More than 17,300 graphics processing units (GPUs) have been reportedly successfully installed under the IndiaAI Mission’s compute infrastructure tender, which has received proposals for a total of 34,333 GPUs across its first two bidding rounds. The third round of bidding has concluded and proposals are currently under technical evaluation. The mission aims to establish a robust cloud compute platform to support artificial intelligence (AI) model training for researchers and start-ups.
Out of the ten cloud service providers (CSPs) selected in the first round, reportedly Jio Platforms and CtrlS Datacenters have yet to install their GPUs. Others, including Yotta, NextGen, and E2E Networks, have made significant progress in both installation and commissioning of GPU infrastructure. Companies have until August 7, 2025, to complete the installation of their allocated GPUs, as per the agreement that required deployments within six months of the February 7 letter of intent.
CtrlS Datacenters has shared purchase orders but has yet to confirm GPU installation or initiate API integration with the compute portal. Meanwhile, Jio Platforms is in the process of procuring 752 Nvidia H200 GPUs and 268 AMD MI300X GPUs, with application programming interface (API) integration expected to complete by the third quarter of 2025.
Further, Yotta Data Services, the largest contributor in terms of GPU numbers, has installed nearly 50 per cent of its allocated GPUs and is awaiting delivery of the rest. Its API integration is fully completed. Yotta received 4,096 Nvidia H100 GPUs for Sarvam’s large language model (LLM) and 200 Nvidia H100 GPUs for Bhashini in early May. An additional 2,592 Nvidia H100 GPUs for Sarvam are set to be delivered by July 10. These units are currently undergoing commissioning and performance testing by Nvidia and Yotta engineers at the Navi Mumbai data centre. In the second round of empanelment, Yotta was also selected to deploy 8,192 Nvidia B200 (Blackwell) GPUs.
Furthermore, NxtGen Cloud Technologies has installed 51 per cent of its proposed GPU units. The remaining 49 per cent are under procurement and expected to be delivered by June 30, 2025. API integration is complete for the installed GPUs.
Moreover, Cyfuture has installed 4 per cent of its allocated GPUs, with 96 per cent still under procurement. API integration is currently in progress. Notably, Cyfuture was the L1 bidder across all the categories it applied for in the second round of bidding.
The IndiaAI Mission is also providing funding and support to start-ups building indigenous LLMs, including Sarvam, Gnani, Gan, and Soket AI Labs, to strengthen India’s AI capabilities.
