Nvidia Forges Major AI Infrastructure Deals in India, Secures Multi-Generational Chip Supply with Meta

Nvidia recently detailed substantial investment strategies and strategic alliances in India during the India AI Impact Summit in New Delhi, signaling an aggressive pursuit of regional artificial intelligence infrastructure dominance. This initiative aligns with a global trend toward establishing localized, high-capacity AI computing centers to address the escalating demand for advanced machine learning capabilities across the subcontinent.

A significant domestic development was announced by Larsen & Toubro (L&T), which confirmed the formation of a joint venture with Nvidia to construct a gigawatt-scale AI factory. This facility, located in Chennai, is set to integrate Nvidia processors to deliver up to 30 megawatts (MW) of computing power, with plans for a subsequent 40 MW facility in Mumbai. This undertaking aims to foster local technological self-reliance by reducing dependence on imported AI hardware solutions.

Further cementing the regional build-out, Yotta Data Services committed capital exceeding $2 billion to deploy over 20,700 Nvidia Blackwell Ultra processors. This procurement is targeted to establish one of Asia's most formidable AI superclusters, with a completion deadline set for August 2026. Separately, Yotta secured a four-year agreement with Nvidia, valued at over $1 billion, to establish one of the Asia-Pacific region's largest DGX Cloud clusters for immediate access to cloud-based AI resources.

The Adani Group declared a long-term investment strategy totaling $100 billion, to be deployed by 2035, focusing on developing AI-ready data centers powered entirely by renewable energy. This plan includes campus developments in collaboration with Google in Visakhapatnam and separate projects with Microsoft in Hyderabad and Pune. Microsoft also reinforced its global AI expansion, pledging $50 billion by the end of the decade to scale AI technology adoption across emerging economies in the Global South.

In a move with implications for the global AI supply chain, Nvidia finalized a multiyear, multigenerational partnership agreement with Meta Platforms. This comprehensive deal mandates the supply of millions of current and next-generation AI chips, explicitly including the Blackwell architecture and the forthcoming Rubin GPUs, ensuring Meta's sustained technological advantage. The agreement also covers the large-scale integration of Nvidia's Grace and forthcoming Vera CPUs, alongside Meta's incorporation of Nvidia's Spectrum-X Ethernet networking technology to optimize hyperscale data center operations for AI training and inference tasks.