Published Aug 13, 2026 | 5:43 PM ⚊ Updated Aug 13, 2026 | 5:54 PM
L&T has called the upcoming plant a big milestone in India's AI journey
Larsen & Toubro (L&T) is set to build India’s largest NVIDIA B300 AI factory after securing a Rs 10,000-15,000 crore order from US-based cloud platform Together AI, the company announced on Thursday.
The AI factory to be set up at L&T’s Vyoma.AI Chennai data centre campus in Tamil Nadu will have the capacity to host 10,000 NVIDIA B300 GPUs.
The NVIDIA B300 GPU (known as Blackwell Ultra) is the flagship data centre GPU released by the company in January 2026. It is said to have 1.5x more compute and memory capacity than the standard Blackwell chip.
L&T, whose AI subsidiary LTN Compute clinched the deal, called it “India’s largest single-cluster AI infrastructure” in its press release.
Together AI, with whom LTN Compute inked the deal, has five founders: Vipul Ved Prakash, Ce Zhang, Chris Ré, Tri Dao, and Percy Liang.
The 49-year-old New Delhi-born Vipul Ved Prakash had dropped out of St Stephen’s College “for want of undisturbed coding time”, an MIT review noted in a profile. He had hit the headlines when Apple acquired Topsy Labs, a social search and analytics firm he co-founded in 2008, for $200 million in 2013.
His current X bio soberly identifies him as Co-founder, CEO@togethercompute, but there was a time when Prakash called himself a “hacker, entrepreneur and desert rat” on the platform.
Commenting on the new deal, Vipul Ved Prakash said, “Making AI globally accessible is going to be the biggest infrastructure build-out in human history, and L&T understands that.”
L&T chairman and managing director SN Subrahmanyan said, “Artificial intelligence is becoming foundational to every industry and AI factories will power this transformation. Our deployment of an Nvidia B300 AI factory for Together AI marks a significant milestone in L&T’s gigawatt AI infrastructure mission.”
Together AI had recently signed a $240-million multi-year agreement with IBM to build a large-scale artificial intelligence cluster on IBM Cloud using NVIDIA chips, Reuters had reported.
The initial cluster with 2,000 NVIDIA B300 chips was to be located in the US, the report said.
The Chennai centre will have five times the GPU capacity of that IBM cluster.
From our archives:
‘Magnifica Humanitas’: What the Pope’s AI Encyclical means for India
The significance of President Trump’s AI security order for India and the rest of the world
