Larsen & Toubro has secured a mega order worth ₹10,000–15,000 crore to build a large-scale AI factory in Chennai for U.S.-based Together AI. The project will deploy 10,000 NVIDIA B300 GPUs, marking L&T’s entry into specialized AI infrastructure. Following the news, the company's stock rose by nearly 2%.
Larsen & Toubro (L&T) has officially entered the high-tech AI infrastructure space by winning a mega order valued between ₹10,000 crore and ₹15,000 crore. The order, placed by U.S.-based cloud platform Together AI, involves building a massive AI factory at L&T’s data center campus in Chennai. This facility will be executed by LTN Compute, a subsidiary under L&T’s Vyoma.AI division.
The project is designed to be a high-performance computing hub, deploying 10,000 NVIDIA B300 graphics processing units (GPUs). These specialized chips are the primary hardware used for training large-scale artificial intelligence models. The campus is planned for a 250-megawatt (MW) capacity in its first phase, supported by a 150 MVA power infrastructure setup. For L&T, this order represents a shift from traditional civil and power infrastructure projects toward the growing demand for specialized digital infrastructure.
By leveraging its experience in large-scale engineering and data center construction, the company is attempting to capture the rising need for high-end computing facilities in India. This move allows L&T to diversify its revenue streams beyond its standard engineering, procurement, and construction (EPC) business. Following the announcement, L&T shares saw positive movement, trading up by approximately 1.2% to 1.9% during the session as investors reacted to the large order size.
While the project is a milestone, investors may want to monitor the specific risks involved in this new business line. Building and running AI factories is a complex, capital-intensive task that requires reliable, high-capacity power and stable connectivity. Unlike traditional infrastructure projects, these facilities face the risk of rapid technological changes, where new GPU generations can quickly make existing hardware less efficient. The company’s success will also depend on its ability to manage technical execution and power requirements at such a massive scale.
The next important developments for investors will be the project's construction timeline and its operational commissioning. Shareholders will likely track how quickly L&T can scale this new AI division and whether it can secure additional orders of this magnitude to build a consistent business in the high-performance computing sector.