From PARAM to a National Computing Powerhouse

S Ahmad

---

India ‘s supercomputing journey has evolved from PARAM 8000 in the early 1990s to a national ecosystem under the National Supercomputing Mission ( NSM ) launched in 2015. With 40 supercomputers delivering 68 PF ( petaflops ) of superfast computing by September 2026 , NSM is expanding into high – performance computing ( HPC ).

When India unveiled the PARAM 8000 in 1991, the achievement carried significance far beyond the machine itself. At a time when access to advanced computing technology was closely controlled and India faced restrictions on acquiring high-performance systems, developing a supercomputer domestically was an assertion of technological capability and strategic confidence.

 

More than three decades later, that early experiment has evolved into something considerably larger.

India is building a national high-performance computing ecosystem under the National Supercomputing Mission (NSM), launched in 2015 with an outlay of about ₹4,500 crore. By September 2026, 40 supercomputers with a combined capacity of 68 petaflops had been deployed across the country, supporting research institutions and applications ranging from weather forecasting and genomics to flood prediction, astrophysics and materials science.

 

The significance of this transition lies not merely in the number of machines or the number of calculations they can perform per second. It lies in a much bigger shift: India is moving from acquiring computing capability to developing an ecosystem around it.

 

That distinction could prove decisive for the country’s technological future.

 

Supercomputing is often presented in terms of enormous numbers. A petaflop represents one quadrillion floating-point operations per second. The figures can be difficult to visualise, but the underlying idea is relatively simple. A supercomputer brings together thousands of processors that work simultaneously on problems too large or complex for ordinary computers.

 

The value of such systems, therefore, is not their speed for its own sake. It is what that speed makes possible.

 

A weather model can process vast quantities of atmospheric data. A flood-prediction system can simulate how water will move through a river basin. Researchers can model molecules in the search for potential drugs, study materials at the atomic level or examine astronomical phenomena that cannot be reproduced in a laboratory.

 

Computing becomes a scientific instrument.

 

This is why India’s supercomputing journey deserves to be viewed as part of the country’s broader development story.

 

For a country of India’s scale, the ability to generate and analyse enormous quantities of data is increasingly becoming a strategic capability. India generates approximately one-fifth of the world’s data, while the expansion of artificial intelligence, digital services, satellite applications and connected infrastructure is creating an even greater demand for computing power.

 

The question is no longer whether India needs advanced computing. It is how quickly it can build sufficient capacity, how widely that capacity can be accessed and how much of the underlying technology can be developed domestically.

 

The National Supercomputing Mission was designed with precisely this challenge in mind.

 

Jointly administered by the Department of Science and Technology and the Ministry of Electronics and Information Technology, the mission seeks to establish a globally competitive high-performance computing ecosystem. Its implementation involves the Centre for Development of Advanced Computing (C-DAC), Pune, and the Indian Institute of Science, Bengaluru.

 

Importantly, the mission follows a phased approach that goes beyond simply assembling machines. It seeks to move towards manufacturing and ultimately design and manufacturing support for indigenous systems.

 

That evolution is perhaps more important than the headline computing capacity.

 

A country can purchase a supercomputer. Building an ecosystem capable of designing servers, developing system software, creating high-speed interconnects, manufacturing components, developing cooling systems and training people to operate and use the machines is considerably harder.

 

India is now attempting the latter.

 

The development of the Rudra series of servers is an important example. Designed and developed indigenously by C-DAC and transferred to Indian electronics manufacturing partners, these servers form a critical part of the PARAM Rudra systems.

 

By September 2026, around 6,000 Rudra servers had been deployed in PARAM Rudra systems, with another 1,500 under construction.

 

This matters because computing sovereignty cannot be measured only by the number of supercomputers installed. It also depends on how much of the technology stack a country can control.

 

Hardware is one part of that equation. Software is another.

 

High-performance computing requires specialised operating environments, compilers, libraries, scheduling systems and applications. It also requires high-speed communication between thousands of computing nodes. A weakness in any one of these layers can limit the effectiveness of the entire system.

 

The development of indigenous high-speed interconnect technology, tested at 100 and 200 gigabits per second, therefore represents an important capability. So does the development of an indigenous HPC software stack and cooling technology.

 

Cooling, in particular, is an easily overlooked aspect of supercomputing.

 

The more computing power a facility contains, the greater the heat generated and the more energy required to maintain operating temperatures. As India expands its HPC infrastructure, energy efficiency will become inseparable from computing capacity.

