NSE Deploys Liquid Cooling for AI Infrastructure, Plans 80+ GPUs This Quarter

NSE Builds a Higher-Capacity Foundation for AI Computing

Liquid cooling and a planned GPU expansion point to a more demanding role for computing infrastructure across the exchange.

Cooling designed for growing AI workloads

The deployment combines two parts of a broader infrastructure push: direct-to-chip liquid cooling and a planned increase in graphics processing unit capacity. GPUs provide the computing resources used for AI workloads, while the cooling system is designed to support high-density computing. Together, the additions are intended to give NSE more capacity as its use of AI expands across the organisation.

The exchange described the installation as the first of its kind in Indian capital markets. That makes the announcement relevant beyond the equipment itself: it highlights the infrastructure choices that financial institutions may need to consider as they increase computing-intensive activity. NSE has not specified individual AI applications or the workload mix, so the deployment’s operational benefits cannot yet be assessed in detail.

GPU targets indicate a staged capacity build-out

NSE plans to scale GPU capacity to more than 80 this quarter and eventually to 224. The two figures describe different stages of the plan: the near-term target is an expansion milestone, while 224 is the longer-term ambition. The announcement does not provide a completion timeline for the latter, nor does it disclose the investment required.

That distinction matters when judging the development. The immediate step indicates that NSE is preparing for higher computing demand, but the eventual scale will depend on execution and the organisation’s evolving requirements. Capacity alone does not establish how quickly AI use cases will grow or what measurable outcomes the exchange expects from them.

Data control remains central to the deployment

NSE said the infrastructure is housed in its on-premises data centre, allowing it to retain control over AI data, computing resources and access. It linked that approach to data sovereignty, security governance and applicable regulatory and compliance requirements as AI adoption scales.

For a financial-market institution, this emphasis is a key part of the infrastructure story. The project is not only about adding processing capacity; it is also about how that capacity is operated and governed. The exchange has not detailed its safeguards or disclosed how the system will be used in specific functions, so those questions remain open.

Co-location extension depends on market requirements

NSE said it plans to extend direct-to-chip liquid cooling to its co-location data centre in the near future, subject to market participants’ requirements. This makes the next phase conditional rather than a confirmed, dated rollout. The statement signals that demand from participants will help shape whether and how the technology is introduced there.

For market users, the potential relevance is access to infrastructure designed for higher-density computing. However, NSE has not provided details on availability, timelines or the services that could use the cooling system. Those details will be important in assessing the practical effect of any co-location expansion.

Key Highlights

• NSE has deployed direct-to-chip liquid cooling in its on-premises data centre.
• GPU capacity is planned to exceed 80 this quarter, with 224 as a longer-term target.
• The infrastructure is intended to support expanding AI workloads and high-density computing.
• NSE says the on-premises setup retains control over data, computing resources and access.
• Extension to the co-location data centre is planned subject to market participants’ requirements.

Investor Note: The announcement is a technology and infrastructure development, not a disclosed financial forecast. Investors and market participants should watch for evidence of the GPU expansion, details on any co-location rollout and clearer information about the AI workloads the new capacity is intended to support.