It’s a no-nonsense tie-up to underpin India’s AI stack with some serious infrastructure and open models. With a combined pot of roughly Rs 855 crore (or $100 million), the two are in the process of standing up this new facility in Jamnagar.
Why this partnership matters
You have Meta’s Llama models on one side and Reliance’s country-wide compute and clout on the other. Put them together and you can put advanced AI in the hands of developers and startups, making India’s digital economy a springboard for the next wave of products.
Mukesh Ambani sees it as a way to ‘democratize enterprise-grade AI for every Indian organization – from the small time to the big boys.’ For Mark Zuckerberg, it’s about getting Llama out in the field and putting down more roots in the enterprise space.
Two-track strategy: software platform plus data centre muscle
The JV has two things to put on the table. One is an AI Platform-as-a-Service for the enterprise to put their generative AI to work in sales, IT, or finance. The other is a batch of pre-made AI solutions for whatever you need, whether it’s across the board or in a specific sector.
On the hardware side, Reliance is putting in a 168 MW data centre for Meta to take a lease on, with room to grow. It’s meant to be the backbone for training and running your AI, right here at home.
Renewable-first build in Jamnagar
Meta has the rights to the Jamnagar site, which is set up to be 100% renewable and will even use desalinated seawater to keep things cool. Meta will foot the bill for the power and water.
To make sure they have the juice, Meta has lined up just under 1 GW of clean energy in India. Some 837 MW of that is from new wind and solar by CleanMax in Rajasthan and Karnataka, pushing the total they’ve put in with them over 900 MW.
Reliance and Meta like to point out that the Jamnagar operation is backed by close to 1 GW of renewables – a sign of an infrastructure play that is as much about efficiency and sustainability as anything else.
Market positioning and enterprise reach
With Llama’s open-source side and Reliance’s door into thousands of companies, the friction to adopt is low. That puts them in a good spot to have some say in how Indian firms go about scaling their AI.
And it’s not just code; there’s the physical plant in Jamnagar to back up the future of AI in one of the most dynamic markets out there. All of this is pending the usual regulatory sign-off and should be done by Q4 2025.
What it offers businesses
When the JV is up and running, this is what you can count on:
– A way to put your own spin on and control generative AI
– Some off-the-shelf AI for common or niche requirements
– To deploy where you like: on-prem, in the cloud, or on our end
– The compute you need to train, infer, and put it to work at scale
Bigger bet on India’s AI infrastructure
They’re touting this as an ecosystem play. You add in Reliance’s distribution and you have a formula to build AI apps faster and without the cost and hassle of doing it alone.
This is the next step after the $5.7 billion Meta put into Jio back in 2020, only now it’s in green data centres. If everything goes to schedule, you can expect to see some of these in action in India come the Jamnagar launch.
