India's NSE to Launch Nanosecond Trading in April
The National Stock Exchange of India (NSE) will roll out nanosecond-level latency for its trading matching engine, effective April 11, 2026. This upgrade represents a significant leap in speed and will impact high-frequency trading firms and others requiring ultra-low-latency execution. The move necessitates that market participants optimize their systems for sub-millisecond response times.
- This upgrade will increase the NSE's order processing capacity from the current 5-6 million transactions per second to nearly 100 million per second. The system's response time is expected to improve by a factor of nearly 1,000 from the current 100 microseconds. - To support this technological leap, the NSE is more than doubling its colocation infrastructure, expanding from the current 2,000 server racks to approximately 4,500. This allows trading firms to house their servers in the exchange's data center, which is critical for minimizing network latency. - The move is a direct response to the increasing dominance of algorithmic and high-frequency trading, which already accounts for over 50% of the total trading volume in the equity derivatives segment on the NSE. - This technological advancement places the NSE in the same league as the world's most technologically advanced exchanges, which have been in a "latency arms race" to attract high-frequency traders. - The Indian high-frequency trading market is on a significant growth trajectory, with projections showing it will reach USD 765.1 million by 2030, growing at a CAGR of 9.6%. - NSE's Managing Director and CEO, Ashishkumar Chauhan, has indicated that alongside the speed upgrade, the exchange is preparing to launch new products, including electricity futures and smaller gold futures contracts, to attract a wider range of investors. - While the upgrade promises greater efficiency, the NSE's CEO has also cautioned that higher speeds introduce heightened cybersecurity risks, urging brokers and vendors to bolster their security frameworks.