Delhivery and Nvidia partner to build AI maps for India’s complex geography
Leveraging billions of shipments, the initiative uses Nvidia’s Nemotron and CV-CUDA models
New tools will resolve address disambiguation and interpret unstructured, landmark-based descriptions
Logistics company Delhivery on Thursday announced that it is advancing the development of digital mapping solutions designed specifically for India’s complex geography and unique addressing systems, in collaboration with Nvidia AI.
The announcement was made at the India AI Summit. Delhivery plans to use Nvidia’s accelerated computing, CV-CUDA and Nemotron open models to create highly accurate, scalable, and locally relevant maps. The goal is to support India’s growing digital economy by using Delhivery’s vast proprietary dataset, built from billions of shipments over several years.
The company is evaluating several key capabilities. These include address disambiguation to distinguish between similar-sounding addresses and locations, contextual inference to interpret unstructured address descriptions, and predictive sequencing to optimise last-mile delivery routes in dense and unplanned urban areas.
Kapil Bharati, Co-founder and Chief Technology Officer of Delhivery, said India needs mapping solutions that can operate at scale and understand the realities of its streets. He noted that global mapping tools often struggle with Indian address formats, and that combining Delhivery’s shipment data with Nvidia’s accelerated computing could help build a powerful location intelligence platform to support India’s digital growth.
India’s geographic diversity, rapid urbanisation, and unique addressing systems make mapping especially challenging. Many addresses in the country are informal and rely on landmarks, local references, and non-standard formats instead of precise coordinates. Through this collaboration, Delhivery aims to use Nvidia’s open AI models to address these challenges.
Vishal Dhupar, Managing Director for Asia South at Nvidia, said Nvidia’s technologies help companies turn complex data into useful insights. He added that Delhivery’s work shows how Nvidia’s AI infrastructure and Nemotron open models can help solve large-scale geospatial challenges and enable new AI-powered applications and services.
Delhivery said its India-focused mapping platform aims to become a core digital infrastructure layer, supporting applications that require accurate, intelligent, and up-to-date location data.




























