Meta and Reliance Partner for AI Data Center in India

Table of Contents

• 1. Meta partners with Reliance for AI data center
• 2. Overview of the Meta-Reliance Partnership
• 3. Details of the AI Data Center in Jamnagar
• 4. Investment and Financial Implications
• 5. Sustainability Features of the Data Center
• 6. India’s Growing Role in AI Infrastructure
• 7. Strategic Importance for Meta and Reliance
• 8. Future Outlook for AI Investments in India
• 9. The Future of AI Infrastructure in India
• 9.1 Strategic Partnerships and Their Implications
• 9.2 The Role of Sustainability in AI Development

---

Meta partners with Reliance for AI data center

Meta Leases AI Capacity in India

• What happened: Meta struck its first AI data center deal in India with Reliance Industries.
• Where / how big: Jamnagar, Gujarat; 168 megawatts (MW) of AI-enabled capacity.
• Commercial structure: Meta leases capacity; Reliance designs, builds, and operates the facility (end-to-end).
• Timeline: Reliance says the site should be ready within ~2 years and expandable.
• Utilities: The companies say it will run on renewable energy and use desalinated seawater for cooling; Meta covers its own energy and water costs.
• Not disclosed (yet): contract value, specific AI workloads (training vs inference), hardware profile, and whether Meta will add more India capacity.

• Meta has struck its first AI data center deal in India, partnering with Reliance Industries on a 168-megawatt facility in Jamnagar, Gujarat.
• Meta will lease capacity, while Reliance will design, build, and operate the site, with delivery targeted within two years.
• The companies say the data center will run on renewable energy and use desalinated seawater for cooling, with Meta covering its energy and water costs.
• The move lands as India’s data center capacity has surged to about 1.5 gigawatts in 2025, with projections of more than 8 gigawatts by decade’s end.

Overview of the Meta-Reliance Partnership

Meta–Reliance AI Partnership Timeline

• 2020: Meta invests $5.7B in Reliance’s Jio Platforms, establishing a long-term strategic relationship.
• 2025: The companies launch a $100M joint venture to build enterprise AI solutions for India and overseas markets.
• 2026: Meta agrees to lease capacity in Reliance’s 168MW AI-enabled data center in Jamnagar (delivery targeted within ~2 years).

Meta’s new agreement with Reliance Industries marks a notable shift in how the U.S. tech giant is approaching AI infrastructure in one of its most important growth markets. The partnership, announced Wednesday, centers on a large AI-enabled data center in Jamnagar, Gujarat—at a time when global demand for computing power is reshaping where data centers get built.

The deal also extends a relationship that has been steadily deepening since 2020, when Meta invested $5.7 billion in Reliance’s Jio Platforms. Since then, the companies have broadened their collaboration beyond consumer-facing digital services. Last year, they launched a $100 million joint venture to develop enterprise AI solutions for customers in India and overseas markets, using Meta’s AI capabilities and Reliance’s reach across Indian industry.

In the Jamnagar arrangement, Reliance will provide end-to-end services—design, construction, renewable power, connectivity, and ongoing operations—while Meta leases capacity at the facility. The companies have not disclosed the value of the agreement, the specific AI workloads planned for the site, or whether Meta intends to pursue additional AI infrastructure investments in India.

What is confirmed is the structure: Meta is leasing capacity, while Reliance is responsible for design, construction, renewable power, connectivity, and ongoing operations, with the facility described as 168 megawatts and targeted to be ready within two years.

Still, the direction is clear: India is moving from being primarily a user market for global platforms to becoming a place where the underlying compute for AI systems is built and operated—often through partnerships with local conglomerates that can deliver power, land, and connectivity at scale.

Details of the AI Data Center in Jamnagar

Data Center Delivery Lifecycle
1. Site + requirements lock: confirm land, grid interconnect, fiber routes, and the initial MW tranche Meta will lease.
– Checkpoint: if power delivery dates or fiber paths slip, the “2-year” target usually slips with them.
2. Design + permitting: facility layout, electrical single-line design, cooling architecture (here: desalinated seawater), and local approvals.
– Checkpoint: cooling-water intake/outfall and desalination integration are often schedule-critical.
3. Build + fit-out: civil works, substations/transformers, generators/UPS, mechanical plant, and data halls.
– Checkpoint: long-lead equipment (transformers, switchgear, chillers) can gate commissioning.
4. Commission + handover to ops: staged energization, integrated systems testing, and reliability validation before production workloads.
– Checkpoint: “AI-ready” typically depends on power density and cooling stability under sustained load.
5. Operate + expand: Reliance runs day-to-day operations; Meta scales leased capacity as demand grows.
– Checkpoint: expansion depends on additional power/water provisioning and available adjacent build space.

