Meta Leases Rather Than Builds Its First India AI Data Center From Reliance

A lease with options to scale, not an owned facility

The central structural fact of this announcement is that Meta is not building a data center in India. Reliance is building it, and Meta is leasing the capacity. The company states plainly that "Reliance will build a data center with 168 MW capacity, which Meta will lease, with options to scale."

That arrangement puts the capital burden of construction on Reliance while giving Meta a call option on future expansion. The first phase delivers 168 MW; the "option to scale" language is repeated twice, signaling that this initial commitment is a foothold on a much larger campus rather than a fixed endpoint.

Reliance describes it as "India's first built-to-suit AI data centre for a global technology leader" — meaning the facility is being designed to Meta's specifications but owned and operated by the local partner. Meta gets tailored infrastructure without carrying the asset.

Meta pays for the energy and water, and the cooling is unusual

One specific and often-overlooked detail: "Meta will cover the full cost of the energy and water supporting the facility." Even in a lease model, Meta is absorbing the operating inputs directly rather than folding them into a bundled rate paid to Reliance.

The Jamnagar site is described as "cooled with desalinated seawater" — a design choice reflecting Gujarat's coastal location and the water intensity of AI compute. That Meta explicitly commits to funding both energy and water suggests the water supply is a material cost line, not an afterthought.

Jamnagar was chosen for a reason the source names directly: "access to the significant energy resources needed to power advanced AI-enabled infrastructure." Reliance's existing industrial complex there is what makes the location viable for high-density compute.

Nearly 1 GW of renewables contracted separately from the lease

Alongside the lease, Meta contracted almost 1 GW of clean energy through two providers, and these agreements are explicitly separate from Reliance's own renewable sourcing for the site. CleanMax accounts for 837 MW of new solar and wind in Rajasthan and Karnataka, bringing cumulative capacity with that provider above 900 MW. Fourth Partner Energy adds 88 MW across Tamil Nadu, Karnataka, Maharashtra, and Uttar Pradesh.

Meta frames this as addressing its "value chain emissions in the region" and consistent with its global goal of matching operations with 100% clean and renewable energy. The scale of contracted renewables — roughly six times the first-phase capacity of the data center itself — is consistent with the "options to scale" language and with matching energy against future growth, not just the initial 168 MW.

The partnership moves from Meta's platforms into Reliance's physical plant

This deal reframes a relationship that until now ran through software and capital. In 2020 Meta made a $5.7 billion investment in Jio Platforms; the two later formed a joint venture bringing Meta's open-source AI models to Indian enterprises and developers. Those were bets on Reliance's reach and Meta's models.

Zuckerberg's own framing ties the facility to product ambition: "This world-class facility in Jamnagar will help us scale our AI infrastructure globally while deepening our long-term investment in India's economy."

The specific implication is a division of labor: Reliance owns the campus, the land, and the energy access; Meta rents the compute and pays the running costs. For applied AI teams watching how hyperscalers enter new markets, this is a template where the platform company avoids owning foreign real estate while still placing capacity — and, via Project Waterworth's subsea connectivity — the network path close to a fast-growing user base. Meta gets local compute without becoming a local infrastructure operator.