Linker Vision has unveiled Linker AI Nexa, a Physical AI platform designed to help cities and enterprises translate real-time environmental data into coordinated operational decisions. Announced at Smart City Expo World Congress 2026, the platform combines agentic AI, digital twins, NVIDIA accelerated computing, and government standard operating procedures (SOPs) to support applications spanning public safety, disaster response, urban planning, and climate adaptation. Rather than limiting AI systems to monitoring and analyzing camera feeds or sensor information, Nexa aims to connect real-world perception with predictive simulation, decision-making, and automated workflows. The launch highlights an emerging direction in enterprise AI: moving intelligent agents beyond digital applications and into the management of physical environments.
Linker Vision Brings Physical AI and Agentic Workflows to Smart Cities
Linker Vision is expanding its Physical AI strategy with Linker AI Nexa, a platform designed to connect real-time city infrastructure data with AI-driven decision-making and operational processes.
The announcement reflects a shift in how cities and infrastructure operators are approaching artificial intelligence. Conventional smart city systems typically use cameras, Internet of Things (IoT) sensors, and analytics software to identify traffic incidents, environmental changes, and potential safety risks. However, converting those insights into coordinated responses across government departments remains a more complicated task.
Linker AI Nexa is designed to address that gap by connecting physical-world data, specialized AI agents, digital twin environments, and established government procedures. Its objective is to move from observing conditions to recommending and coordinating actions.
The platform builds on Linker Vision’s existing AI capabilities and incorporates NVIDIA technologies for accelerated computing, visual understanding, simulation, and reasoning.
Four-Stage Architecture Connects Perception With Action
Linker AI Nexa organizes its operational framework around four stages: Sense, Predict, Decide, and Act.
The Sense stage processes information from cameras and IoT sensors to interpret real-world conditions. The Predict stage uses environmental models and digital twin simulations to evaluate possible outcomes. During Decide, AI agents connect those findings with organizational knowledge and government SOPs to identify responsible teams and recommend responses. Finally, Act coordinates workflows and triggers actions through connected systems.
This architecture is intended to reduce the gap between identifying an incident and initiating an appropriate response.
For example, during a potential flooding event, a system could interpret camera and sensor information, assess environmental conditions, identify relevant emergency procedures, and recommend actions to the appropriate municipal departments.
The approach also introduces an important governance consideration. AI-generated recommendations and actions need to remain aligned with established operational authority, security requirements, and human oversight. Linker Vision says Nexa incorporates security policies and operational controls to govern AI-driven activity.
NVIDIA Technologies Form the Platform’s Technical Foundation
Nexa brings together several NVIDIA technologies, each serving a different part of the Physical AI workflow.
NVIDIA Metropolis’ Video Search and Summarization (VSS) blueprint helps process and interpret large volumes of camera footage. NVIDIA Cosmos provides world foundation models designed to help AI systems understand physical environments and real-world conditions.
For spatial computing, Linker Vision uses NVIDIA Omniverse libraries and OpenUSD to combine city infrastructure, vegetation, and environmental information into digital twins. The company uses 3D Gaussian Splatting (3DGS) to construct detailed three-dimensional representations from visual data.
NVIDIA Nemotron reasoning models provide another layer, allowing planners and operators to explore scenarios and adjust design parameters through natural-language interactions.
The combination connects visual intelligence, spatial simulation, and AI reasoning. Rather than treating video analytics and urban planning as separate software functions, Nexa aims to connect them through a shared digital representation of the physical environment.
This architecture also aligns with NVIDIA’s broader Physical AI strategy, which combines simulation, synthetic data generation, AI model development, and deployment across real-world environments.
MVRDV Collaboration Brings AI Digital Twins Into Urban Planning
Alongside the platform launch, Linker Vision announced a collaboration with Dutch architecture and urban design firm MVRDV. The partnership will apply AI-powered digital twins to urban planning and climate adaptation, with potential projects across Poland, the Netherlands, Ireland, and Berlin.
