Meta Bets on Jamnagar: Scaling AI Infrastructure Through Reliance Partnership
Table of Contents
The Compute Land Grab Moves East
The global race for AI supremacy is no longer just about who has the best model, but who has the most efficient way to power it. Meta is now moving its infrastructure bets into India, announcing a strategic partnership with Reliance Industries to establish a 168-megawatt (MW) AI-enabled data center in Jamnagar, Gujarat. This marks Meta’s first dedicated AI infrastructure footprint in the country, signaling a transition from mere software deployment to deep-tier hardware integration.
The deal is more than a simple lease agreement; it is a tactical pivot. While Meta’s 2020 investment of $5.7 billion into Jio Platforms was primarily about capturing the Indian consumer market via WhatsApp and Facebook, this new venture focuses on the ‘industrial’ side of AI. By securing a high-capacity facility in Jamnagar, Meta is effectively diversifying its global compute grid, reducing its reliance on North American and European hubs while tapping into India’s growing role as a global backend for AI training and inference.
Engineering the Jamnagar Facility
The technical specifications of the Jamnagar site highlight the immense energy and cooling demands of modern Large Language Models (LLMs). AI workloads generate significantly more heat and consume more power than traditional cloud computing. To combat this, the facility will utilize desalinated seawater for cooling—a necessity in the arid climate of Gujarat—and will be powered by renewable energy. Meta has reportedly committed to covering the full cost of the water and energy required to sustain its operations.
The facility is expected to be operational within two years. Reliance will act as the end-to-end service provider, handling everything from the physical construction and connectivity to the renewable power sourcing. This positions Reliance not just as a telecom giant, but as a comprehensive infrastructure partner for the ‘compute-hungry’ era of Big Tech.
The Geopolitical Compute Shift
Meta is not alone in this migration. The rush toward Indian soil reflects a broader trend where Microsoft, Google, and Amazon are aggressively expanding their local data center footprints. The incentive is twofold: domestic demand and government policy. New Delhi has introduced aggressive tax exemptions through 2047 for foreign cloud providers, provided the workloads are processed within Indian borders.
The scale of this shift is evident in the numbers. India’s total installed data center capacity grew from 375MW in 2020 to roughly 1.5GW in 2025. Industry projections suggest this will surge to over 8GW by 2030. When paired with Blackstone-backed AirTrunk’s recent $30 billion pledge to build 5GW of capacity, it’s clear that the ‘AI Gold Rush’ has a significant physical footprint in South Asia.
Sourcing Green Energy for Black Boxes
Beyond the Jamnagar site, Meta is aggressively securing its energy supply chain to meet sustainability targets. The company has signed contracts for nearly 1GW of new renewable energy capacity via CleanMax and Fourth Partner Energy. These agreements ensure that as Meta scales its AI operations in India, it doesn’t collide with the country’s existing power grid constraints.
While the financial specifics of the Reliance deal remain undisclosed, the strategic implication is clear: Meta is building a hedge. By integrating its AI requirements into Reliance’s ecosystem, Meta gains a streamlined pathway to deploy next-gen AI tools in one of the world’s most populous digital markets, all while securing the raw compute power necessary to keep its models competitive against the likes of OpenAI and Google.
