Larsen & Toubro Lands $1.57 Billion AI Data Centre Contract in India

Infrastructure Giant Enters AI Factory Business

Larsen & Toubro announced Thursday it has secured a contract valued between $1.05 billion and $1.57 billion to build and operate an AI data centre for U.S.-based cloud platform Together AI. The deal represents the Mumbai-headquartered engineering conglomerate's entry into the AI infrastructure sector and marks the largest deployment of advanced AI computing hardware in India to date.

The contract, awarded to L&T's newly established AI infrastructure subsidiary LTN Compute, will see the company deploy 10,000 Nvidia B300 chips at its Vyoma.AI facility located on the company's campus in Chennai. According to L&T, the centre will support Together AI's cloud platform for AI inference, fine-tuning, and training workloads.

Shares of Larsen & Toubro climbed 1.2 percent in midday trading following the announcement, reflecting investor confidence in the company's diversification into high-margin technology infrastructure.

Nvidia's Most Powerful Chips Head to South India

The Nvidia B300 chip ranks among the most powerful processors currently available for AI inference tasks, the process by which trained AI models generate outputs from new data. The deployment of 10,000 units positions L&T's Chennai facility as a significant node in the global AI compute infrastructure network.

Together AI, the California-based client, operates a cloud platform that provides developers and enterprises with access to large language models and other AI capabilities. The partnership gives Together AI dedicated compute capacity in South Asia, a region where demand for AI infrastructure is accelerating but supply remains constrained.

For L&T, the contract represents a strategic pivot toward digital infrastructure at a time when traditional engineering and construction markets face margin pressure. The company has been exploring opportunities in data centres, cloud infrastructure, and telecommunications as part of a broader technology services expansion.

India's AI Infrastructure Race Heats Up

The contract arrives as India intensifies efforts to build domestic AI capabilities amid global competition for compute resources. Several Indian conglomerates, including Reliance Industries and the Adani Group, have announced plans for large-scale data centre investments over the past year, though few have secured commitments of this scale with international cloud providers.

India's data centre capacity is projected to grow significantly through 2027, driven by both domestic digital transformation initiatives and multinational companies seeking alternatives to concentrated infrastructure in the United States and China. Chennai, Hyderabad, and Mumbai have emerged as key hubs, benefiting from relatively stable power supply, technical talent pools, and supportive state-level policies.

The deployment timeline and commercial terms of the L&T contract were not disclosed. However, the facility's integration with Together AI's platform is expected to support enterprise customers across sectors including financial services, healthcare, and software development, all of which are scaling AI adoption.

L&T's move also underscores a broader shift in the company's portfolio. Long dominant in heavy engineering, power transmission, and infrastructure construction, the conglomerate has been investing in technology verticals including cybersecurity, semiconductor manufacturing support, and now AI compute infrastructure. The AI factory model, which bundles hardware, cooling, power management, and software orchestration, offers higher returns than traditional construction projects but requires deeper technical partnerships and operational expertise.

The Chennai facility will need to address challenges common to high-density compute environments, including power consumption, thermal management, and network latency. Nvidia's B300 chips, while powerful, generate substantial heat and require sophisticated cooling systems to maintain performance and reliability.

As AI workloads continue to grow globally, contracts of this scale are likely to become more common in Asia. India's combination of cost advantages, technical workforce, and improving digital infrastructure positions it as a contender in the race to host the next generation of AI factories.