Berlin – October 19, 2026 -- Linker Vision launched an ESG AI agent that, in a pilot, projected urban shade coverage rising from 15.7% to more than 50% within 15 years through targeted tree planting.
Agent simulates 15 years of planting scenarios in minutes
The agent combines environmental predictions with AI reasoning to assess urban heat and shade coverage. It identifies heat hotspots, selects suitable tree species and simulates the results. A planner asks where new trees would best reduce street-level heat. The agent then tests planting options and models shade and temperature over 15 years in minutes. Planners review the recommended plan before an urban heat hotspot project moves forward.
MVRDV supplied scientific data and evaluation criteria
Dutch architecture and urban design firm MVRDV contributed scientific data, shaped the criteria the agent uses to assess planning options and tested the agent during development. The two companies are collaborating on a digital twin that supports predictions of urban heat island effects and shade coverage. Urban forest planning is available today.
NVIDIA stack underpins the digital twin
The digital twin is built with 3D Gaussian Splatting on NVIDIA accelerated computing. NVIDIA Omniverse libraries and OpenUSD unify city, tree and environmental data into a single scene. NVIDIA Nemotron reasoning models let users run planting scenarios and optimize shade through natural language. The NVIDIA Metropolis Blueprint for video search and summarization and the Cosmos open world foundation model help interpret video and real-world conditions.
Shared twin targets additional government applications
Linker Vision designed the city digital twin as a common foundation for multiple planning needs, including climate adaptation and disaster resilience. Government agencies could build further applications on it, such as analyzing pedestrian flows and preparing venues for major sporting events.
Linker Vision founder and CEO Paul Shieh said the aim is to help cities explore options and make better choices by connecting real-world understanding with simulation and AI reasoning.