PARAM PRAGYA: AI-POWERED SUPERCOMPUTING FACILITY AT IIT DELHI

PRERNA FOR IAS

         Mehnat Aapki, Guidance Humari

         Roz ki Prelims Practice — Experts ke Saath

PARAM PRAGYA: AI-POWERED SUPERCOMPUTING FACILITY AT IIT DELHI

Context: Prime Minister Narendra Modi inaugurated PARAM Pragya on 8 August 2026 during IIT Delhi’s 57th Convocation. The AI-powered high-performance computing facility has been established at IIT Delhi’s Sonipat Campus and is intended to strengthen research in artificial intelligence, data science, advanced computing and interdisciplinary science.

Background: India has steadily built indigenous supercomputing capability since PARAM 8000 in 1991. This effort later expanded through the National Supercomputing Mission, which seeks to provide advanced computing infrastructure to research and academic institutions. PARAM Pragya represents the growing shift from conventional high-performance computing towards AI-intensive computing infrastructure.

Salient Features: According to IIT Delhi, PARAM Pragya is powered by 400 NVIDIA A100 GPUs with 80 GB memory each and provides around 250 AI petaflops of computing capability. This makes it one of the largest AI-oriented supercomputing facilities in Indian academia.

The facility can support large-scale machine learning, deep learning, data analytics and scientific simulations. Potential applications include climate modelling, material science, healthcare, genomics, engineering design and other computation-heavy research.

India’s Approach and Significance: India increasingly views advanced computing as a form of strategic digital infrastructure. Domestic access to powerful computing resources can reduce dependence on overseas cloud infrastructure, improve research capability and support the development of Indian AI models.

Current Status: PARAM Pragya is now operational at IIT Delhi’s Sonipat campus. IIT Delhi describes it as the latest academic supercomputing facility of its scale and expects it to expand access to high-end AI computing for researchers.

UPSC Note: Official sources confirm the 400 A100 GPUs and 250 AI-petaflop capability. Claims regarding an ₹120-crore cost, 8-PFLOPS conventional HPC rating, 800-kW power requirement, or that it was specifically built under NSM by C-DAC should be used cautiously unless separately confirmed by an official source.

Analytical Questions

1. Question: Why does India need its own high-end AI computing infrastructure when global cloud services are already available?

Answer: Foreign cloud services are useful, but complete dependence can create cost, access and data-security concerns. Domestic computing capacity gives researchers greater control over sensitive datasets and long research projects. It also helps build local skills. The aim should be strategic capacity, while still using global platforms where they offer clear advantages.

2. Question: How can PARAM Pragya contribute to public policy and not just academic research?

Answer: Its value can extend to areas like weather forecasting, disease modelling, traffic planning, agriculture and disaster management. Large datasets can be analysed faster and better models can be built. But policy decisions should not blindly follow computer output. Administrative judgement, ground data and public accountability must remain central.

3. Question: What are the main limitations of building powerful supercomputers for AI?

Answer: High-end systems are expensive to operate and need large amounts of electricity, cooling and skilled manpower. Hardware also becomes outdated quickly. So investment should include maintenance, energy efficiency, software capability and regular upgrades. The real test is not peak computing power, but how effectively researchers actually use the facility.

4. Question: Can a facility like PARAM Pragya make India self-reliant in artificial intelligence?

Answer: It can strengthen self-reliance, but it cannot achieve it alone. India also needs advanced chips, good datasets, skilled researchers, strong software ecosystems and sustained research funding. Dependence on imported hardware may continue. Self-reliance should therefore mean reducing critical vulnerabilities and building domestic capability across the full AI ecosystem.

5. Question: If you were managing access to PARAM Pragya, how would you ensure that the facility creates maximum public value?

Answer: I would use a transparent system to allocate computing time. Projects with strong scientific merit and public relevance should receive priority. Universities, startups and government research agencies should also get fair access. Usage, outcomes and idle capacity should be reviewed regularly so that expensive computing resources do not remain underused.