Voxelmaps and NVIDIA partner to give city teams real-time 3D visibility into streets, infrastructure, and urban change – with natural-language querying built in.
SAN JOSE, CA / ACCESS Newswire / March 17, 2026 / Voxelmaps today announced the deployment of a real-time city digital twin for the City of San José, developed in collaboration with NVIDIA. The initiative combines continuous street-level LiDAR and imagery capture through the Voxelmaps Reality Capture Network with NVIDIA Jetson platform and NVIDIA accelerated computing to create a living digital model of the city that updates continuously as urban environments change.
The data captured through the Reality Capture Network is processed through Voxelmaps City Insights, its 4D real-time city digital twin platform, enabling municipal teams to work from an always-current view of streets, infrastructure, and urban change over time.
The deployment demonstrates how AI-powered digital twins can help cities better understand and manage complex urban environments. By combining continuous sensing, high-resolution spatial data, and accelerated AI computing, the platform creates a living model of the city that can support infrastructure monitoring, urban planning, and new forms of human-AI interaction with the built environment.
“At city scale, the challenge is not creating a 3D model once – it’s keeping it live and current,” said Peter Atalla, CEO of Voxelmaps. “Voxelmaps Reality Capture Network, using our SYMBO Duo Realtime 3D Mapping Sensors, combined with NVIDIA Jetson at the edge and NVIDIA-accelerated computing across the workflow, enables a city digital twin that stays measurable, refreshable, and practical for day-to-day city operations.”
Multimodal City Intelligence at Scale
The digital twin is built from continuously captured street-level data collected through the Voxelmaps Reality Capture Network, which uses SYMBO Duo Realtime 3D Mapping Sensors to capture LiDAR and high-resolution imagery.
Within Voxelmaps City Insights, this data is fused into a continuously updated urban model that combines:
-
LiDAR point clouds for precise geometry and measurement of infrastructure and urban assets
-
High-resolution imagery to enable feature detection and semantic understanding
-
3D Gaussian Splatting (3DGS) workflows to generate photorealistic visualization of city environments
Unlike traditional mesh-based modeling approaches, which are difficult to maintain at city scale, 3D Gaussian Splatting enables continuously refreshed visual representations that better support incremental updates as new data is captured.
Natural-Language Interaction with City Data
Voxelmaps has integrated NVIDIA Cosmos Reason 2, an open, customizable reasoning vision-language model (VLM), enabling natural-language interaction with visual and temporal city data within Voxelmaps City Insights.
Authorized city users can query the digital twin using everyday language to support operational awareness, planning, and infrastructure management.
Example queries include:
-
“Show me where roadworks have changed curb access since last month.”
-
“Identify corridors with low streetlighting coverage relative to comparable neighborhoods.”
-
“List blocks with repeated pavement defects and estimate the affected surface area.”
Looking forward, Voxelmaps is working through integration with latest version of NVIDIA Metropolis Blueprint for video search and summarization (VSS) to enable the deployment of video analytics AI agents.
Supporting City Operations and Planning
Voxelmaps’ work with NVIDIA and the City of San José emphasizes measurable outputs, traceability to source data, and clear governance over how insights are generated and used.
The platform is designed to support a range of municipal workflows, including:
-
Improved visibility into street-level conditions and urban change
-
Proactive infrastructure maintenance and asset monitoring
-
Broader access to spatial intelligence beyond specialist GIS teams
-
Data-driven planning and decision-making to improve residential services
San José – home to many leading AI and technology companies – provides a real-world testbed for next-generation AI-powered city infrastructure and digital twin technologies.
Scaling the Reality Capture Network
Voxelmaps plans to expand the Voxelmaps Reality Capture Network, using its SYMBO Duo Realtime 3D Mapping Sensors, and deploy Voxelmaps City Insights across 100 U.S. cities over the next two years, supporting a new generation of real-time digital twins designed for urban operations.
About Voxelmaps
Voxelmaps digitalizes the physical world for Physical AI, enabling the creation of real-time, living models of cities and complex environments. Through the Voxelmaps Reality Capture Network and Voxelmaps City Insights, the company produces high-fidelity, time-aware (4D) digital representations that support continuous measurement, monitoring, and analysis of the physical world.
Voxelmaps’ data and platforms help power Large World Models (LWMs) and the next generation of robots, autonomous systems, and AI technologies that need to perceive, navigate, and interact with real-world environments.
Media Contact (Voxelmaps)
Peter Atalla
CEO, Voxelmaps
+44 7970 717141
peter@voxelmaps.com
Forward-Looking Statements
This press release contains forward-looking statements, including statements regarding planned deployments, anticipated capabilities, and potential impacts. These statements are based on current expectations and are subject to risks and uncertainties that could cause actual results to differ materially.
SOURCE: Voxelmaps Inc.
View the original press release on ACCESS Newswire
