Larsen & Toubro’s newly formed subsidiary, Vyoma.AI, just secured an order worth Rs 10,000-15,000 crore from Together AI. The deal, which will establish India’s largest Nvidia B300 AI facility in Chennai, marks L&T’s decisive step into building core computational infrastructure for artificial intelligence. It follows an earlier, equally strategic move by the conglomerate to manufacture rare earth magnets in India, aligning with government objectives for critical materials.
The Chennai facility, first reported by Startuptalky, will house up to 10,000 Nvidia B300 GPUs. These units are designed for large-scale AI training, fine-tuning, and inference workloads. This kind of compute capacity is fundamental for any nation’s AI ambitions, providing the raw horsepower needed for developing sophisticated models and deploying them at scale. The order from Together AI gives Vyoma.AI a foundational project as it consolidates L&T’s data centre, cloud, and AI infrastructure businesses, as Inc42 reported.
From Construction to Compute
L&T’s traditional strength lies in large-scale engineering and construction projects, from power plants to metro systems. The formation of Vyoma.AI and the immediate securing of a multi-billion-dollar AI factory contract indicate a strategic pivot. It shows L&T is applying its expertise in managing complex, capital-intensive infrastructure projects to the digital realm. This isn’t just about building data centers; it is about constructing specialized facilities optimized for the specific demands of AI workloads, which require immense power, cooling, and network capabilities distinct from general-purpose computing. The Nvidia B300 GPUs, central to this facility, represent the cutting edge in AI accelerators. Deploying 10,000 of them in a single location means creating a compute cluster capable of training foundation models or running large-scale inference for diverse applications.
Securing Critical Materials
This push into AI infrastructure is not L&T’s only recent strategic play for national self-reliance. On August 12, 2026, Startuptalky reported that L&T aims to manufacture rare earth magnets in India. The government’s goal is to minimize dependence on China, the world’s largest producer of these critical materials. L&T is likely to seek an INR 72.8 billion incentive package for this initiative, and may partner with a technology specialist.
Rare earth magnets are indispensable components in high-tech industries, including electric vehicles, wind turbines, and advanced electronics. China’s near-monopoly on their production and processing has long been a point of strategic concern for many countries, including India. By stepping into this manufacturing space, L&T is addressing a fundamental supply chain vulnerability. This move requires substantial investment and specialized metallurgical expertise, akin to the technical and logistical challenges of building an AI factory.
A Coherent Strategy: Foundational Infrastructure for a Digital India
Taken together, L&T’s moves into constructing India’s largest Nvidia B300 AI facility and planning a rare earth magnet plant reveal a coherent, long-term strategy. Both ventures are about building foundational, strategic infrastructure. They are not merely commercial projects, but initiatives that directly support India’s broader ambitions for technological independence and industrial capability.
The common thread is L&T’s role as a builder of core national assets, whether physical or digital. For the AI factory, it is about providing the essential compute power that will drive innovation across sectors, from deep tech research to enterprise applications. For rare earth magnets, it is about securing the material inputs necessary for advanced manufacturing and clean energy technologies. Both require substantial upfront capital, deep engineering know-how, and a long-term vision that aligns with national strategic mandates. L&T, a company long associated with tangible, large-scale engineering, is now clearly positioning itself as a key architect of India’s future technological underpinnings, from the very components that power AI to the compute clusters that run it. This signals a broadening of its mandate, focusing on areas critical for a modern, technologically self-sufficient economy.
Compiled by Launch91 Desk from the sources linked above. More about Launch91.
