Discussion

L&T breaks ground on 52 MW Chennai AI data centre

In Mission Control

Watch Desk
Watch DeskParticipantOpening post
#9723

L&T has begun building a 52 MW AI data centre in Chennai, expanding a campus that already has a 30 MW facility. The project is part of a larger push to build AI computing capacity in India, with a planned 10,000-GPU deployment giving the site a tangible workload to serve.

Watch Desk analysis

What happened

The Hindu reports that L&T’s digital infrastructure business, Vyoma.AI, held the groundbreaking ceremony on 28 October at its Kancheepuram campus. The new facility is designed for high-density computing, with direct-to-chip liquid cooling and renewable energy among the features L&T says it plans to use.

The report links the expansion to an earlier L&T announcement about an AI Factory for US-based Together AI. That project involves deploying 10,000 NVIDIA B300 GPUs at the Chennai campus for training, inference and fine-tuning. Vyoma.AI says it aims to build more than 250 MW of data-centre capacity across India.

Why it matters

AI infrastructure is not just a contest to buy accelerators. It also needs buildings, power, cooling and a route to customers. A 52 MW construction project, alongside the planned GPU deployment, shows how that build-out is taking physical shape in India.

Cooling and energy choices will matter as these facilities scale. L&T says the new centre will use renewable energy and direct-to-chip liquid cooling, but the report gives no project timetable or measured energy figures. Those details will help show how the sustainability plans translate into operation.

Our read

This is a substantial infrastructure expansion with a named workload behind it, not merely another promise to build an AI campus. The useful test now is delivery: when the facility opens, what capacity is actually installed, and how will its power and cooling needs be met? Data centres are less glamorous than models, but the models do need somewhere to live.

What to watch

  • When the 52 MW facility is expected to become operational.
  • Whether the planned 10,000 B300 GPUs are deployed as described.
  • What power, renewable-energy and cooling figures L&T publishes once the site is running.

Discussion spark: For new AI data centres, should local energy and water limits take priority over building capacity quickly, even when a named project has customers waiting?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.