共有:
Meta · India AI Datacenter

Meta is finally putting
an AI home in India.

With hundreds of millions of users across WhatsApp, Instagram, and Facebook, Meta's AI infrastructure has long been centered on the U.S. That balance is shifting—teaming up with Reliance to build a 168MW AI data center in the state of Gujarat. We unpack, in diagrams, what it means for the "first Meta AI data center" to rise in India.

AI Navigate Editorial·2026.06.11·6 min read
BEFORE U.S. data centers Training and inference pooled in the U.S. AFTER India · 168MW Inference placed near demand
01
The Gap

Hundreds of millions of users,
a brain only in the U.S.

For Meta, India is one of its largest markets by user count. Combine WhatsApp, Instagram, and Facebook and monthly users reach into the hundreds of millions. Yet the AI "brain" those people interact with—the data centers that handle training and inference—has long remained U.S.-centered.

Over the past year or two, Google, Microsoft, and AWS have announced cloud investments in India running into the tens of billions of dollars. Through all of that, Meta held back on expanding its Indian data centers. Its partnership with Reliance Jio had existed for some time, but it had never been tied directly to AI infrastructure—and that is what changes now, for the first time.

Until nowAfter this announcement
AI infrastructure pooled in the U.S.A new AI-ready site built within India
Reliance partnership centered on apps/telecomThe partnership extended to an AI data center
Inference routed through the distant U.S.Inference right next to where demand is
Lagging behind the other majorsSecuring the "first Meta AI home"

The users are already there.
What was missing was a brain running close by.


02
The Deal

What, exactly, was decided

The focus is "location" and "partner," and "scale." It's about lifting the existing Reliance relationship from the app layer up to the infrastructure layer.

Meta has officially announced that it will partner with Reliance Industries to build its first AI data center in India. The location is Jamnagar, in the state of Gujarat. It extends the existing relationship with Reliance / Jio Platforms from the consumer-app layer up to the AI-infrastructure layer. Against the backdrop of U.S.–China tensions and India's enormous demand, U.S. hyperscalers are all racing to place AI training and inference capacity in India—and this lands as the "first Meta AI data center" within that wave.

01

Who to partner with

Rather than seeking out a new partner, Meta chose Reliance / Jio Platforms, with whom it already has deep ties in telecom and apps. That brings land, power, and on-the-ground operational know-how on board all at once.

02

Where to put it

The site is Jamnagar, Gujarat. Adjacent to Reliance's industrial cluster, it's a place where power and land are easy to secure. It becomes a site that is "close" to India's users.

03

What it changes

Meta steps into the AI domain, previously outside the partnership's scope. The essence this time is that an app partnership has been elevated into a data center partnership handling training and inference.

03
By The Numbers

168MW, as a
measure of scale

The impact of a "first build" can be measured in figures. Scale, location, partner—three axes give us the outline.

JAMNAGAR · GUJARAT POWER Supplies 168MW COMPUTE AI training & inference racks REACH Hundreds of millions of India's users
FIG. Turning power into compute, and compute into reach for the region's hundreds of millions of users—the site's role
168MW
Power scale of the AI-ready data center
First
As Meta's India AI home
Hundreds of millions
India's monthly users

A power scale of 168MW signals a "serious building" for an AI data center. Even compared with the India sites Google and Microsoft set up earlier, the figure shows that Meta has gone in with both feet on the ground, not just dipping a toe.

And what matters is that this isn't a solo investment from scratch—it is built on top of the existing relationship with Reliance. Because the heavy parts—land, power, and local operations—can be shared with a partner, the speed and feasibility of getting it off the ground increase.

04
So What

Who does it affect, and how

The gradient of impact is clear. For those watching India and Asia expansion it's practical; for Japan and the West it's an indirect implication.

Users in India

Meta AI's latency (response delay) improves, and tuning models for the region becomes easier too. It points toward gains in perceived speed and accuracy.

People running WhatsApp businesses in India

For businesses built on WhatsApp, the response speed and accuracy of AI features trend upward. It's a direct tailwind for operators using AI for customer service and support on the ground.

People in Japan and the West

The direct impact is slight. Still, Meta consolidating its footing in India as "a player with AI infrastructure" is worth keeping in mind for anyone considering Asia expansion.


05
Frontier

This is a redrawing of the map

This announcement is more than the construction of a single data center. It's a geopolitical signal that amid the trend of U.S. hyperscalers starting to shift AI training and inference capacity to India, Meta has officially joined that ranks. The map of AI infrastructure, once U.S.-centered, is beginning to be redrawn toward where the demand is.

That said, businesses and developers in Japan or the West don't need to change anything right away. This is a story about a shift in the center of gravity of "where you run AI". For anyone with India or Asia expansion in their sights, the fact that Meta now has a brain on the ground there adds one more option going forward.

AI Navigate — Daily Update · 2026.06.11