India has completed the installation of over 17,300 graphics processing units (GPUs) as part of the IndiaAI Mission, Economic Times reported.
The IndiaAI Mission, with an outlay of more than ₹10,000 crore, has approved a proposal to purchase 34,333 GPUs in its first two rounds. It aims to create a cloud-computing platform for entrepreneurs and researchers to train their AI models. This will help the country develop its own large-language model (LLM).
During a meeting on 16 June 2025, IndiaAI CEO Abhishek Singh and all empanelled cloud-service providers (CSPs) reviewed GPU-installation progress, integration with the IndiaAI compute site and GPU-time allocation to end-users.
The objective is to create a scalable cloud-computing platform for researchers and start-ups to train artificial-intelligence models. The third round of bidding has been completed and the offers are undergoing technical examination.
Two of the ten CSPs selected in the first round, Jio Platforms and CtrlS Datacentres have yet to deploy their GPUs, but providers such as Yotta, NextGen and E2E Networks have made significant progress in installing and commissioning GPUs, according to sources.
The IndiaAI Mission raised this issue during the review meeting but the corporations claimed they have until 7 August 2025 to install the GPUs. The agreement required installation within six months of receiving the letter of intent on 7 February 2025.
Foundational Models Full Subsidy
The Indian government is providing a 100% subsidy on compute-infrastructure costs for companies developing foundational AI models, Moneycontrol reported.
It is critical to clarify that this 100% subsidy applies solely to core-model development. The existing 40% subsidy supports other GPU-intensive operations, such as inferencing and applications. This distinction is significant, as it was previously assumed that the 40% subsidy for GPU access also covered foundational-model training.
“This is not a 40% scheme like others foundational-model efforts are being given 100% subsidy on compute,” said a person directly familiar with the development. “Other AI workloads, like inferencing or vertical applications, will fall under the 40% bracket.”
Four companies, Sarvam, Gnan.AI, GAN.ai and Socket have been approved to develop foundational models. Sarvam, led by co-founder Pratyush Kumar, secured the largest single allocation, receiving 4,096 NVIDIA H100 SXM GPUs from Yotta Data Services, with a subsidy of ₹98.68 crore against a project cost of ₹246.71 crore. The other three start-ups are expected to receive comparable GPU allocations.
