techcrunch.com via Reddit

Meta Lands First India AI Data Center With Reliance

meta ai infrastructure ai-infrastructure data-centers india hyperscalers

Key insights

  • Meta's 168 MW Jamnagar facility will run on renewables with desalinated seawater cooling, and Meta covers all energy and water costs.
  • Meta separately secured nearly 1 GW of new renewable capacity in India through CleanMax and Fourth Partner Energy.
  • India's data center capacity grew from roughly 375 MW in 2020 to 1.5 GW in 2025, projected to exceed 8 GW by decade's end.

Why this matters

Meta's decision to build owned capacity in India rather than procure third-party compute signals that hyperscalers now treat India as a tier-one AI infrastructure market, not an overflow region. The renewable-energy-plus-desalinated-cooling model, with Meta absorbing all operating costs, establishes a template for how Western AI companies can satisfy India's sustainability expectations while retaining direct capacity control. With India's data center footprint projected to grow more than fivefold by decade's end and Microsoft, Amazon, Google, OpenAI, Adani, and Tata Consultancy Services all building simultaneously, this deal marks the opening of a multi-year infrastructure land grab that will determine which AI stacks dominate the world's most populous country.

Summary

Meta is committing to India's AI infrastructure at scale through a 168-megawatt data center in Jamnagar, Gujarat, built in partnership with Reliance Industries. The facility runs on renewable energy, uses desalinated seawater for cooling, and Meta absorbs all energy and water costs. A separate agreement adds nearly 1 gigawatt of new renewable capacity through CleanMax and Fourth Partner Energy. Essentially: (Meta, Reliance Industries) are extending a relationship built on Meta's multibillion-dollar Jio Platforms investment. - Jamnagar facility is expandable, expected online within two years. - India's data center capacity grew from roughly 375 MW in 2020 to around 1.5 GW in 2025, projected above 8 GW by decade's end. - Microsoft, Amazon, Google, OpenAI, Adani, and Tata Consultancy Services are all building simultaneously. India has graduated from emerging market to primary AI infrastructure target.

Potential risks and opportunities

Risks

  • Reliance Industries faces construction and grid-connection delays in Jamnagar that could slip the two-year completion window, leaving Meta's India AI roadmap behind Microsoft, Google, and Amazon.
  • Meta's commitment to cover all energy and water costs creates direct exposure to cost overruns if renewable procurement through CleanMax or Fourth Partner Energy encounters pricing or permitting setbacks.
  • Concentrating Meta's entire Indian AI infrastructure in a single Reliance-operated facility creates vendor dependency; regulatory or geopolitical friction between the Indian government and Meta could disrupt operations with no alternative site in place.

Opportunities

  • CleanMax and Fourth Partner Energy gain a marquee hyperscaler anchor for nearly 1 GW of renewable projects, strengthening their capital-raising position and ability to attract additional enterprise anchor tenants.
  • Adani and Tata Consultancy Services, already expanding data center capacity in India, can use the Meta-Reliance deal as a benchmark to attract similar anchor-tenant agreements from other hyperscalers entering the market.
  • Indian AI companies and cloud resellers gain a new source of domestic GPU capacity within two years, reducing dependence on US-based hyperscaler regions and potentially lowering inference latency for India-facing applications.

What we don't know yet

  • Total contract value and lease terms between Meta and Reliance have not been disclosed in public reporting.
  • Whether the two-year completion timeline covers the full 168 MW capacity or only an initial phase is not specified in the announcement.
  • How Indian data sovereignty regulations will apply to Meta's AI workloads running on Reliance-operated infrastructure has not been addressed publicly.