Meta & Reliance Plan Massive AI Data Centre

India has been described as the world's next great AI infrastructure market for years. In June 2026 that stopped being a forecast. Meta Platforms signed its first major data centre agreement in the country with Reliance Industries, and the numbers are not symbolic: 168 megawatts of AI-ready capacity, live within two years, with contractual room to grow.

1. The deal at a glance

Under the agreement, Reliance Industries takes on the entire build and run: site design, construction, utility management, renewable power supply, network connectivity and managed operational services. Meta leases the full 168 MW and pays for the energy and water its workloads consume. The split matters — Meta gets hyperscale capacity without owning Indian physical infrastructure, and Reliance keeps the operating asset.

The facility is expected to be operational within two years, with contractual options to scale beyond the initial 168 MW. It becomes Meta's first built-to-suit data centre capacity in India, specified from the outset for the compute density that AI training and inference demand rather than retrofitted from a general cloud footprint.

2. Why Jamnagar — the strategic logic

Jamnagar is not an arbitrary pin on a map. The Gujarat city has become one of India's most significant locations for heavy industrial and infrastructure projects, and Reliance has decades of presence there — the Jamnagar refinery complex is among the largest in the world. Five factors make it work for AI infrastructure specifically:

• Land at campus scale. A 168 MW facility needs a large contiguous footprint, which Jamnagar can offer without the land and planning constraints of Mumbai, Bengaluru or Chennai.
• Energy access. The region has strong grid infrastructure and is well positioned for the large renewable power purchase agreements that AI operators now need to sign years ahead of capacity going live.
• Coastal location. Proximity to the Arabian Sea makes seawater desalination cooling viable — the single design decision that makes a facility this size defensible in India's climate.
• An existing Reliance campus. Meta is plugging into a proven environment with power, connectivity and operations already established, rather than starting greenfield.
• State policy support. Gujarat has been deliberate about attracting global technology investment, with dedicated infrastructure zones and a favourable data centre policy framework.

3. Renewable power and seawater cooling

The environmental profile is the most technically interesting part of the announcement. AI workloads are power-hungry and, historically, water-hungry. Jamnagar is being designed to attack both constraints at the same time.

Roughly 1 GW of contracted renewable capacity

Alongside the data centre, Meta confirmed renewable energy partnerships with two Indian firms, CleanMax and Fourth Partner Energy. Together those agreements add approximately 1 GW of new renewable capacity dedicated to supporting the facility. At that ratio of contracted clean generation to leased load, Jamnagar would sit among the most renewably powered AI infrastructure projects in Asia.

Seawater desalination instead of freshwater cooling

Conventional data centres consume large volumes of potable water for evaporative cooling — a serious problem in water-stressed regions. Jamnagar will meet its cooling requirement from desalinated seawater, removing freshwater abstraction from the equation. It is operationally more complex and more energy-intensive than a standard cooling loop, but the coastal location makes it practical, and it sets a water usage benchmark other Indian projects will now be measured against.

4. Meta and Reliance: six years of build-up

Jamnagar is not a standing start. It is the latest step in a relationship that now spans six years and several billion dollars of committed capital.

• 2020Meta invests $5.7 billion in Jio Platforms

  Meta acquired a 9.99% stake in Reliance's digital and telecoms arm — its largest single investment in a company at the time, and a decisive entry into the Indian digital market.

• 2025$100 million enterprise AI joint venture

  The two companies established a joint venture building AI solutions for enterprises across India and selected international markets — the commercial groundwork for the infrastructure commitment that followed.

• 2026168 MW Jamnagar AI data centre

  Meta's first built-to-suit data centre capacity in India, operated end to end by Reliance. Formal entry into Indian AI infrastructure at scale.

For Reliance, the deal converts it into an anchor AI infrastructure provider with global tenants — a position worth considerably more over time than the lease revenue from one campus. For Meta, India stops being a market served from overseas capacity and becomes an AI infrastructure base in its own right, with the data residency and latency advantages that follow.

