NSE Liquid Cooling Upgrade Enables AI Expansion With Capacity for Up to 224 GPUs
The National Stock Exchange has deployed direct-to-chip liquid cooling as it prepares to expand its GPU infrastructure for growing artificial intelligence workloads.
The National Stock Exchange of India (NSE) has deployed Direct-to-Chip Liquid Cooling (DCLC) technology at its on-premises data centre to support its expanding artificial intelligence infrastructure.
Announced on November 9, 2026, the NSE liquid cooling deployment creates room for future computing expansion, with infrastructure designed to accommodate up to 224 graphics processing units (GPUs) across high-density racks.
The exchange already operates multiple GPUs and plans to increase its deployment to more than 80 GPUs during the current quarter. The 224-GPU figure represents the cooling infrastructure’s design capacity for further expansion.
NSE Liquid Cooling: Key Details
NSE described the deployment as the first of its kind in Indian capital markets. The upgrade addresses the cooling requirements associated with a growing concentration of powerful processors inside its data centre.
| Component | Announced details |
|---|---|
| Cooling technology | Direct-to-Chip Liquid Cooling |
| Deployment environment | NSE’s on-premises data centre |
| Existing GPU deployment | Multiple GPUs; exact count not disclosed |
| Near-term expansion | More than 80 GPUs during the current quarter |
| Infrastructure design capacity | Up to 224 GPUs across high-density racks |
| Planned extension | Co-location data centre, based on participant requirements |
The phased approach gives NSE scope to increase computing resources as its AI requirements develop.
How Direct-to-Chip Liquid Cooling Works
Direct-to-chip liquid cooling removes heat close to the processors that generate it.
In a typical system, cold plates are attached to components such as CPUs and GPUs. Coolant flows through channels within these plates, absorbs heat and carries it to a heat exchanger or cooling unit.
This arrangement allows heat to be collected directly from demanding components. Depending on the system’s design, air cooling may continue to handle other parts of the server.
For readers unfamiliar with the technology, the basic principle is straightforward: powerful processors generate heat while working, and that heat must be removed continuously to keep them within suitable operating temperatures.
Direct-to-chip cooling can help data centres accommodate more computing power within a limited rack footprint. Its effectiveness depends on the complete installation, including coolant circulation, heat removal and operating conditions.
Why Cooling Matters for AI Infrastructure
AI workloads can place sustained demands on GPUs. As more processors are installed together, the amount of heat generated within each rack can increase substantially.
Cooling therefore becomes part of capacity planning. Adding processors requires suitable power delivery, thermal management and supporting infrastructure.
Effective cooling also helps reduce the risk of thermal throttling, which occurs when a processor lowers its operating speed to control its temperature. Maintaining suitable temperatures supports more consistent performance during demanding workloads.
NSE Chief Technology Officer Sampath Manickam said increasing AI adoption was creating higher requirements for computing density, power and cooling. He positioned the deployment as preparation for the exchange’s evolving technology needs.
On-Premises Deployment Supports Operational Control
NSE has installed the cooling infrastructure within its own data centre as part of its approach to managing AI resources.
According to the exchange, the arrangement supports data sovereignty while allowing it to retain control over AI data, computing resources and access.
NSE also said the approach supports its security governance and alignment with applicable regulatory and compliance requirements as AI adoption expands.
The infrastructure decision consequently involves both processing capacity and how that capacity is managed. Keeping the environment on premises gives the exchange direct responsibility for its operation and access controls.
Liquid Cooling Could Extend to Co-Location Facilities
NSE plans to extend DCLC technology to its co-location data centre in the near future, depending on market participants’ requirements.
Co-location allows participants to house their systems within designated exchange premises and connect to trading systems and market data services. NSE provides supporting facilities such as rack space, power, connectivity and cooling, while participants manage the hardware and software within their allocated racks.
Introducing liquid cooling into that environment could support participants whose equipment requires more intensive thermal management.
The exchange has not announced a specific implementation date for this extension.
Efficiency Is a Goal, With Performance Figures Awaited
NSE Executive Director Viral Mody linked the investment to resilient infrastructure, operational efficiency and responsible resource use.
Liquid cooling can reduce some of the demands placed on conventional air cooling. However, the energy performance of an installation depends on its equipment, operating temperatures, utilisation and wider cooling system.
NSE’s announcement did not provide measured energy savings, installation costs or details of the GPU models being used. It also did not specify when the infrastructure would reach its full 224-GPU design capacity.
For the exchange, the immediate milestone is expansion beyond 80 GPUs during the current quarter. The larger cooling capacity provides scope to accommodate subsequent growth as its AI computing requirements increase.
For more insights navigate: PressPire | Stories That Shape the Future
Read previous article here : manastu-space-116-crore-green-propulsion/
Follow PressPire on Facebook | Instagram | LinkedIn | YouTube
PressPire | Stories That Shape the Future