotron 3.5 Super VL, released Oct. 15, is a 120-billion-parameter model aimed at multimodal agent workflows, while the accompanying Bolt models target local and edge workloads.
The article provides named projects and technical components, but it does not quantify adoption beyond the figures in the source, and it does not supply timelines for completion for most of the deployments described.
Be the first to comment