NxtGen launches M, India's first sovereign, enterprise-grade agentic AI inferencing platform built on NVIDIA Nemotron, enabling production-scale AI reasoning and autonomous decision-making within India's sovereign cloud.

The infrastructure layer features India’s dedicated AI Factory powered by NVIDIA Blackwell GPUs, Spectrum-X networking, and BlueField-3 DPUs, designed for large-scale model training, high-performance inference, and AI-as-a-Service inside the sovereign cloud.

The interview highlights the importance of training the trainers for AI education, the evolving role of coding with AI assistance, and balancing leadership, learning, and life in the field.

M acts as an intelligent inferencing layer that connects AI models with enterprise systems (ERP, CRM, HRMS, ITSM, document repositories) and supports both global and India-developed models with local-language and region-specific capabilities.

Safety and governance are addressed by quarantining models, restricting internet access, disabling memory, and preventing data leakage, with domain-specific experts helping to limit risk and misuse.

The sovereign design centers on data residency, security, and compliance, incorporating persistent memory, auditability, and governance guardrails suitable for regulated industries and large deployments.

Sustainability efforts include achieving net-zero status, exploring future energy with Small Modular Reactors, and prioritizing liquid cooling for data centers due to higher power densities, while noting talent gaps in this technology in India.

Open-source is central to M, enabling cost reduction and rapid iteration, with developers able to contribute via GitHub and IDE integrations, backed by active community engagement.

M emphasizes practical utility over generic chat, including domain-specific models such as an Indian legal model and a chartered-accountant level financial assistant to reduce hallucinations and boost accuracy.

M for Travel demonstrates a vertical application that translates natural-language intents into executable travel journeys using live inventory through ITQ’s Travelport APIs for planning, booking, and post-trip services.

The first use case, M for Travel, enables natural-language queries to generate fully planned and executable travel journeys with real-time flight and hotel data, moving from planning to booking within a single system.

M for Travel showcases vertical capability by turning intents into structured travel plans with live inventory via Travelport APIs, supporting planning, booking, and post-trip services at scale.