ATLAS focused on autonomous navigation and 3D reconstruction for minimally invasive surgery; MICROCARD applied GPU-based solvers to cardiac modeling.
OROBIX SRL
Italian AI and GPU computing SME applying computer vision and high-performance computing to scientific data, surgical robotics, and cardiac simulation.
Their core work
Orobix is an Italian AI and computer vision SME based in Bergamo that applies machine learning, GPU computing, and 3D reconstruction techniques to demanding scientific and medical domains. Their contributions span from processing large-scale astronomical data and building virtual observatory tools, to enabling autonomous navigation for surgical robots, to accelerating cardiac electrophysiology simulations on GPU architectures. They act as a specialized technology provider, bringing AI and high-performance computing capabilities into large research consortia that need these skills but lack them in-house.
What they specialise in
Both MICROCARD (GPU solvers, exascale computing) and ESCAPE (big data, data lake processing) required HPC expertise.
ESCAPE project contributed to EOSC and virtual observatory tools for major European research infrastructures (CERN, ESO, SKA).
ATLAS project specifically required 3D reconstruction and autonomous navigation for continuum robotics in surgical settings.
How they've shifted over time
Orobix entered H2020 in 2019 with contributions to large-scale scientific data processing for astronomy and particle physics (ESCAPE project, working with CERN and ESO data). By 2021, their focus had shifted decisively toward medical and computational biology applications — surgical robotics with AI-driven navigation (ATLAS) and GPU-accelerated cardiac simulation (MICROCARD). The common thread throughout is AI and GPU computing, but the application domain has moved from fundamental physics infrastructure toward healthcare and life sciences.
Orobix is moving toward AI-powered medical applications, particularly where GPU computing, 3D reconstruction, and autonomous systems intersect with healthcare — expect them to seek projects in surgical AI, medical imaging, or computational medicine.
How they like to work
Orobix has never coordinated an H2020 project, consistently joining as participant or third party — the role of a specialist contributor brought in for specific technical capabilities. Their 63 unique partners across 11 countries indicate they work in large consortia (typical of RIA and MSCA projects), contributing focused AI/computing expertise rather than leading the research agenda. This makes them a low-risk, high-value addition to any consortium needing applied AI or GPU computing skills.
Despite only 3 projects, Orobix has built a network of 63 partners across 11 countries, a consequence of participating in large European research infrastructure and training network consortia. Their network spans major European research nations but is rooted in Italy.
What sets them apart
Orobix bridges the gap between AI/computer vision research and real-world scientific and medical applications — a rare combination for an SME. While many AI companies focus on commercial products, Orobix has proven experience embedding its technology into demanding research environments from particle physics to surgical robotics. For consortium builders, they offer a compact, flexible team that can deliver GPU computing and AI components without the overhead of a large partner.
Highlights from their portfolio
- MICROCARDPushes GPU computing to exascale for cardiac simulation — positions Orobix at the frontier of computational medicine.
- ESCAPEMassive open science infrastructure project connecting Europe's top research facilities (CERN, ESO, SKA) — places Orobix in the EOSC ecosystem.
- ATLASAmbitious MSCA training network applying AI to autonomous surgical robotics across vascular surgery, colonoscopy, and ureteroscopy.