--- title: "How Indian Startups Can Get Subsidised GPUs Under the IndiaAI Mission" description: "The IndiaAI Mission offers GPU computing at heavily subsidised rates. Here is how many GPUs are available, what they cost and how startups can apply." author: "Ravi Karn" published: 2026-11-09 updated: 2026-11-09 canonical: https://answers.mindstick.com/blog/1532/how-indian-startups-can-get-subsidised-gpus-under-the-indiaai-mission category: "Latest Development" tags: ["IndiaAI Mission", "startups", "GPUs"] reading_time: 5 minutes --- # How Indian Startups Can Get Subsidised GPUs Under the IndiaAI Mission Building AI products requires computing power, especially graphics processing units, or GPUs. Renting high-end GPUs from global cloud providers can be very expensive, which puts many Indian startups and researchers at a disadvantage. The IndiaAI Mission aims to change that by offering GPU access at heavily subsidised rates. Here is how it works and how startups can benefit. ## What the IndiaAI Mission is The IndiaAI Mission is a government programme approved in March 2024 with an outlay of over ?10,000 crore, often cited as about ?10,372 crore. It aims to build India's AI capabilities across several pillars: computing infrastructure, datasets, foundation models, AI skills, startup financing, applications and safe and trusted AI. Computing infrastructure is one of the biggest parts. According to reporting in June 2026, nearly half of the mission's budget goes toward compute. ## How many GPUs are available The mission's initial target was 10,000 GPUs. Through competitive bidding, the government empanelled data centre and cloud providers to offer GPUs at fixed, subsidised rates. Reports describe more than 18,000 GPUs empanelled in early phases, with over 38,000 GPUs available through the broader compute ecosystem at subsidised rates as of mid-2026. These include high-end GPUs used for training and running AI models. Providers named in reports include companies such as Yotta, E2E Networks and NxtGen, among others. ## How much it costs Competitive bidding brought rates far below typical international prices. Reports put the subsidised rate at around ?65 per GPU-hour in some cases, well below global market rates for comparable hardware. Rates vary by GPU type and provider. In addition, the government has used a voucher-style system covering a significant portion of compute costs, reported at roughly 40 to 60 percent, with users paying the rest. Exact rates and subsidy levels can change, so check the official IndiaAI Compute Portal for current figures. ## Who can apply According to reports, eligible applicants include: - Startups. - MSMEs. - Researchers and academic institutions. - Government departments and agencies. - Other approved entities working on AI. Priority may be given to projects aligned with national goals, such as Indian language AI, healthcare, agriculture and public services. ## How to apply The process generally works like this: 1. Register on the IndiaAI Compute Portal. 2. Provide details about your organisation and eligibility, such as startup registration. 3. Describe your project and how you will use the compute. 4. Specify the GPU type and number of compute hours you need. 5. **Submit for review.** Larger requests go through additional evaluation. 6. Once approved, access GPUs through an empanelled provider. Allocation depends on eligibility, requested hours and subsidy requirements. ## What startups can use it for - Training or fine-tuning AI models, especially for Indian languages or specific industries. - Running inference for AI products, if allowed under the programme's terms. - Research and experimentation. - Building and testing prototypes before raising investment. ## Things to keep in mind Plan your compute needs carefully. Estimate how many GPU hours you need for training, fine-tuning or experiments. Over-requesting may slow approval, while under-requesting may leave you short. **Check lease terms.** Reports have mentioned limits such as a seven-day lease cap on some fine-tuning runs, which may affect longer training jobs. Plan training in stages if needed. Understand data and security requirements. Check where data will be stored and the provider's security practices, especially if you handle sensitive data. **Compare options.** For some workloads, using APIs from AI providers may be cheaper or simpler than training your own model. Subsidised GPUs are most valuable when you genuinely need to train, fine-tune or host models yourself. **Build the skills.** Using GPUs effectively requires engineering skills in machine learning, distributed training and infrastructure. Factor in time to learn or hire. ## Other IndiaAI support The IndiaAI Mission also supports: - AIKosh, a platform for datasets and AI resources. - Development of Indian foundation models, with several companies selected to build models for India. - AI skills programmes and fellowships. - Startup financing initiatives. - Safe and trusted AI projects. ## Challenges Reports have noted a gap between announced capacity and actual deployment at times, as well as questions about whether enough funding goes to high-quality datasets compared with hardware. Access can also depend on administrative processes. These are common challenges for large programmes, and the system continues to evolve. ## Why it matters For Indian startups, compute costs can make or break an AI idea. Subsidised GPUs lower the barrier to experimenting, training models for Indian languages and building products for Indian and global markets. Combined with very cheap AI APIs and open-weight models, they help level the playing field with well-funded companies abroad. ## The bottom line The IndiaAI Mission offers Indian startups, researchers and institutions access to tens of thousands of GPUs at heavily subsidised rates, reported at around ?65 per GPU-hour in some cases, through the IndiaAI Compute Portal. Plan your compute needs carefully, check lease terms and data requirements, and use the programme for work that genuinely requires your own GPU access. For many Indian AI startups, it can turn an expensive idea into an affordable one. --- Original Source: https://answers.mindstick.com/blog/1532/how-indian-startups-can-get-subsidised-gpus-under-the-indiaai-mission Copyright © MindStick Software Pvt. Ltd. This Markdown version is provided for developers, AI systems, and offline reading.