 

The future supercomputer cannot simply be faster. It must also be more energy-efficient, reliable and economically sustainable.

 

This is especially important as artificial intelligence and high-performance computing increasingly converge.

 

The computing requirements of advanced AI models are enormous. At the same time, many scientific problems increasingly require a combination of machine learning and conventional numerical simulation. Weather forecasting, drug discovery, materials science, astronomy and engineering can all benefit from this convergence.

 

India therefore has an opportunity to build an ecosystem in which HPC is not treated as an isolated scientific facility but as infrastructure supporting the next generation of AI and scientific research.

 

Yet the real test of any technology mission is ultimately its impact beyond the laboratory.

 

Here, India’s emerging supercomputing ecosystem offers some encouraging examples.

 

One of the most consequential applications is weather and climate modelling.

 

India’s vulnerability to floods, cyclones, heatwaves, changing rainfall patterns and other climate-related events makes better prediction a public necessity. More powerful computing allows scientists to work with increasingly detailed models and larger datasets.

 

Similarly, flood-prediction systems can provide advance warnings by modelling river-basin behaviour. Under the NSM, an early-warning system has been developed for the Mahanadi basin that can predict floods up to two days in advance.

 

Two days may sound like a small window. For a district administration, however, two days can mean time to move people, protect livestock, secure essential infrastructure and prepare relief systems.

 

This is where computing power becomes public infrastructure.

 

The same principle applies to forest fires. By combining satellite remote sensing with computational models, researchers can simulate the spread of fires and support response planning. In a country with fragile Himalayan and forest ecosystems, the ability to anticipate rather than merely react can save both lives and natural resources.

 

Urban environmental management is another area where advanced computing can have a direct public impact.

An Urban Environmental Decision Support System can model weather and air pollution and help anticipate heavy rainfall and pollution episodes. Such tools can help cities move towards evidence-based environmental management rather than relying solely on historical patterns or reactive measures.

 

In healthcare, the potential is equally significant.

 

The genomics and drug-discovery platform developed under the mission can analyse large numbers of molecules to support the search for potential medicines. During the Covid-19 period, computational approaches were used to examine existing drugs and predict possible side effects, including cardiovascular risks.

 

Computational chemistry and materials science open another frontier. Simulating atoms, molecules and alloys allows researchers to investigate materials and their properties before expensive physical experiments are undertaken.

 

In energy and natural-resource exploration, the indigenous seismic imaging suite demonstrates another application. Advanced algorithms can map subsurface structures and support oil and gas exploration.

 

These applications illustrate a larger point: the true measure of supercomputing is not how fast a machine performs a calculation but whether that calculation helps solve a problem that matters.

 

India’s mission is also notable because it is building capacity across institutions rather than concentrating everything in a handful of laboratories.

 

The NSM infrastructure has supported more than 16,000 researchers, including more than 2,900 PhD scholars, across over 400 institutions. By September 2026, the systems had executed more than 15 million compute jobs and contributed to more than 1,990 research publications.

 

Those figures suggest that the mission is gradually becoming an enabling platform for a wider research community.

That is crucial.

 

A supercomputer locked inside one institution can be a powerful machine. A network of researchers capable of using advanced computing can become a national scientific capability.

 

The National Knowledge Network plays an important role in this regard. By connecting academic and research institutions through high-speed infrastructure, it allows computing resources and scientific expertise to be shared across geographical boundaries.

 

The challenge now is to deepen this democratisation.

 

India’s premier institutions have considerable research capabilities, but the country’s scientific potential cannot be confined to a few metropolitan centres. Universities, engineering colleges and research institutions in smaller cities and regions need access to HPC resources, training and computational expertise.

 

The development of PARAM Cub, described as a “supercomputing-in-a-box” system for students and researchers, points in this direction.

 

The idea is significant because the next generation of Indian scientists must not encounter high-performance computing only after entering elite research institutions. They should encounter it during their education.

 

That requires a parallel investment in people.

 

Computing infrastructure without skilled users is expensive hardware waiting for problems to solve.

 

The NSM has therefore invested in awareness programmes, hackathons, bootcamps, faculty development programmes and structured courses. Training has extended beyond computer science to faculty from other disciplines, recognising that HPC is increasingly relevant to physics, chemistry, biology, engineering, environmental science and other fields.

 

The annual EduHPC workshop, HPC courses on the SWAYAM platform, the HPC Education Portal and the NSM User Forum all contribute to building this wider ecosystem.

