NSM 2.0 to chart path for indigenous exascale system with 7-year HPC infra plan, says leader

Indigenously built compute systems will be the flavour of the upcoming National Supercomputing Mission (NSM) 2.0, which is set to be rolled out shortly. This is the second of the National Supercomputing Mission that was launched in 2015 with the aim of building high-performance computing (HPC) capabilities.

In an interaction with TechCircle, Dr Hemant Darbari, mission director of NSM, said that key goal of NSM 2.0 will be to develop Exascale systems to meet the advanced computing needs of various applications like weather forecasting, astronomy, and drug discovery that require massive computational capabilities.

“We are building exascale systems which is 10^18 compute power. Our upcoming system, PARAM Shankh, is named after Shankh, which in the Indian numbering system denotes 10^18. This exascale system will deliver unprecedented performance, and several smaller yet powerful systems are also under development. These systems are expected to roll out around the 2027–28 timeline.”

To be sure, PARAM is a series of Indian supercomputers designed and assembled by the Centre for Development of Advanced Computing (C-DAC) in Pune. Darbari added that NSM 1.0 successfully laid the foundation for indigenous systems, leading to platforms like Rudra, and going ahead, all systems, including the exascale machines, will be designed and built indigenously. “This is part of a seven-year mission to achieve complete capability in high-performance computing infrastructure,” he said.

Rudra is a server developed by C-DAC, along with a high-speed interconnect called Trinetra, which connects the supercomputers.

“We are also working on digital twins, starting with process plants, in collaboration with NCL. On the smart solutions front, we are implementing advanced weather prediction and traffic management systems in Varanasi. Astronomy and astrophysics, which require enormous compute power, are key focus areas as well,” Darbari further added. To be sure, so far NSM has deployed over 37 supercomputers with a combined capacity of 40 petaflops. One petaflop is 10^15 FLOPs.

Darbari was speaking on the sidelines of the flagship Supercomputing India 2025 (SCI 2025) event in Bengaluru, which saw participation from 5000 delegates from HPC, AI, and semiconductor domains. Organised by C-DAC under the NSM and supported by MeitY, this event featured the signing of 12 memoranda of understanding, particularly for training and upskilling in advanced technologies.