Milestone Transforms Urban Traffic in Europe With NVIDIA, Launches Project Hafnia in Genoa
Milestone Expands Project Hafnia to Europe: Partnering With NVIDIA and Nebius to Pioneer Responsible AI for Urban Infrastructure, Starting in Genoa After its initial rollout in the United States, Milestone’s groundbreaking Project Hafnia has made its European debut, beginning with a strategic deployment in Genoa, Italy. The project represents a significant step forward in the evolution of urban infrastructure, combining advanced AI technologies with ethically sourced and regulation-compliant video data to improve traffic management systems and contribute to the development of smarter, safer cities.
At the heart of this expansion is a robust collaboration with NVIDIA, a global leader in accelerated computing and AI, and Nebius, a European cloud provider. Together, the organizations are building a next-generation AI-powered data platform capable of transforming how cities operate, leveraging computer vision, synthetic data generation, and advanced modeling to make urban environments more efficient and responsive to citizens’ needs.
Driving AI-Powered Smart Cities With Project Hafnia
Project Hafnia is centered on a bold vision: to create AI systems that are not only powerful and effective but also transparent, fair, and fully compliant with European regulations. By integrating advanced AI solutions with city infrastructure, the initiative seeks to improve urban mobility, reduce congestion, and enhance public safety.
Milestone is using NVIDIA’s cutting-edge Maestro™ platform to manage and train large-scale visual language models (VLMs). These models learn the connections between text and visual information such as images and video footage, enabling them to deliver actionable insights, generate detailed summaries, and detect events or anomalies with a high degree of accuracy.
The development and training of these VLMs are being conducted on NVIDIA DGX Cloud using the NVIDIA NeMo Curator — a powerful tool for preparing high-quality datasets — ensuring that all inputs are traceable, diverse, and aligned with the EU’s strict data protection rules, including GDPR and the recently passed AI Act.
Genoa Becomes the Launchpad for European Expansion
The city of Genoa is the first European location to participate in Project Hafnia, providing an initial dataset centered on transportation and traffic systems. This collaboration will serve as the pilot for a much broader initiative that aims to deploy AI in urban centers across Europe. By focusing on real-world challenges such as traffic congestion, pedestrian safety, and public transport efficiency, Project Hafnia is laying the groundwork for meaningful innovation with tangible community impact.
Genoa’s participation reflects its ongoing commitment to digital transformation, particularly within public administration. “AI is achieving extraordinary results, unthinkable until recently, and the research in the area is in constant development,” said Andrea Sinisi, Information Systems Officer for the City of Genoa. “We enthusiastically joined forces with Project Hafnia to allow developers to access fundamental video data for training new Vision AI models. This data-driven approach is a key principle in the Three-Year Plan for Information Technology, aiming to promote digital transformation in Italy and particularly within the Italian Public Administration.”
NVIDIA Omniverse and Synthetic Data Capabilities
Milestone is among the first organizations to adopt the newly launched NVIDIA Omniverse Blueprint for Smart City AI — a reference architecture designed to help cities optimize their operations using AI agents and digital twins. This platform enables the creation of realistic, scalable simulations of urban environments, facilitating the development and testing of AI applications in a risk-free, virtual setting.
To enrich the training datasets further, Milestone is using NVIDIA Cosmos, a solution that generates synthetic video data based on real-world footage. This approach not only enhances the volume and variability of training data but also helps address privacy concerns by reducing reliance on personally identifiable information. Synthetic data generation plays a key role in refining VLMs while ensuring ethical data usage.
By combining real and synthetic datasets, Milestone is able to develop highly accurate and adaptable AI models that reflect the complexities of real-life urban systems. These models are then fine-tuned to run efficiently on NVIDIA GPUs, optimizing them for deployment in city-scale environments.
A European Visual Language Model for Transportation
One of the key outcomes of the Genoa initiative is the creation of a fully European Visual Language Model tailored for transportation management. Trained on a vast repository of responsibly sourced traffic data from the city, this VLM is the first of its kind to fully adhere to European regulatory and ethical standards.
This transportation VLM enables city administrators to derive meaningful insights from video surveillance and traffic monitoring systems. By automating the analysis of visual data, municipalities can identify traffic patterns, predict congestion, respond to accidents more effectively, and even fine-tune public transport scheduling.
The model is made available to other European cities through a controlled access license model, allowing them to benefit from high-quality AI tools without compromising on data ethics or privacy. This approach aligns with the European Union’s broader push toward digital sovereignty and responsible AI innovation.
Ensuring Compliance Through Sovereign Cloud Infrastructure
A key component of Project Hafnia’s responsible AI approach is its partnership with Nebius, which provides the cloud infrastructure required for model training and deployment. As a European-based cloud provider, Nebius guarantees that all data remains within the EU’s jurisdiction and is handled in accordance with its stringent privacy laws.
“Project Hafnia is exactly the kind of real-world, AI-at-scale challenge Nebius was built for,” said Roman Chernin, Chief Business Officer at Nebius. “Supporting AI development today requires infrastructure engineered for high-throughput, high-resilience workloads, with precise control over where data lives and how it’s handled. From our EU-based data centers to our deep integration with NVIDIA’s AI stack, we’ve built a platform that meets the highest standards for performance, privacy and transparency.”
By leveraging Nebius’ sovereign cloud environment, Milestone ensures that AI development remains fully anchored in Europe, reinforcing public trust and supporting the EU’s regulatory and ethical frameworks.
A Platform for Scalable, Ethical AI in Europe
Milestone’s CEO, Thomas Jensen, emphasized the broader implications of Project Hafnia’s European rollout. “I’m proud that with Project Hafnia we are introducing the world’s first platform to meet the EU’s regulatory standards, powered by NVIDIA technology. With Nebius as our European cloud provider, we can now enable compliant, high-quality video data for training vision AI models — fully anchored in Europe,” Jensen said. “This marks an important step forward in supporting the EU’s commitment to transparency, fairness, and regulatory oversight in AI and technology — the foundation for responsible AI innovation.”
The framework developed by Project Hafnia is not limited to transportation. Its architecture is designed to be extensible, supporting a wide range of use cases such as public safety, environmental monitoring, infrastructure planning, and emergency response. With the ability to scale across multiple domains and integrate data from diverse modalities, the platform offers long-term flexibility and value.
Scaling Smart City AI Across Europe
With its initial success in Genoa, Project Hafnia sets the stage for a new era in AI-enabled urban transformation across Europe. The project showcases how cross-sector collaboration — involving city governments, cloud providers, AI leaders, and regulatory experts — can foster responsible innovation at scale.
As Milestone continues to expand the platform to other cities, the goal remains clear: to empower European municipalities with the tools they need to build intelligent, responsive, and ethically governed urban environments. By combining cutting-edge technology with a strong commitment to data integrity and compliance, Project Hafnia offers a replicable model for AI-driven progress that other regions around the world may look to emulate.