At the center of the partnership is a 168-megawatt AI-enabled data center planned for Jamnagar, a city in Gujarat on India’s western coast. The size matters: megawatt capacity is a key yardstick for data centers, reflecting the power available to run servers and the cooling systems that keep them operating reliably—especially important for AI workloads that can be exceptionally energy-intensive.

Reliance says the facility will be ready within two years and can be expanded over time. That timeline places delivery around 2028, assuming construction and commissioning proceed as planned. The expansion option is significant because AI demand is volatile and fast-growing; companies increasingly prefer sites that can scale without starting from scratch in a new geography.

Meta’s role is that of an anchor tenant rather than an owner-operator. Under the agreement, Meta will lease capacity at the Jamnagar facility, and the companies say it will support Meta’s global infrastructure and AI computing requirements—plugging India more directly into Meta’s worldwide network of AI facilities.

The partnership also signals Reliance’s ambition to become a “one-stop shop” for AI infrastructure for global technology firms. Beyond the physical build, Reliance is positioning itself to deliver the full stack of requirements hyperscalers care about: reliable power (in this case, renewable), network connectivity, and operational management over the life of the facility.

What remains undisclosed is equally telling. The companies have not said what kinds of AI workloads will run there—training, inference, or a mix—nor have they detailed the hardware profile. That opacity is common in hyperscale infrastructure deals, where competitive advantage often lies in the specifics of model development and deployment.

Investment and Financial Implications

Leasing Versus Owning Capacity
Leasing capacity (what Meta is doing here)

• Upside: faster time-to-capacity; less construction/permitting execution risk; capex shifts to the operator; easier to scale up/down by contracting more/less MW.
• Tradeoffs: less control over facility design choices and expansion pacing; long-term costs depend on lease terms; still exposed to operating inputs (power/water), which Meta explicitly pays.

Building/owning (the alternative)

• Upside: maximum control over design, security posture, and expansion sequencing; potentially lower unit costs at very large scale.
• Tradeoffs: higher upfront capex; longer lead times; greater exposure to local delivery risks (land, approvals, grid interconnect, supply chain).

What this deal structure does not answer yet: total contract value, the exact MW Meta has committed to lease initially vs later, and whether Jamnagar is meant for training, inference, or both.

Neither Meta nor Reliance has disclosed the commercial value of the Jamnagar leasing agreement, leaving analysts to infer the financial stakes from the scale and from the companies’ broader commitments. Even without a price tag, the structure of the deal highlights how AI infrastructure is increasingly financed and de-risked: a local operator builds and runs the facility, while a global tech firm commits to leasing capacity—helping underwrite the investment.

Meta has also committed to covering the entire cost of the energy and water required to support its operations at the Jamnagar site. That pledge matters for two reasons. First, it clarifies that Reliance is not subsidizing Meta’s operating inputs; second, it underscores that energy and water are now central line items in AI-era data center economics, not peripheral utilities.

The agreement builds on earlier capital ties between the two companies. Meta’s $5.7 billion investment in Jio Platforms in 2020 established a long-term strategic relationship. More recently, the companies launched a $100 million joint venture to develop enterprise AI solutions, signaling that the partnership spans both “software layer” ambitions (AI tools and platforms) and the “infrastructure layer” required to run them.

The Jamnagar deal also lands amid a broader investment wave in India. Blackstone-backed AirTrunk has announced plans to invest $30 billion to build 5 gigawatts of data center capacity in India by 2030. Indian conglomerates including Adani and Tata Consultancy Services have also unveiled major data center expansion plans aimed at supporting AI workloads. For Meta, leasing in India can be read as a hedge against global capacity constraints and a way to diversify where compute is sourced.

For Reliance, the financial logic is equally straightforward: landing a global anchor tenant can improve utilization, support future expansion, and strengthen its pitch to other hyperscalers looking for turnkey AI-ready capacity in India.

