MUMBAI: As rivals race into billion-dollar artificial intelligence (AI) infrastructure bets, Larsen & Toubro Ltd is taking a slower, engineering-led approach, wagering that disciplined scaling will win in a rapidly expanding market.
India’s largest engineering conglomerate expects its data centre business to generate $1 billion, or approximately ₹9,200 crore, in annual revenue by 2030, by which time it aims to build 350 megawatts (MW) of operational capacity. The strategy contrasts with peers announcing gigawatt-scale investments, even as demand for AI infrastructure accelerates across the country.
“We of course hope that in the next five years, we can scale our business up to $1 billion in annual revenue. That’s the goal, and given our growth targets coupled with how fast the industry is growing, I don’t think this should be improbable,” said Seema Ambastha, chief executive of Larsen & Toubro Vyoma, the company’s data centre division.
India’s largest engineering and construction group reported ₹2.55 trillion in annual operating revenue last fiscal.
Measured buildout
Currently, L&T operates a 30MW data centre at its Chennai campus and is expanding capacity in phases. “We’re currently building additional capacity of 30MW in Chennai, as well as a further 40MW in Mumbai. By end-2027, we’ll therefore have 100MW in data centre capacity built and operational,” she added.
In November, the company said it plans to spend nearly $2.5 billion over five years to set up 300MW in data centre capacity across five facilities, with potential to expand more.
The data centre targets that Ambastha shared with Mint show L&T increasing capacity to 180MW by 2028, followed by 250MW by end-2029 and 350MW by 2030. The push comes after L&T, on 18 February, announced a partnership with Nvidia to build gigawatt-scale data centres in the country.
The phased expansion stands in contrast to the scale of investments being announced across the sector. In October last year, Tata Consultancy Services announced plans to spend $6.5 billion to build a 1 gigawatt (GW) data centre by 2030. Tata’s data centre division, branded as HyperVault, later received $1 billion in funding from US-based asset manager TPG in November, and on 17 February signed a partnership with the world’s largest private AI firm OpenAI for data centre usage.
On 14 October, Google announced a $15-billion, 1GW data centre investment in Visakhapatnam, with domestic conglomerate Adani Group as partner. A month later, Reliance Industries’ joint venture with asset manager Brookfield announced a similar 1GW data centre, also in Vizag, at a net investment outlay of $11 billion. Meta Platforms, through a 30% stake in its step-down subsidiary Reliance Enterprise Intelligence Ltd, is a partner in Reliance’s data centre venture.
L&T, however, says its integrated engineering, energy and technology capabilities allow it to scale capacity in phases rather than commit upfront to giga-scale projects. Despite competitors announcing significantly larger investments, Ambastha said the company does not see itself as falling behind.
L&T’s strengths across engineering, procurement and construction, power transmission and distribution, and green energy allow it to rely on in-house capabilities for building and operating data centres, she said. “This enables our data centres to use in-house capabilities from within the group.”
“Alongside, we have access to two technology and engineering services integrators—L&T Technology Services (LTTS) and LTM (formerly LTIMindtree). We have good quality capital that helps us proactively scale our own data centre capacity at the right time, aligned with market opportunity and strategic priorities.”
Much of the expansion, she said, will depend on how AI demand evolves in India.
“As per current projections, India will see 3.5-5GW in data centre capacity by 2030. By 2032, this has been projected to rise to 8GW, driven by demand for digital transformation and technology services. If AI grows the way initial projections were made, this capacity will straight-up double in volume, and India will quite likely see up to 16GW in net data centre capacity in the next six years,” Ambastha said.
“Our goal is to capture 10% of the market. If this kind of demand growth does indeed happen, we’ll be there to capture it.”
AI land rush
Overall demand for data centres in India is rising steeply. On 20 February, at the India AI Impact Summit event in New Delhi, union IT minister Ashwini Vaishnaw said the country had already received $250 billion in AI infrastructure investment commitments.
Between October and February, companies including Reliance Industries, Adani Enterprises, TCS, Amazon, Google and Microsoft announced plans to commit nearly $259 billion over five to seven years.
Demand has been driven by increasing consumption of 5G services, rising use of AI applications such as ChatGPT and favourable state and central government policies encouraging investment in digital infrastructure.
Executives across the industry broadly share expectations of sustained AI-driven growth. At brokerage firm Kotak’s annual investor conference, Jugeshinder Singh, group chief financial officer of Adani Group, said that “AI’s short-term impact is often overestimated, but its long-term impact is significantly underestimated.”
“He highlighted the trajectory of data centre capacity addition in the US and suggested the large expansion potential in India, well into the 2030s. In his view, India is likely to reach, and potentially surpass, US-level capacity over time. This will require a step-change in national energy infrastructure,” the brokerage firm cited Singh as saying at the event.
Bharti Enterprises has also outlined expansion plans. On 6 February, in a post-earnings conference call, executive vice-chairman Gopal Vittal said the company aims to expand its Nxtra data centre business to 1GW in capacity within three years, though he did not disclose planned investment.