Ben Guerir, Morocco — November 5, 2025 — Akila has launched a campus-scale digital twin for Mohammed VI Polytechnic University (UM6P), establishing an AI infrastructure that spans 50 buildings across a 55-hectare site, home to more than 7,000 students. Built leveraging NVIDIA Blueprint for smart city AI, a reference architecture to build, train, and deploy city-scale AI. The project maps to the blueprint workflow showcasing a new benchmark for how simulation, real-time intelligence and semantic data can be fused at scale to enable smarter, more responsive environments.

This is the first time Akila’s full platform has been deployed at campus scale, transforming UM6P into a living laboratory for future smart-city systems. A deep technical integration was needed to make the solution possible including NVIDIA Omniverse libraries for synthetic data generation, vision AI ingestion from real-world sensors with NVIDIA Metropolis, the NVIDIA AI Blueprint for Video Search and Summarization (VSS) and NVIDIA Cosmos Reason to deploy video analytics AI agents for ongoing alerts and efficient operations.

The digital twin integrates diverse data sources — from building systems to utilities and mobility flows — within a structured, continuously updated Omniverse environment. It also introduces new layers of mobility and vehicle data, intelligently linked to the 3D model and natural language interfaces, giving campus operators unprecedented operational insight.

A distinctive capability of the platform is its use of a custom video language models (VLMs) trained with NVIDIA Cosmos Reason, enabling a contextually aware AI agent to deliver real-time video intelligence across the entire campus. Combined with Akila’s semantic graph, these models can analyze crowd movement, traffic flows and incident patterns, then feed the results back into simulations to support predictive and operational decision-making.

Stakeholders can also query the twin through LLM-powered interfaces, asking complex questions in natural language and receiving contextualized answers and automated reports. Whether it’s occupancy trends, mobility anomalies or energy deviations, campus managers can get precise, actionable intelligence on demand.

“What we’re building at UM6P is the AI infrastructure of the future — one that connects simulation, real-time intelligence and semantic data to create something far more powerful than the sum of its parts,” said Philippe Obry, Vice-President of Akila. “NVIDIA Blueprint for smart city AI and the incorporated technologies have accelerated our ability to integrate these layers at campus scale, turning the digital twin into a living, intelligent system.”

“With UM6P, we’re creating a uniquely data-rich and queryable digital environment,” said Pierre-Manuel Patry, Head of EMEA at Akila. “Any stakeholder can retrieve detailed, AI-generated insights about their area of focus, or generate automated reporting on demand. It’s a degree of control made possible by Akila’s unified AI infrastructure — and a perfect showcase of what a smart campus can be.”

By linking simulation, video intelligence, semantic structuring and natural language interfaces, Akila and UM6P are demonstrating how NVIDIA technologies can be operationalized to build next-generation AI infrastructure for the built environment — scalable from a single campus to entire smart cities.

Akila will be presenting this deployment live at the Smart City Expo World Congress 2025 in Barcelona from November 4-6, 2025. Akila can be found at the Mindware booth, and encourages all visitors to join us for a preview of our work in campus-scale digital twin.

Barcelona, 4 November 2025 — Akila, together with K2K and ThinkDeep, will showcase their shared vision for how semantically rich digital twins, video-language models (VLM) and retrieval-augmented AI systems can form an integrated intelligence chain that powers the future of Physical AI in cities and the built environment.

In line with NVIDIA Blueprint for smart city AI – a physical AI framework that combines simulation, training with VLMs, and deploying video analytics AI agents – the three companies are demonstrating how complementary technologies — spanning contextual intelligence, perception, and interaction, can be combined to create a new generation of city-scale AI systems.

• Akila provides the semantically rich digital twin platform, built on OpenUSD and integrating NVIDIA Omniverse libraries as its core environment for GPU-intensive simulation and spatial intelligence. Akila serves as the central layer where upstream data is contextualized in a unified operational model, enabling advanced analysis and scenario simulation at building and city scale.

• K2K deploys video analytics AI agents that generate structured, semantic narratives from city-wide visual data built with the NVIDIA Blueprint for video search and summarization (VSS) and powered by NVIDIA Cosmos Reason, an advanced reasoning VLM.
• ThinkDeep enriches the ecosystem with visual data ingestion Agentic AI and Graph RAG (retrieval-augmented generation) pipelines using NVIDIA Metropolis – transforming unstructured operational documents and records into structured knowledge that can be accessed interactively through AI assistants.

Together, these capabilities create a context-to-perception-to-interaction chain: from understanding and modeling the physical world through Akila’s digital twin, to extracting real-time intelligence from VLM, to making this knowledge accessible to diverse stakeholders through retrieval-augmented AI systems.

