Meta Platforms signed its first AI data-centre deal in India on June 10, 2026, leasing 168 megawatts of capacity at a new Reliance Industries facility in Jamnagar, Gujarat. The partnership, announced simultaneously by both companies, is the latest step in a relationship that began with Meta's $5.7 billion investment in Jio Platforms in 2020 and deepened with a $100 million enterprise-AI joint venture in August 2025. The Jamnagar facility will be powered entirely by renewable energy and cooled with desalinated seawater — a design choice that, alongside Meta's separate 1-gigawatt renewable PPA with CleanMax and Fourth Partner Energy, makes the deal one of the most sustainability-front-loaded hyperscaler build-outs in India to date.
The Deal: What Was Announced
| Capacity | 168 MW AI-enabled data centre, leased to Meta |
| Location | Jamnagar, Gujarat (Reliance complex) |
| Powered by | 100% renewable energy, desalinated seawater cooling |
| Readiness | ~2 years (per Reliance) |
| Additional renewable PPA | ~1 GW with CleanMax and Fourth Partner Energy |
| Predecessor deals | Meta $5.7B in Jio Platforms (2020); $100M JV (Aug 2025) |
| Scope of services (Reliance) | Design, construction, power, connectivity, operations |
Why Jamnagar
Jamnagar is the site of Reliance's flagship oil refinery complex — the world's largest single-site refinery. The conglomerate has spent two decades building out the surrounding ecosystem: captive power plants, port facilities, water desalination, fibre connectivity, and large industrial land holdings. A new greenfield AI data centre sits naturally on that footprint. Reliance's offer to Meta is end-to-end: site, shell, power, cooling, connectivity, and ongoing operations. In effect, Reliance is positioning itself as a one-stop shop for hyperscaler build-outs in India — competing directly with specialist colocation operators like CtrlS, Yotta, and STT GDC.
Why Meta Chose India Now
The Jamnagar lease is the first materialisation of Meta's India infrastructure strategy. The company's 2026 capital expenditure guidance — raised in late April to as much as $145 billion — includes capacity for international expansion. India is the natural next stop for two reasons. (1) It is Meta's largest single user base by country. WhatsApp alone serves more than 500 million Indian users. (2) Indian AI demand is rising sharply: enterprise AI inference, Indic-language model hosting, and the IndiaAI Mission's sovereign-compute programme are all pulling hyperscalers to add capacity locally. Meta's existing arrangement with Reliance — via the Jio investment and the enterprise-AI JV — gave it a partner with the land, the power, and the balance sheet to deliver 168MW in 24 months. The deal also dovetails with the US-India trade and technology dialogue, which has framed digital infrastructure as a strategic priority.
What This Means for India
Three things matter for Indian readers beyond the headline. (1) The ecosystem play: The Jamnagar facility is the anchor for a broader Reliance industrial-AI build-out. Reliance has talked openly about becoming a hyperscale-grade infrastructure provider — this deal is the proof point. (2) Renewable and water constraints: The seawater-cooling design is a Gujarat-specific solution, but the renewable PPA is portable. If the Jamnagar model scales, it becomes a template for hyperscaler build-outs in other coastal Indian states (Andhra Pradesh, Tamil Nadu, Karnataka). (3) Competition for IndiaAI compute: The IndiaAI Mission is funding sovereign compute capacity for Indian startups and researchers — and the same pool of skilled AI engineers, GPUs, and renewable power will be contested by private hyperscalers. The risk is that Indian startups get priced out of their own market. The opportunity is that the same global capacity accelerates Indian-language AI products to global scale.
Sources: CNBC, TechCrunch, about.fb.com (Meta press release), Liquide, Business Standard
See also: PM Modi Launches Bharat Innovates 2026 in France — 120 India · US Inflation Hits 4.2% in May 2026 — Highest in Three Years.
Sources
- Bloomberg — bloomberg.com
- Reuters Business — reuters.com/business
- Voxlogue editorial research
