NPCI embeds sovereign AI into India’s real-time payments infrastructure with Nvidia

Focuses on building a scalable AI layer for population-scale payment systems.

The National Payments Corporation of India (NPCI) has announced a collaboration with Nvidia to build and scale a sovereign, payments-native AI foundation model built for India’s digital payments ecosystem.

As part of this engagement, NPCI will use Nvidia Nemotron - a family of open models with open weights, training data and recipes - in its model development journey.

NPCI states that the collaboration focuses on embedding AI into population-scale, real-time payment systems while maintaining security, trust and resilience.

The immediate focus is on strengthening existing AI deployments.

The organisation has already deployed AI in grievance redressal through the UPI Help Assistant. The assistant is powered by FiMI (Financial Model for India), a fine-tuned small language model developed for the payment ecosystem.

It supports grievance resolution for UPI users by enabling timely and consistent responses at scale.

As part of the next phase of its AI journey, NPCI aims to evolve from use-case–specific agents to a foundational, scalable AI layer for the payment ecosystem.

The proposed model will explore architectures such as Mixture of Experts (MoE) to support high-volume, low-latency payment environments, while gradually expanding capabilities across multilingual datasets and agent-optimised systems.

NPCI’s chief technology officer, Vishal Kanvaty, said, through this collaboration with Nvidia, NPCI aims to advance AI capabilities designed specifically for India’s payments ecosystem.

Drawing from the experience of operating population-scale, real-time payment systems, this initiative is designed to create a sovereign, payments-native AI foundation that strengthens trust, resilience, and security, while remaining aligned with India’s regulatory and data sovereignty requirements, said Kanvaty.

“As we evolve from use-case–led AI deployments to a foundational and scalable AI layer, our focus remains on enabling the broader ecosystem to innovate responsibly through robust governance frameworks and secure, future-ready infrastructure,” Kanvaty added.

Over time, the initiative intends to provide a platform that can be leveraged by banks, fintechs, and other participants in the payment ecosystem, while maintaining a strong focus on data security, sovereignty, and responsible use of AI.

The development indicates that AI deployment is moving beyond standalone enterprise use cases and into regulated, high-volume real-time payments infrastructure.

This could translate into opportunities around AI infrastructure integration, domain-specific model tuning for BFSI, compliance-led AI architecture and secure deployment services.

Such integration will require sustained investment in accelerated computing, model optimisation and secure data handling frameworks that align with regulatory requirements.