IndiaAI Mission Deploys 18,693 GPUs to Boost AI Research and Build India’s Own Language Model.
India is stepping up its AI game in a big way! The country has decided to host China’s open-source AI model, DeepSeek, on local servers while actively working on its own large language model (LLM).
The goal?
To have India’s very own AI model up and running within the next ten months.
Union Minister for Electronics and IT, Ashwini Vaishnaw, made this announcement just days after DeepSeek launched its budget-friendly AI model, shaking up the global AI industry.
Hosting DeepSeek on Indian Servers: A Smart Move?
Since DeepSeek is an open-source AI model, India sees an opportunity to host it locally, similar to what it did with Meta’s Llama AI model.
This move ensures that India can test data privacy measures and strengthen AI infrastructure.
“Since DeepSeek is open source, we can host it on our servers, just like Llama.
This will allow us to test and ensure data privacy,” Vaishnaw stated, addressing concerns about using AI models developed in China.
India’s Ambitious AI Plan: 18,600 GPUs for AI Research
While hosting DeepSeek is a short-term solution, India’s real focus is on creating its own AI language model.
This effort will be powered by the newly established IndiaAI Compute Facility, which has secured a whopping 18,600 GPUs—including 15,000 high-performance processors.
These GPUs are the backbone of AI model training and are crucial for India’s AI ambitions.
“In the next few months, India will have a world-class AI model. DeepSeek was trained on 2,000 GPUs, ChatGPT used 25,000 GPUs, and we now have 15,000 high-end GPUs ready. This puts us in a strong position to develop cutting-edge AI,” Vaishnaw explained.
Making AI Computing Affordable for Everyone
One of the key highlights of India’s AI plan is making high-performance computing accessible at a fraction of the global cost.
The government is offering a 40% subsidy to make AI computing more affordable for students, researchers, and startups.
“The average cost per GPU hour in India will be ₹115.85, compared to the global average of $2.5-$3 per GPU hour for low-end GPUs.
For high-end computing, it will be ₹150 per GPU hour, but after subsidies, students and researchers can access it for under ₹100,” Vaishnaw added.
To further support AI development, the government is also considering direct funding through grants and equity-based investments in companies building AI models.
Prioritizing AI Safety and Ethical Development
India is also taking AI safety seriously. The government is setting up eight institutions to focus on AI safety and ethical development.
These institutions will work together to address concerns like AI bias, machine unlearning, synthetic data generation, and privacy protection.
Some key projects include:
- Machine unlearning (IIT-Jodhpur)
- Synthetic data generation (IIT-Roorkee)
- AI bias mitigation and explainability (Defence Institute of Advanced Technology, Pune)
- Privacy-enhancing strategies (IIT-Delhi, IIIT-Delhi, and TEC)
- AI governance and ethical frameworks (IIT-Jodhpur, IIT-Roorkee, IIT-Delhi)
India’s Growing Role in AI Innovation
With its strong software expertise and booming startup ecosystem, India is positioning itself as a global AI leader.
By combining affordable AI computing, ethical AI governance, and a dedicated research framework, the country is making bold strides in the AI space.
“India will play a major role in the AI revolution because of our strong software capabilities and innovation ecosystem,” Vaishnaw emphasized.
With these efforts, India isn’t just keeping up with the AI boom, it’s setting the stage to be a powerhouse in AI development.