5. What the leaders are saying

The framing from both sides is deliberate. Neither describes a vendor-client transaction; both describe a long-term strategic commitment to India's position in global AI. That signalling is part of the point — it tells every other hyperscaler that India is now competing for the largest AI infrastructure investments, not just the regional ones.

6. India's AI infrastructure outlook to 2030

The Meta–Reliance deal lands against extraordinary momentum in Indian data centre investment. The headline numbers set the context:

Going from roughly 1,700 MW to 8,000 MW by 2030 is close to a fivefold increase. AI is the driver: AI workloads need far more power per rack, far more cooling and far more specialised networking than the cloud and e-commerce workloads that powered the previous decade of Indian data centre growth. Government policy has been an active enabler through dedicated data centre frameworks, tax incentives and a deliberately welcoming stance towards global cloud and AI operators. Jamnagar is, in part, a return on that policy.

7. What this means for technology businesses

For AI and software engineering teams

The facility is specifically intended to support Llama model localisation for the Indian market, starting with finance and healthcare. In practice that means Indian-market applications built on Llama get dedicated, low-latency in-country compute. If you are shipping on Meta's open model stack for Indian users, this changes your latency budget, your data residency story and eventually your unit economics.

For UK businesses watching India

A build-out this size creates demand well beyond the concrete: integration work, platform engineering, AI application development and the long tail of services that surrounds a major campus. UK technology firms that establish Indian relationships while the infrastructure is being laid will be better placed when the capacity actually comes online. If you are considering an offshore engineering base, the compute story is now part of the decision, not just the cost story — and AI engineering talent tends to cluster where the compute lands.

For the global AI infrastructure race

Jamnagar is one of several announcements putting India alongside the US, Europe and Southeast Asia as a core geography for AI compute. Google, Microsoft and Amazon have all committed large-scale Indian investment. The pace has no real precedent in digital infrastructure, and it feeds directly back into cloud pricing, model availability and the economics of building software anywhere in the world.

8. Frequently asked questions

What is the Meta and Reliance AI data centre deal?

Meta Platforms and Reliance Industries have agreed to build a 168 MW AI-enabled data centre in Jamnagar, Gujarat. Reliance designs, builds and operates the facility end to end; Meta leases the full 168 MW and pays for the energy and water it uses. It is Meta's first built-to-suit data centre capacity in India.

Why is the data centre being built in Jamnagar, Gujarat?

Jamnagar combines large-footprint land, coastal access for seawater cooling, strong energy infrastructure and an existing Reliance data centre campus. Gujarat's state government has also been proactive in attracting global technology investment through dedicated infrastructure zones and favourable data centre policy.

What is the capacity of the Jamnagar AI data centre?

It launches with 168 megawatts of initial capacity and contractual options to expand beyond that. Power comes from approximately 1 GW of renewable capacity contracted through CleanMax and Fourth Partner Energy.

How will the Jamnagar data centre be cooled?

Through seawater desalination rather than freshwater abstraction. The Arabian Sea location makes that viable and removes potable water consumption from the cooling loop — a material advantage in a water-stressed region.

What is the history between Meta and Reliance?

Meta invested $5.7 billion in Jio Platforms for a 9.99% stake in 2020. In 2025 the two formed a $100 million joint venture focused on enterprise AI. Jamnagar is the largest commitment in that multi-year relationship.

What does the deal mean for India's AI infrastructure future?

India had roughly 1,700 MW of live capacity at the end of 2025 and is projected to reach 8 GW by 2030. A single 168 MW campus is close to 10% of today's total, so the deal both accelerates that trajectory and signals to other hyperscalers that India can host world-class AI compute.

How does this affect UK businesses and technology teams?

UK businesses serving Indian users gain lower-latency in-country AI compute and better data residency options. UK companies with Indian engineering teams get access to stronger infrastructure. And the wider race for AI capacity is reshaping cloud pricing, model availability and software development economics worldwide.