Key Highlights
- Joint venture between Larsen & Toubro and NVIDIA to create AI‑centric data facilities totaling gigawatt‑level power capacity.
- Initial sites include a 30 MW campus in Chennai spanning 300 acres and a 40 MW data centre under construction in Mumbai.
- The infrastructure will host high‑density GPU clusters, accelerated storage, and the NVIDIA AI Enterprise software suite for enterprise‑grade workloads.
- Project aligns with IndiaAI's vision of sovereign, secure, and globally interoperable AI computing resources.
Detailed Insights
The partnership capitalises on L&T’s proven capability to deliver large‑scale engineering projects and NVIDIA’s leadership in GPU‑driven compute. By deploying massive GPU farms, the duo intends to provide on‑premises AI training and inference capacity that rivals leading global cloud providers, while ensuring data remains within Indian jurisdiction. The Chennai campus is designed as a 300‑acre, expandable zone that can be scaled to gigawatt output as demand grows. The Mumbai facility, slated for completion later this year, adds another 40 MW of dedicated AI compute power. Together, these sites will support everything from large language model training to real‑time inferencing for sectors such as manufacturing, healthcare, and finance.
Key Concepts
- AI Factory: A purpose‑built data environment optimised for accelerated computing, featuring dense GPU arrays, high‑speed networking, and AI‑specific storage.
- Gigawatt‑Scale Infrastructure: Power capacity on the order of one billion watts, enabling simultaneous operation of thousands of high‑performance accelerators.
- Sovereign AI Computing: Computing resources that reside under national regulatory control, guaranteeing data residency and security.