NVIDIA Backs Europe’s RAISE Science Push With 35 Supercomputers
NVIDIA joins the EU's RAISE-linked science initiative with open models, free training, startup compute and 35 supercomputers for 3 million researchers.
Summary
On October 15, 2026, the European Commission launched the AI in Science Enablers Initiative at Dublin’s AI in Science Summit. Linked to the Resource for AI Science in Europe, or RAISE, it connects researchers with private sector tools, compute, data, training, technical support, collaborative projects and startup assistance. NVIDIA joined with more than 250 open models, 250 open datasets and 1,400 open source software projects on Hugging Face and GitHub. Its research stack includes NVIDIA Nemotron architectures, weights, models, data and recipes, plus BioNeMo Agent Toolkit, PhysicsNeMo libraries, Earth-2 models and ALCHEMI NIM microservices for biology, chemistry, molecular design, climate, energy, engineering, physics and materials. Commissioner Ekaterina Zaharieva said the campaign converts Europe’s ambitions into concrete commitments.
NVIDIA Deep Learning Institute also opened free, openly licensed courseware for accelerated computing and deep learning, with no enrollment or deadline and adaptable for universities, supercomputing bootcamps and developer events. NVIDIA and European AI factories are expanding sovereign compute access for startups and SMEs: France’s AI2F has worked with NVIDIA Inception for four months, Spain’s BSC AI Factory is launching its first full accelerator, with replication interest for 2027, and Sweden AI Factory, one of 19 EuroHPC-funded facilities, will serve sectors including life sciences, materials, autonomous systems and gaming. CINECA and JSC provide free NVIDIA GPU access for open research, supplemented by office hours, bootcamps and hackathons. An October 2026 European AI hackathon with Austrian Scientific Computing, Leibniz Supercomputing Centre, High-Performance Computing Center Stuttgart and The Cyprus Institute supports Pharos-CY. NVIDIA is powering 35 supercomputers under development, Europe’s largest single-year capacity buildout, intended to serve more than 3 million researchers.
Positives
- More than 250 open models, 250 datasets and 1,400 software projects give researchers reusable resources through Hugging Face and GitHub.
- Deep Learning Institute courseware is free, openly licensed and adaptable for universities, supercomputing centers and developer communities.
- CINECA and JSC provide academia and startups with free NVIDIA GPU infrastructure and services for open research.
- Sweden AI Factory will open infrastructure and expertise to Inception startups and SMEs across life sciences, materials, autonomous systems and gaming.
- Thirty-five supercomputers under development are intended to equip more than 3 million European researchers with next-generation infrastructure.
Risks & concerns
- Smaller laboratories, early-career scientists and science startups still face barriers assembling tools, obtaining compute and developing expertise.
- All 35 supercomputers remain under development, so the planned capacity is not yet fully available to researchers.
- AI2F’s four-month collaboration is described as producing strong results, but no performance figures are provided.
- BSC AI Factory is only beginning its first full accelerator, while wider replication remains prospective for 2027.