NVIDIA Pledges Open Models and Compute Access for European Scientific Community
At the AI in Science Summit, NVIDIA announced expanded open-source resources and compute partnerships to support European researchers and startups.
Key points
- NVIDIA released over 250 open models and 1,400 software projects for scientific research.
- The company partnered with European AI factories in France, Spain, and Sweden to provide compute access.
- New open-source courseware from the NVIDIA Deep Learning Institute offers free training materials.
NVIDIA announced a comprehensive expansion of technology access for Europe’s scientific community during the AI in Science Summit in Dublin today. The initiative includes open-source software releases, educational resources, and new partnerships with national AI factories to support researchers and startups across the continent.
The announcement aligns with the European Commission’s launch of the AI in Science Enablers Initiative, a framework linked to the Resource for AI Science in Europe (RAISE). This shared framework aims to connect private-sector support with the European research community, providing access to accelerated tools, compute resources, and collaborative projects for early-career scientists and smaller labs.
Context and Background
NVIDIA stated that its pledge begins with open-source technology, releasing more than 250 open models and 250 open datasets. The company also shared over 1,400 open-source software projects spanning large language models, healthcare, robotics, climate science, and physical AI. These resources are available on Hugging Face and GitHub, allowing researchers to build agentic reasoning capabilities using NVIDIA Nemotron architectures and open weights.
Researchers can utilize scientific software such as the NVIDIA BioNeMo Agent Toolkit, PhysicsNeMo libraries, Earth-2 models, and ALCHEMI NIM microservices. These tools are designed to accelerate research across biology, chemistry, molecular design, weather, climate, energy, engineering, physics, and material sciences. The availability of these open assets aims to lower barriers for smaller labs and science-based startups that struggle with assembling tools and finding compute.
Who Is Affected
The NVIDIA Deep Learning Institute launched a new open-source courseware hub featuring hands-on materials on accelerated computing and deep learning. These courses are freely downloadable, openly licensed, and require no enrollment, fees, or deadlines. The modular design allows universities, supercomputing centers, and developer communities to adapt the content for their specific curricula or events, serving as a bridge to broader training programs.
NVIDIA is collaborating with cloud infrastructure partners and European national AI factories in France, Spain, and Sweden to open new routes for startups and small-to-medium enterprises to access sovereign European compute. The NVIDIA Inception program provides these startups with compute resources and technical support. AI2F in France was the first AI factory to collaborate with Inception for four months, demonstrating strong results, while the BSC AI Factory in Spain is rolling out its first full accelerator program after a successful pilot.
What should you do?
Researchers and startup founders in Europe should review the newly released open-source models and datasets on Hugging Face and GitHub to identify relevant tools for their projects. Institutions interested in the new courseware should visit the NVIDIA Deep Learning Institute hub to download materials for integration into their training programs. Startups seeking compute access should explore the NVIDIA Inception program and contact national AI factories in France, Spain, or Sweden for partnership opportunities.
Background: Large language models
Large language models are a specific type of neural network trained on text. Generative AI is the broader umbrella term for any system that creates new content, including images and code. Not all generative AI uses large language models, but most modern text generators do.
What to do and how to stay safe: NVIDIA
- Review the 250+ open models and 1,400+ software projects on GitHub and Hugging Face for research applicability.
- Download the free, open-source courseware from the NVIDIA Deep Learning Institute hub for team training.
- Contact national AI factories in France, Spain, or Sweden to inquire about compute access via the Inception program.
- Monitor the rollout of the BSC AI Factory accelerator program for potential startup support opportunities in 2027.
Step-by-step guide: Mobile Device Security: How to Protect Data Beyond Passwords
General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
What specific resources did NVIDIA release for European scientists?
NVIDIA released over 250 open models, 250 open datasets, and more than 1,400 open-source software projects on Hugging Face and GitHub.
Which countries are involved in the new AI factory partnerships?
NVIDIA is collaborating with national AI factories in France, Spain, and Sweden to provide compute access and support for startups.
Is there a cost associated with the new educational materials?
No, the new courseware from the NVIDIA Deep Learning Institute is freely downloadable, openly licensed, and has no enrollment fees or deadlines.
