Meta to lease Reliance-built AI data center in Gujarat as India compute demand grows

Meta Platforms has agreed to lease an AI-enabled data center from Reliance Industries in Gujarat, according to a report citing the terms of the deal. The facility is described as 168 megawatts in size, positioning it as part of Meta’s effort to lock in computing capacity for training and running AI systems across its family of apps.

The move is notable because it ties Meta’s next phase of AI infrastructure directly to India’s expanding data center supply. Meta is already investing in AI models and AI-powered features across Facebook, Instagram and WhatsApp, and the report frames the Reliance arrangement as a step toward meeting rapidly rising compute requirements.

Reliance, for its part, is increasingly positioning itself as a key builder and partner for large technology customers that want locally hosted AI computing. While the specific contract terms beyond the lease and the facility size were not detailed in the report, the partnership is presented as a first-of-its-kind data center relationship between Meta and Reliance in India.

Secondary coverage also characterizes the project as a strategic capacity expansion with room for further build-out, reflecting a broader industry pattern in which hyperscalers and major AI buyers seek flexible, power-focused infrastructure. However, those accounts are not primary disclosures from Meta or Reliance, and they should be treated as unverified until confirmed by an official statement or filing.

The broader context is that AI workloads are forcing a shift in where compute is sourced. Data centers have become a central bottleneck because they require large quantities of electricity, high-speed networking and specialized equipment. India is emerging as a destination for new supply, supported by rising demand and expanding grid capacity, even as global markets remain constrained.

For Meta specifically, the timing matters because AI model development and deployment tend to translate into a steady stream of infrastructure demand. Data centers also sit at the intersection of AI and regulatory scrutiny, as governments and industry groups increasingly focus on energy use, grid impacts, and data localization rules. Leasing capacity through a local partner can be faster than building from scratch, though it can also shift long-term cost and operational risk to the customer.

Still, important details were not disclosed in the materials surfaced for this report. The company did not specify in the cited coverage when the 168 MW facility is expected to go live, how much of it will be dedicated to Meta versus shared workloads, or the duration and pricing structure of the lease. It also did not provide clarity on whether the data center will support a particular generation of AI hardware or any specific model training or inference commitments.

What to watch next is confirmation from Meta and Reliance through official communications, such as investor materials, project announcements, or regulatory filings. Industry watchers will likely focus on (1) project timeline and phasing, (2) any expansion options tied to the initial 168 MW, and (3) whether Meta indicates similar capacity deals in other Indian states as it scales AI infrastructure.