India Looks to Orbit for Its Next Data Centre
Agnikul Cosmos and NeevCloud plan to test an AI data-center module in low Earth orbit before year-end.
Topics
News
- Cyber Attacks Hit Gulf Institutions as Regional Conflict Moves Online
- xAI and Tesla Launch “Macrohard,” a Playful Dig at Microsoft
- ‘Right Decision, Not Easy’: Atlassian Cuts 1,600 Jobs Ahead of AI Push
- AI-Assisted DevOps Under Spotlight After Amazon Cloud Disruption
- Saudi Arabia Doubles Down on AI Ambitions; Declares 2026 as the Year of AI
- LeCun’s New Startup Raises $1.03 Billion for AI Beyond LLMs
An Indian rocket startup and a domestic AI cloud provider plan to deploy computing capacity in low Earth orbit in what they describe as India’s first attempt at a space-based AI data center.
Chennai-based Agnikul Cosmos, which builds small satellite launch vehicles, has signed a commercial agreement with NeevCloud, an AI infrastructure arm of RackBank Datacenters.
Under the deal, an AI data-center module will be deployed on Agnikul’s rocket and hosted on its upper stage after launch.
A pilot mission is scheduled before the end of this year. If validated, NeevCloud plans to scale the concept into a network of more than 600 orbital edge data centers over the next three years.
The model relies on Agnikul’s extendable upper-stage design. After placing its primary payload in orbit, the rocket’s upper stage would remain operational and be repurposed as a computing platform.
“As with all rockets, with more launches we leave upper stages in orbit,” Srinath Ravichandran, co-founder and chief executive of Agnikul. “Our convertible upper-stage technology allows them to remain functional and host compute hardware in space.”
NeevCloud will deploy and operate the AI hardware on the orbital platform, powered by solar power and integrated into its cloud orchestration system to handle inference workloads.
“We are not just building a data center in space, we are building an entirely new layer of orbital inferencing infrastructure,” said Narendra Sen, founder and chief executive, NeevCloud.
He added that advanced AI systems need to be “decoupled from terrestrial limitations” to broaden access.
The companies said the orbital infrastructure is intended to support latency-sensitive applications such as defense, maritime operations, energy management and industrial automation.
By shifting some computing tasks into space, they said, it may be possible to reduce duplication of terrestrial edge data centres and associated costs, including land, cooling and energy consumption.