Sustainability Features of the Data Center

Jamnagar Energy and Water Details

• Confirmed by the companies: renewable energy for the facility; desalinated seawater used for cooling.
• Confirmed cost responsibility: Meta says it will cover the full cost of energy and water for its operations at Jamnagar.
• Related (but separate) renewable sourcing: Meta says it contracted nearly 1GW of new renewable energy capacity in India via CleanMax and Fourth Partner Energy.
• Not disclosed yet (key context for “how green”): the specific renewable procurement structure for Jamnagar (on-site vs off-site, matching approach), expected efficiency metrics (e.g., PUE), and water system details (intake/outfall, recycling, blowdown handling).
• Not disclosed yet (operational reality): what portion of the 168MW will run at high utilization (AI training) vs bursty utilization (inference), which can materially change power and cooling profiles.

The companies are framing Jamnagar not just as a capacity play, but as a sustainability-forward build—an increasingly important requirement as AI pushes power consumption higher and scrutiny of data center footprints intensifies.

Reliance and Meta say the facility will be powered by renewable energy and cooled using desalinated seawater. Using desalinated seawater for cooling is a notable design choice in a country where freshwater constraints can complicate large industrial projects. While the companies have not provided engineering specifics, the headline approach suggests an attempt to reduce reliance on local freshwater supplies while still meeting the cooling demands of dense compute.

Meta has also said it contracted nearly 1 gigawatt of new renewable energy capacity in India through agreements with CleanMax and Fourth Partner Energy. Those renewable contracts are separate from the Jamnagar facility itself but are intended to supplement the renewable power supporting Meta’s operations there. In practice, such agreements are a common mechanism for large buyers to match electricity consumption with renewable generation at scale.

Sustainability features are not being treated as optional add-ons, but as core operating inputs that must be budgeted and secured over time.

The broader context is that sustainability is becoming a competitive differentiator for data center locations. As more AI workloads move into production, operators face pressure to demonstrate credible power sourcing and water management—especially in markets where grid reliability and resource constraints can be concerns. Jamnagar’s design choices appear tailored to that reality, even as many details remain undisclosed.

India’s Growing Role in AI Infrastructure

Meta’s move comes as India is rapidly emerging as a hub for AI and cloud infrastructure, driven by a combination of demand growth, policy incentives, and the search by global tech firms for new geographies to build data centers.

Government data shows India’s installed data center capacity has risen sharply since 2020. Industry estimates project that capacity could grow to over 8 gigawatts by the end of the decade, fueled by cloud adoption, AI workloads, and rising demand for local data processing.

In other words, the baseline (around 1.5 gigawatts in 2025) is presented as a government-reported snapshot, while the end-of-decade figure is an estimate—useful for direction, but not a committed build schedule.

Global technology companies have been increasingly vocal about India in their infrastructure plans. Microsoft, Amazon, Google, OpenAI, and Uber have recently announced AI and cloud infrastructure investments in the country, reflecting both the scale of the market and the strategic need to place compute closer to users and customers.

Policy is also playing a role. New Delhi has sought to attract investment through incentives, including tax exemptions through 2047 for foreign cloud providers on services sold overseas—so long as those workloads are run from Indian data centers. That kind of long-horizon incentive is designed to make India not only a destination for domestic demand, but also a base for exporting cloud services.

The rush extends beyond hyperscalers. AirTrunk’s planned $30 billion investment for 5 gigawatts by 2030 is one of the clearest signals that global capital sees India as a long-term data center market, not a short-term bet. Meanwhile, Indian conglomerates are positioning themselves as builders and operators of AI-ready capacity, with Reliance’s end-to-end pitch in the Meta deal serving as a prime example.

Strategic Importance for Meta and Reliance

Strategic Benefits for Both Sides

For Meta, leasing AI-ready capacity in India is a strategic infrastructure decision with global implications. The company is racing—like its peers—to secure the compute needed to train and deploy AI systems. By tying Jamnagar into its global infrastructure and AI computing requirements, Meta is effectively adding India as a node in its worldwide AI footprint, rather than treating the country solely as a market for apps and advertising.

The deal also reflects a pragmatic approach: instead of building and operating a facility alone, Meta is leaning on Reliance to deliver the project end to end. In a market where land, permitting, power procurement, and network buildouts can be complex, partnering with a local conglomerate can reduce execution risk and speed time to capacity.

For Reliance, the partnership is a high-profile validation of its ambition to become a turnkey provider of AI infrastructure. Reliance is signaling it wants to compete not just as a landlord, but as an integrated infrastructure operator—bundling energy, connectivity, and operations into a single offering that global tech firms can plug into.