One focus is urban forestry, where planners need to understand how tree placement affects shade, thermal comfort, and the wider built environment.
Linker Vision says its ESG agent combines scientific evaluation criteria from MVRDV with digital twin simulations to identify urban heat hotspots, evaluate planting opportunities, and model long-term environmental outcomes.
In Taipei pilot applications, the companies report that 3DGS-based scanning increased tree survey throughput from approximately 100 to more than 1,000 trees per person per day. Simulations also projected an increase in shade coverage from 15.7% to more than 50% over 15 years.
The company says the platform can simulate those 15-year outcomes in approximately 10 minutes and support multiple design iterations within an hour, compared with around two weeks for an iteration using the previous process.
These figures are company-reported pilot results and projections, rather than independently verified outcomes from completed citywide deployments. Their practical value will depend on local environmental data, model assumptions, planting conditions, and implementation.
The collaboration is expected to expand beyond urban forests into residential communities, public spaces, urban regeneration, and rooftop utilization.
From Smart City Analytics to Operational AI
Linker Vision’s existing work with the Kaohsiung City Government in Taiwan provides an operational foundation for the new platform. Its Lighthouse Project has focused on using AI to monitor traffic incidents and road hazards, while the Nexa announcement identifies flood and typhoon response as additional application areas.
The distinction is significant. Detecting a flooded road is one problem; coordinating transportation agencies, emergency responders, and municipal teams around that information is another.
Nexa’s proposed architecture attempts to connect both stages through shared operational context and predefined procedures. Its modular design also follows a reusable deployment model, allowing cities to adapt the underlying digital twin and AI framework to different local datasets and governance requirements.
The next challenge is proving that this approach can operate reliably across different municipal systems, maintain accurate environmental representations, and coordinate responses without bypassing human accountability.
For Linker Vision, the launch represents a move from AI-powered city monitoring toward agentic systems that help manage physical infrastructure. For the wider industry, it highlights how AI agents, digital twins, and accelerated computing are beginning to converge around real-world operational problems.
Market Landscape
Physical AI is emerging as an important direction in enterprise AI, extending machine learning beyond digital applications into infrastructure, robotics, industrial environments, and smart city operations.
Linker Vision’s announcement reflects five market developments:
- Physical AI: AI systems are increasingly being designed to understand and interact with real-world environments.
- Agentic AI: Autonomous agents are moving from information analysis toward coordinated operational workflows.
- Digital twins: Cities are using 3D representations to simulate infrastructure changes, environmental conditions, and planning scenarios.
- AI-powered urban governance: Municipalities are exploring ways to connect fragmented camera, sensor, and departmental systems.
- Climate intelligence: AI simulations are helping planners evaluate urban heat, tree coverage, shade, and sustainability initiatives.
The competitive ecosystem includes NVIDIA, Microsoft, Google, and specialist digital twin and geospatial technology providers. NVIDIA’s Physical AI ecosystem is particularly relevant because it connects accelerated computing, simulation, foundation models, and video intelligence.
Linker Vision’s differentiation is its attempt to connect these capabilities directly with municipal SOPs and operational decision-making rather than treating AI analytics as a standalone function.
Top Insights
- Linker AI Nexa combines Physical AI, agentic AI, NVIDIA technologies, and digital twins to coordinate real-world city operations.
- Its Sense, Predict, Decide, and Act framework connects camera and sensor information with government procedures and operational workflows.
- NVIDIA Cosmos, Metropolis, Omniverse, OpenUSD, and Nemotron provide the foundation for video understanding, simulation, spatial intelligence, and reasoning.
- Linker Vision and MVRDV are applying AI digital twins to urban forestry, climate adaptation, shade optimization, and sustainable city planning.
- Taipei pilot results reported by Linker Vision include 10-times-plus tree survey throughput and projected shade coverage exceeding 50% within 15 years.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