“This collaboration reflects a shared ambition to push the boundaries of how AI understands, simulates, and acts upon the physical world,” said Philippe Obry, Vice President at Akila. “By aligning semantically rich digital twins with advanced VLM and RAG technologies — and by leveraging NVIDIA Blueprint for smart city AI Omniverse for high-fidelity simulation — we’re laying the groundwork for a new generation of Physical AI capabilities in cities and infrastructure.”

Akila, K2K and ThinkDeep will jointly exhibit at the Smart City Expo World Congress 2025 in Barcelona, from 4–6 November, 2025.

Paris, June 11 2025 –– SNCF Gares&Connexions, France’s national railway operator, has joined forces with Akila, a leader in digital twin and AI platforms to deploy a groundbreaking real-time AI and simulation platform at Monte-Carlo train station in Monaco. This project is a major milestone for smart transportation infrastructure and smart cities, and was achieved by implementing the Akila digital twin platform while integrating NVIDIA Omniverse libraries and the newly announced NVIDIA Omniverse Blueprint for Smart City AI to deliver an application of NVIDIA 3-computer Physical AI architecture. 

Download the Press Release

Akila’s deployment at SNCF Monte-Carlo train station merges three powerful computing layers: simulation with digital twins, AI training, and deployment of AI agents — into a unified system that delivers real-time intelligence and automation to the built environment. By accelerating Akila’s digital twin platform with NVIDIA Omniverse technologies and NVIDIA Omniverse Blueprint for Smart City AI, SNCF is revolutionizing how critical infrastructure is monitored, managed, and integrated with powerful AI agent-driven urban transformation.

“This is not just a digital upgrade—it’s a leap into the future of infrastructure,” said Fabrice Morenon, Managing Director, SNCF Gares&Connexions. “With Akila and NVIDIA, we’re unlocking real-time understanding of our stations to boost efficiency, lower emissions, and ultimately serve the public better. This is a major milestone for SNCF group”

A Digital Twin Revolution in Rail and Transport Hubs

Akila’s platform ingests and consolidates diverse data sources—such as crowd movement, solar heating, airflow, and IoT sensors—to simulate building operations at high fidelity. Through NVIDIA Omniverse technologies, SNCF Gares&Connexions can visualize and interact with 3D digital twins enabling:

• Emergency simulation to improve safety
• Energy optimization that has already delivered a 20% reduction in energy usage
• Faster responses to operational issues, with a 50% cut in intervention times
• Preventive maintenance to achieve 50% downtime reduction and 100% on-time completion during deployment
• €30,000 in annual energy savings per site
• Enhanced training for autonomous systems
• New use cases in safety, security, and crowd control

This platform also provides a secure, sovereign digital environment thanks to Akila and NVIDIA’s accelerated computing both on-premise and in the cloud, ensuring compliance with SNCF Gares & Connexions’ data governance and cybersecurity standards.

Transforming Real Estate Operations in Real Time

Akila’s real-time AI and simulation platform is helping SNCF Gares&Connexions shift from static building management to dynamic, predictive infrastructure operations. SNCF’s service providers, like French computer vision specialist XXII, are now fully integrated into the Akila platform—offering advanced monitoring and data insight in a user-friendly interface.

Real-Time AI, From Edge to Cloud

This integrated platform lays a foundation closely aligned with the NVIDIA Omniverse Blueprint for Smart City AI, a reference framework for building, testing, and optimizing AI agents in SimReady digital twins:

• Simulate: Build a SimReady digital twin of specific locations and facilities with the Omniverse blueprint. Remote operators and partners can visualize and interact with infrastructure across locations in a digital twin. Powered by high-speed GPU, this computing enables deep simulations, operational forecasting, and predictive analytics.
• Train AI models: Currently, Akila uses SNCF’s historical and real-time data to train its AI Assistant agent. It plans to augment its training data with synthetic data generated from NVIDIA Cosmos.
• Deploy AI Agents: With real-time sensor fusion and computer vision technology from NVIDIA Metropolis, AI agents can generate and act on insights derived from cross-system analytics while maintaining a safe and healthy environment for occupants. This includes understanding station energy, thermal loads, crowd movement, airflow, and IoT sensors.

From Monaco to the World

This initiative is more than a local success — it is a model for how cities and transport networks can embrace physical AI to build resilience, sustainability, and efficiency.

Akila and its partners are already adapting this model to airports, ports, and logistics hubs across Europe, the United States, and the Middle East — where the appetite for smart infrastructure is growing rapidly.

“The experience and insights gained through this project are already proving invaluable,” said Philippe Obry, Vice President of Akila. “We are now working with cities and asset owners to scale this platform across global portfolios — enabling a real-time digital nervous system for the built environment.”