Strategically, the Jamnagar facility also builds on the companies’ enterprise AI joint venture launched last year. Taken together, the moves suggest a two-layer strategy: develop AI solutions for enterprises (the “product” layer) while simultaneously ensuring access to the compute needed to run those solutions (the “infrastructure” layer).

What’s missing—details on workloads, contract value, and Meta’s next steps—doesn’t diminish the strategic message. The partnership indicates that India is now part of the global AI infrastructure chessboard, and that local industrial players like Reliance can play a central role in hosting and operating that compute.

Future Outlook for AI Investments in India

Key Deal Watchpoints Ahead
What to watch next (these details will change how the deal should be interpreted):

• Workload mix: whether Jamnagar is positioned for training, inference, or a hybrid (different power density, networking, and utilization patterns).
• Commercial scope: initial MW leased vs expansion options, contract duration, and whether Meta is the sole/primary tenant.
• Delivery signals: permitting milestones, grid interconnect progress, and commissioning phases that confirm the “within two years” target is holding.
• Power matching reality: how renewable sourcing is structured over time (and whether additional renewable contracts are announced).
• Follow-on deals: whether other hyperscalers sign similar end-to-end leases with Indian conglomerates, indicating a repeatable market template.

The Meta-Reliance agreement is likely to be read by the market as both a milestone and a signal. It is a milestone because it marks Meta’s first AI data center deal in India. It is a signal because it suggests that future AI capacity in the country may increasingly be delivered through partnerships—global tech firms leasing from local operators that can assemble power, water strategy, connectivity, and construction at scale.

India’s capacity trajectory provides the backdrop. With installed capacity rising to around 1.5 gigawatts in 2025 and projections pointing to more than 8 gigawatts by the end of the decade, the country is in a buildout phase where early anchor tenants can shape where clusters form and how quickly they expand.

The investment pipeline is already crowded. Alongside announcements from Microsoft, Amazon, Google, OpenAI, and Uber, the entry of large-scale investors like AirTrunk—planning $30 billion for 5 gigawatts by 2030—raises the competitive bar for speed, efficiency, and sustainability. Indian conglomerates’ expansion plans add further momentum, suggesting that domestic capital will compete with, and complement, foreign investment.

Policy incentives could further accelerate the trend, especially the tax exemptions through 2047 for foreign cloud providers on services sold overseas if workloads run from Indian data centers. That framework encourages India-based compute not only for domestic consumption, but also for export-oriented cloud services.

At the same time, key uncertainties remain. The companies have not said whether Meta will pursue additional AI infrastructure investments in India, nor what workloads will run in Jamnagar. Those details will matter for understanding whether India becomes a major training hub, an inference hub, or a mixed environment—each with different power, networking, and operational profiles.

The Future of AI Infrastructure in India

Strategic Partnerships and Their Implications

The Meta-Reliance deal illustrates a broader pattern in AI infrastructure: partnerships that combine global demand with local execution. Meta brings a need for large-scale AI compute and a global infrastructure footprint; Reliance brings the ability to deliver end-to-end services—design, construction, renewable power, connectivity, and operations—within India’s regulatory and industrial landscape.

If more deals follow this template, India’s AI infrastructure growth may be shaped less by single-company campuses and more by ecosystems of operators and anchor tenants. That could speed deployment, but it also concentrates influence among a smaller set of infrastructure providers capable of delivering at hyperscale.

The Role of Sustainability in AI Development

Jamnagar’s renewable power and desalinated seawater cooling highlight how sustainability is becoming inseparable from AI infrastructure planning. As AI workloads expand, the limiting factors are increasingly physical: power availability, cooling capacity, and water strategy.

Meta’s separate contracting of nearly 1 gigawatt of renewable energy capacity in India—through CleanMax and Fourth Partner Energy—underscores that energy procurement is now a strategic function, not a back-office utility decision. In that environment, data centers that can credibly align scale with renewable sourcing and responsible cooling approaches are likely to attract the next wave of AI tenants.

Perspective: This analysis is written from the lens of Martin Weidemann (weidemann.tech), focused on how large-scale infrastructure deals are structured—capacity leasing, end-to-end operators, and the operational economics of power and water that increasingly shape AI deployment decisions.

This piece reflects publicly available information and company statements about the Meta–Reliance Jamnagar data center partnership at the time of writing. Key commercial and operational details—such as pricing, workload mix, and precise build or commissioning milestones—have not been disclosed and may change as new information emerges. Estimates about India’s end-of-decade capacity are directional and should not be read as guaranteed outcomes.