AImagelab at the University of Modena and Reggio Emilia (UniMoRe) has announced a multi-year technology alliance with AMD Silo AI aimed at advancing multimodal Vision Language Action (VLA) systems for robotics and autonomous driving. This collaboration, which kicks off in 2026, will leverage AMD Instinct™ GPUs to enhance the capabilities of these systems and is poised to extend into additional domains over time.
The alliance will initially focus on developing open-source AMD ROCm™ software-native VLA models and world models. Located at UniMoRe’s AImagelab and aligned with European initiatives such as MINERVA EuroHPC and the EU AI Factory for Science, the partnership aims to establish a World Models Foundry. This will serve as a collaborative hub for engaging further ecosystem partners from both academia and industry, delivering an open-source VLA stack and demos optimized for AMD hardware.
UniMoRe’s involvement underscores its prominent role in European AI research, particularly in computer vision and multimodal learning. The collaboration also has an educational dimension, as it plans to launch a three-year PhD program in 2026 to nurture the next generation of AI talent.
The partnership’s significance lies in its commitment to integrating research with industry collaboration. By teaming up with Generative Bionics, an Italian company specializing in humanoid robotics, the collaboration aims to bridge theoretical research with practical deployment. Generative Bionics maintains a close engineering relationship with both UniMoRe and AMD, contributing valuable insights into real-world applications of their research.
Niko Vuokko, Senior Director of Software Development at AMD, emphasized the collaborative nature of this effort. “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,” he stated. “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.” He also highlighted the importance of having industry leaders like Generative Bionics involved in the task force.
The collaboration will focus on three core research thrusts. First, it will develop multimodal world models that integrate vision, language, and action for closed-loop control in both robotics and autonomous driving. Second, it will optimize training and inference processes on ROCm, including memory and compute-optimized kernels and high-throughput data pipelines. Lastly, there will be a strong emphasis on evaluation methods tailored to robotics and autonomous driving, leveraging both simulation and real data to enhance multimodal perception and human-environment interaction.
As the collaboration unfolds, it aims to produce impactful research contributions that will be presented at top-tier venues while simultaneously developing open-source artifacts. This dual focus ensures that academic advancements and engineering deployment progress in tandem.
AImagelab is recognized as a leading European center for computer vision and multimodal AI, with a strong publication record across major conferences such as CVPR, ICCV, and ECCV. The lab, under the leadership of Prof. Rita Cucchiara, who is also the Rector of UniMoRe, plays a crucial role in anchoring the university within the European Laboratory for Learning and Intelligent Systems network. Its research encompasses various fields including vision-language and VLA systems, robotics, and autonomous driving, combining foundational methodologies with applied, human-centered AI approaches.
AMD’s collaboration with UniMoRe aligns with its commitment to fostering innovation in Physical AI and Robotics. The partnership is part of broader European initiatives focused on high-performance computing and sovereign AI, reflecting a concerted effort to advance research and development in these rapidly evolving fields. As the alliance progresses, it is expected to catalyze advancements not only in robotics and autonomous driving but also in other sectors that may benefit from enhanced multimodal AI capabilities.
This partnership represents a forward-looking effort to address significant technical challenges in AI, aiming to develop machines that can effectively understand and interact with the physical world. As the collaboration moves into its implementation phase, the outcomes are anticipated to have a lasting impact on the landscape of robotics and AI in Europe.
See also
Sam Altman Praises ChatGPT for Improved Em Dash Handling
AI Country Song Fails to Top Billboard Chart Amid Viral Buzz
GPT-5.1 and Claude 4.5 Sonnet Personality Showdown: A Comprehensive Test
Rethink Your Presentations with OnlyOffice: A Free PowerPoint Alternative
OpenAI Enhances ChatGPT with Em-Dash Personalization Feature
