L&T partners US chipmaking giant NVIDIA to build India’s largest gigawatt-scale AI factory

Gigawatt-scale AI factory data centre infrastructure developed by L&T and NVIDIA in India
High-density AI server racks representing gigawatt-scale AI factory infrastructure. Representative image

Construction and engineering major Larsen & Toubro (L&T) has announced a proposed venture with US chipmaking giant NVIDIA to build sovereign, scalable gigawatt-scale AI factory infrastructure aimed at reinforcing India’s position as a global AI powerhouse.

The announcement was made at the India AI Summit on 18 February 2026, with the initiative aligned to the IndiaAI Mission and designed to support production-grade AI capacity anchored in the country’s digital and industrial transformation.

The partnership integrates L&T’s engineering and infrastructure execution capabilities with NVIDIA’s AI infrastructure, including GPUs, CPUs, networking, accelerated storage platforms, enterprise software stack, and reference architectures to enable rapid and secure AI adoption.

As part of the venture, AI-ready datacenter infrastructure and advanced computing platforms will be deployed to support large-scale AI workloads across priority sectors.

Read also : PM Modi to inaugurate India AI Impact Summit 2026 at Bharat Mandapam on February 19

In alignment with the IndiaAI Mission, the initiative is intended to create sovereign AI infrastructure that allows critical data, models, and workloads to be built, trained, and deployed within India while remaining interoperable with global ecosystems.

The framework is expected to serve domestic requirements as well as global hyperscalers, cloud providers, and enterprises seeking large-scale AI capacity from India as a strategic hub.

The venture plans to develop a gigawatt-scale AI data center factory providing AI-ready capacity for high-density, next-generation workloads to help customers expand in India efficiently and sustainably.

L&T will scale NVIDIA GPU cluster deployment at its Chennai data centre up to 30 MW capacity within a 300-acre gigawatt-scalable campus, while a new 40 MW datacenter is currently under execution in Mumbai.

“This AI factory model will be able to deliver advanced AI services to global off takers, hyperscalers & India Inc, including manufacturing, infrastructure, energy, financial services, healthcare, and public services thereby moving from experimentation to production-scale deployment,” L&T said in a statement.

It is designed to provide standardized, enterprise-grade AI capabilities, enabling predictable performance, security, and time-to value for industrial and services use cases, the company added.

The infrastructure will also allow L&T and its group companies to deploy live AI agents on a sovereign cloud, fostering a self-sustaining innovation hub through frameworks such as LTTS’s Lights-Out Factory, LTM’s Blueverse, LTFS’s agentic AI deployments, and internally developed AI agents.

Commenting on the development, S. N. Subrahmanyan, Chairman & Managing Director of L&T, said, “India’s enterprises are ready to move from AI pilots to production-scale deployment. The investment establishes the foundation – secure, scalable, and sovereign infrastructure – required to power manufacturing, energy, financial services, healthcare, and public services.”

Jensen Huang, Founder and CEO of NVIDIA, added, “AI is driving the largest infrastructure buildout in human history – everyone will use it, every company will be powered by it, and every country will build it… Together with L&T… we are laying the foundation for world-class AI infrastructure that will power India’s growth and help realize the full vision of India AI.”

“The venture is enabling AI factories at national scale – sovereign by design, optimized for accelerated computing, and ready to serve global and domestic AI demand, thereby advancing India’s AI Mission and its vision of “Making AI in India and Making AI Work for India,” the company said.

Read also : India to expand national AI infrastructure with addition of 20,000 GPUs to existing 38,000 base