AMD Silo AI and UniMoRe Launch a Technology Alliance to Advance Physical AI

AImagelab at the University of Modena and Reggio Emilia (UniMoRe), one of Europe's leading computer vision and multimodal learning groups, will collaborate with AMD Silo AI, to advance multimodal VLA systems for robotics and autonomous driving on AMD Instinct™ GPUs starting in 2026.

AMD Silo AI and the University of Modena and Reggio Emilia are establishing a broad, multi-year technology alliance spanning robotics, autonomous driving, and multimodal AI with the ambition to extend into additional domains over time. The alliance's first chapter centers on co-developing open-source AMD ROCm™ software-native Vision Language Action (VLA) models and world models. 

Centered at UniMoRe’s AImagelab and aligned with European initiatives such as MINERVA EuroHPC and the EU AI Factory for Science, the collaboration will establish the World Models Foundry for collaboration with additional ecosystem partners from academia and industry forerunners. It aims to deliver an open source VLA stack, open-source demos and toolkits optimized for AMD Instinct™ GPUs and cultivate next-generation talent through a three-year PhD program starting in 2026.

Integrating Research, Innovation, and Active Industry Collaboration

The effort aligns with the AMD Silo AI contribution in Physical AI and Robotics and builds on ongoing European collaborations in high-performance computing and sovereign AI.

The collaboration has direct industry relevance and product insight through collaboration with Generative Bionics, a leading Italian robotics company that designs and deploys full-stack humanoid systems for real-world applications. Generative Bionics maintains a tight engineering collaboration with UniMoRe and AMD, bringing deployment experience that bridges research into working robotic platforms.

“This collaboration accelerates and broadens our work in providing robotics and physical AI developers with a great and highly performant experience on AMD Instinct GPUs. The co-design work with AImagelab is proof of our deep commitment to open collaboration. By bringing together optimized model implementations, data pipelines, and evaluations into a single World Models Foundry, we’re creating a clear path from ideas to full-scale digital twin capabilities. Additionally, we are ensuring strong industry relevance by inviting technology forerunners to the task force, with Generative Bionics leading the way as a founding member,” highlights Niko Vuokko, Sr. Director of Software Development at AMD.

Collaboration Scope and Thrusts

UniMoRe and AMD Silo AI will co-design architectures, training recipes, and evaluation pipelines targeting AMD MI‑series GPUs and future ROCm™ platforms. 

Prof. Rita Cucchiara, Rector of UniMoRe and founder and head of AImagelab, underscored the strategic importance of the initiative, “The quest to build machines that truly understand and act in the physical world is one of the defining scientific challenges of our time. World Models and Vision Language Action systems are at the frontier of this endeavor. With AMD Silo AI, we are combining state-of-the-art research with scalable, open engineering to push that frontier forward while nurturing a new generation of European AI talent at UniMoRe and across our partner network.” 

The research work will emphasize three core thrusts:

1. Multimodal world models that integrate vision, language, and action for closed-loop control in both robotics and autonomous driving.

2. Efficient training and inference on ROCm, including memory and compute-optimized kernels and high-throughput data pipelines.

3. Robotics and autonomous driving-centric evaluation using both simulation and real data, with a strong focus on multimodal perception and human-environment interaction.

The collaboration is designed to produce state-of-the-art research contributions at top-tier venues alongside open-source artifacts, ensuring that academic impact and engineering deployment advance in lockstep.

About AImagelab and UniMoRe’s European Leadership in AI Research

AImagelab at the University of Modena and Reggio Emilia is recognized as a leading European center for computer vision and multimodal AI, with a strong publication record across CVPR, ICCV, and ECCV. Under the leadership of Prof. Rita Cucchiara, founder of AImagelab, Director of the Modena ELLIS Unit, and Rector of UniMoRe since late 2025, the lab anchors UniMoRe within the European Laboratory for Learning and Intelligent Systems network. Its research spans vision-language and Vision Language Action (VLA), robotics, and autonomous driving, combining foundational methods with applied, human-centered AI. UniMoRe and AImagelab are active in major European initiatives, including EuroHPC efforts such as MINERVA and the EU AI Factory programs (e.g., ITAL4IA), and collaborate broadly across academia, industry, and cultural institutions. With a vibrant community of ELLIS scholars and PhD candidates, AImagelab continues to drive European AI excellence through open science, high-impact research, and cross-disciplinary innovation.

AMD CAUTIONARY STATEMENT   

This blog contains forward-looking statements concerning Advanced Micro Devices, Inc. (AMD) such as the features, functionality, performance, availability, timing and expected benefits of AMD products; and the collaboration between AMD Silo AI and the University of Modena and Reggio Emilia (UniMoRE) and the expected benefits thereof, which are made pursuant to the Safe Harbor provisions of the Private Securities Litigation Reform Act of 1995. Forward-looking statements are commonly identified by words such as "would," "may," "expects," "believes," "plans," "intends," "projects" and other terms with similar meaning. Investors are cautioned that the forward-looking statements in this presentation are based on current beliefs, assumptions and expectations, speak only as of the date of this presentation and involve risks and uncertainties that could cause actual results to differ materially from current expectations. Such statements are subject to certain known and unknown risks and uncertainties, many of which are difficult to predict and generally beyond AMD's control, that could cause actual results and other future events to differ materially from those expressed in, or implied or projected by, the forward-looking information and statements. Investors are urged to review in detail the risks and uncertainties in AMD’s Securities and Exchange Commission filings, including but not limited to AMD’s most recent reports on Forms 10-K and 10-Q.  

AMD does not assume, and hereby disclaims, any obligation to update forward-looking statements made in this presentation, except as may be required by law.