NeevCloud and Agnikul Cosmos to build India’s first AI data centre in space
Move aims to deploy AI inference infrastructure directly in space for latency-sensitive applications.
NeevCloud, an Indian cloud infrastructure provider, has signed a memorandum of understanding (MoU) with Agnikul Cosmos to build what the companies describe as India’s first AI-powered data centre in space.
The initiative aims to deploy high-performance AI inference nodes in low-Earth orbit (LEO) to support real-time, latency-sensitive workloads.
Under the agreement, Agnikul will provide the launch vehicle and orbital hosting platform using its extendable upper-stage rocket architecture.
Unlike conventional missions where upper stages are discarded after payload deployment, Agnikul’s upper stage will remain in orbit and be repurposed as an operational hosting platform for NeevCloud’s AI computing infrastructure.
The lower stage will return to Earth.
NeevCloud will deploy and manage AI inference, storage and compute systems within this orbital platform through its cloud orchestration framework.
The first satellite configuration is expected to weigh between 300 kg and 350 kg within a total payload capacity of up to 500 kg.
Around 100 kg will account for the hosting platform. The system is expected to support approximately 500 high-performance AI chips.
According to the companies, a single satellite could handle up to 100,000 concurrent users or process around 10 million AI-driven requests per day, depending on workload intensity.
The satellite will operate at an altitude of 350 to 500 kilometres in LEO and will be powered entirely by solar energy with onboard battery storage.
The platform is designed to support applications requiring real-time AI processing, including defence and border surveillance, maritime and energy operations, industrial automation, autonomous systems, disaster response and remote healthcare.
The first proof-of-concept launch is scheduled before the end of this year, with commercial operations expected in 2027.
Following validation, the companies plan to scale the network to 30 to 40 satellites by 2027. In the longer term, NeevCloud aims to expand to over 600 orbital edge data centres by 2030 to create a continuous, near real-time AI inferencing network.