Meta’s India Bet in the AI Infrastructure Race

A New Front in the AI Race Opens in India

Meta has agreed to lease capacity at a new artificial-intelligence-enabled data center being built by Reliance Industries in Jamnagar, Gujarat, a deal that gives India one of its clearest signs yet that global technology companies are beginning to treat the country not only as a vast consumer market but also as a serious location for the physical infrastructure behind the AI boom.

The agreement, announced Monday, is Meta’s first AI data-center deal in India. The first phase of the Jamnagar site is expected to provide 168 megawatts of capacity, according to Meta, with room to expand further. The company also said it was supporting nearly 1 gigawatt of renewable energy in India alongside the project, underscoring how tightly the future of AI computing is now bound to power availability.

The arrangement gives fresh momentum to India’s effort to position itself as a major destination for AI investment at a moment when its broader economic story has become more complicated. For years, India has sold itself to global capital as the world’s fastest-growing major economy and an increasingly attractive alternative base for digital and industrial investment. But that narrative has recently come under greater scrutiny, as investors weigh signs of resilience against a softer outlook, inflation risks and global uncertainty.

From Policy Ambition to Concrete Capacity

For Indian officials, the Meta-Reliance agreement represents something more tangible than another policy announcement. It is a hyperscaler commitment to computing capacity inside India — the kind of long-duration investment that can signal confidence in a country’s infrastructure, regulation and demand.

That matters because AI infrastructure is unlike the software-led tech expansion of the past decade. Large language models and other advanced AI systems require enormous amounts of computing power, land, cooling and electricity, along with clear rules on taxation, data handling and investment. Countries hoping to capture a larger share of this industry must prove they can offer not just engineers and users, but also energy, transmission and construction at scale.

India has spent the past year trying to make that case. In February, the government said it would add 20,000 graphics processing units, or GPUs, to an existing base of 38,000, framing compute capacity as one of the foundational layers of a national AI stack that also includes chips, networks, energy and data centers. Officials have also pointed to a long-dated tax framework for foreign cloud providers using Indian facilities, presenting it as a source of policy visibility stretching to 2047.

The government’s own projections show why the stakes are so high. India’s installed data-center capacity now stands at about 960 megawatts, according to official background materials, but could rise to 9.2 gigawatts by 2030. Major investments have already been announced by Microsoft, Amazon and Google. Meta’s move suggests that competition for India’s AI buildout is now entering a new phase, in which compute-intensive projects may become as important as cloud expansion once was.

Economic Promise Meets New Doubts

The timing is especially significant for Prime Minister Narendra Modi’s government, which is trying to preserve India’s appeal to foreign investors in his third term as the country’s growth outlook grows less straightforward.

On June 5, the Reserve Bank of India cut its growth forecast for the 2026-27 fiscal year to 6.6 percent, citing both domestic and external risks. On the same day, official data showed that gross domestic product had grown 7.7 percent in 2025-26, a reminder that India remains one of the world’s strongest large economies even as forward-looking indicators point to a more uneven road.

That tension — strong recent performance, but rising questions about durability — has become central to how investors view India. Slower global demand, pressure on the rupee, volatile energy prices and concerns about private investment have all made capital more selective. In that environment, a large AI infrastructure commitment carries weight beyond the technology sector. It suggests that some of the world’s biggest digital companies are willing to place long-term bets on India despite the more mixed macroeconomic backdrop.

For Reliance, controlled by Mukesh Ambani, the agreement also fits a broader strategy of building out India’s digital and energy infrastructure simultaneously. AI data centers are increasingly judged not only by server capacity, but also by whether operators can secure reliable low-cost power, preferably from cleaner sources that appeal to global customers and shareholders.

The Real Test: Execution

Still, the deal raises as many practical questions as it answers.

It remains unclear how quickly the Jamnagar capacity will come online, how much of the site’s expansion option Meta will ultimately use and whether other hyperscalers will follow with similarly large AI-focused commitments in India. A single project can signal momentum, but it does not by itself establish a durable market.

There is also the harder issue of execution. Building AI data centers at scale requires a difficult combination of land, water, chip access, grid connectivity and uninterrupted electricity. India’s policymakers have increasingly emphasized the role of clean energy in supporting data-center growth, but turning that vision into competitive operating economics will be crucial. If power costs rise or transmission bottlenecks emerge, the advantage of locating compute capacity in India could narrow.

That is why the renewable-energy component of the Meta-Reliance arrangement may prove almost as important as the lease itself. In the global contest for AI infrastructure, power is no longer a supporting detail; it is often the decisive factor.

Why This Matters Now

For much of the past decade, global technology companies approached India mainly as a huge market for smartphones, social media, e-commerce and digital payments. What is changing now is the possibility that India could also become a place where the heavy machinery of the AI era is built and run.

If more deals follow, that could deepen India’s role in the global technology supply chain and attract the kind of capital that tends to stay invested through economic cycles. It could also reinforce the government’s argument that India’s next phase of growth will be driven not just by consumption, but by strategic infrastructure tied to advanced industries.

Whether that promise is realized will depend on more than one agreement in Gujarat. It will hinge on whether India can deliver the basics of the AI age — abundant power, fast approvals, stable policy and competitive costs — even as its economy faces new tests. But with Meta’s decision to commit capacity in Jamnagar, the contest to host the world’s AI infrastructure has gained a formidable new arena.