 

This may ultimately prove to be one of the mission’s most important contributions.

Machines become obsolete. Skills compound.

 

A researcher trained to formulate problems for high-performance computing can continue contributing to science even as hardware changes. A university that develops a strong computational research culture can produce generations of researchers capable of working across disciplines.

 

This is why India’s supercomputing strategy should increasingly be understood as a human-capital strategy as much as a technology strategy.

 

There is, however, an important caution.

 

Building computing capacity is expensive. Operating large HPC facilities requires electricity, cooling, specialised staff, maintenance and continuous technological upgrades. The rapid pace of change in processors, accelerators and AI hardware means that today’s leading system can quickly become yesterday’s technology.

 

India must therefore avoid measuring progress simply by installed petaflops.

 

The more meaningful indicators will be utilisation, research outcomes, patents, scientific discoveries, industrial applications, start-ups, training and measurable improvements in public services.

 

A machine that operates below capacity is not necessarily a failure, but a national programme must continually ask whether computing resources are reaching the problems that need them most.

 

This is where industry becomes important.

 

Government and academic institutions can establish foundational infrastructure, but commercial applications can help create a sustainable ecosystem. Pharmaceutical companies, climate-tech firms, semiconductor companies, engineering businesses, space enterprises and AI start-ups all have potential uses for HPC.

 

Greater interaction between government, universities, industry and start-ups could help convert computing capability into economic value.

 

India’s ambitions for a developed economy by 2047 will depend not only on building physical infrastructure but also on building technological infrastructure capable of supporting high-value innovation.

 

Supercomputing belongs to that category.

 

There is also a strategic dimension.

 

Dependence on foreign high-performance computing technology can become a vulnerability when access to advanced components, software or technical support is affected by geopolitical considerations. India’s experience in the early years of supercomputing demonstrated this reality.

 

The lesson from the PARAM 8000 era was that restrictions can force innovation.

 

The lesson today should be different but complementary: indigenous capability provides strategic resilience.

 

Self-reliance, however, should not mean technological isolation.

 

India will continue to work within a global scientific ecosystem. International collaboration, access to global research and participation in international technology standards remain essential.

 

The objective should be the ability to choose, innovate, collaborate and compete from a position of strength.

That is a more mature understanding of technological self-reliance.

 

India’s supercomputing story has therefore travelled a long distance—from a 1-gigaflop PARAM 8000 in 1991 to a national ecosystem involving dozens of systems, indigenous servers, software, networking and cooling technologies.

The journey from PARAM to PARAM Rudra is not simply a story of faster machines.

 

It is a story of institutional learning.

 

The country has gradually moved from demonstrating that it can build a supercomputer to asking what an indigenous computing ecosystem should look like.

 

The next stage should be even more ambitious.

 

India must develop systems that are faster and more energy-efficient, expand access to researchers, strengthen domestic manufacturing, deepen software capabilities and integrate HPC with artificial intelligence.

 

But above all, it must connect advanced computing to national priorities.

 

A farmer does not need to know how many petaflops a machine can deliver. What matters is whether better computing can improve a crop forecast.

 

A family threatened by floods does not need to understand numerical modelling. What matters is whether the warning arrives early enough.

 

A patient does not need to understand molecular simulation. What matters is whether computational research contributes to a safer and more effective medicine.

 

A city resident may never see a supercomputer. What matters is whether its computing power helps predict pollution or extreme rainfall.

 

That is the ultimate test.

 

Technology becomes transformative when it disappears into the quality of everyday life.

 

India’s supercomputing ecosystem is still evolving. The planned expansion to 50 systems with cumulative capacity exceeding 123 petaflops represents another step in that journey. But capacity alone will not determine success.

 

The real opportunity lies in building an ecosystem where machines, software, researchers, universities, industry and public institutions work together.

 

The country that once struggled to acquire advanced computing technology is now developing its own servers, software, networks and applications.

 

The achievement is substantial.

 

But the larger ambition should be to ensure that India’s computing power produces something more valuable than speed: knowledge, resilience, innovation and solutions at national scale.

 

The first chapter of India’s supercomputing story was about proving that the country could build.

 

The next must be about proving what the country can do with what it has built.

 

If that transition succeeds, the National Supercomputing Mission will be remembered not merely for the machines it installed, but for the scientific and technological capabilities it helped unleash.

 

That is the real promise of India’s emerging supercomputing ecosystem—and perhaps one of the most important building blocks for the India of 2047.