The National Stock Exchange of India (NSE) has deployed Direct-to-Chip Liquid Cooling (DCLC) technology at its on-premises data centre as it expands its artificial intelligence computing capabilities. Announced on November 9, 2026, the deployment is designed to support high-density graphics processing unit (GPU) computing and prepare the exchange for increasingly demanding AI workloads. NSE describes it as the first deployment of its kind in India’s capital markets.
The exchange has already installed multiple GPUs and plans to expand its computing environment to more than 80 GPUs during the current quarter. Its new cooling infrastructure has been designed to accommodate up to 224 GPUs across high-density racks. NSE also plans to extend the technology to its co-location data centre in the future, depending on the requirements of market participants. The investment reflects the growing importance of specialised computing infrastructure as financial institutions adopt AI for data analysis, operations and other technology applications.
Key Takeaways
- NSE has deployed Direct-to-Chip Liquid Cooling technology at its on-premises data centre.
- The exchange plans to expand GPU capacity beyond 80 units during the current quarter.
- The cooling infrastructure is designed to support up to 224 GPUs across high-density racks.
- NSE says the deployment is the first of its kind in Indian capital markets.
- The infrastructure is intended to improve cooling capabilities for compute-intensive AI workloads.
- NSE plans to extend the technology to its co-location data centre based on market participants’ requirements.
- The exchange says keeping AI infrastructure on-premises strengthens data sovereignty and operational control.
What Is NSE’s Direct-to-Chip Liquid Cooling Technology?
Direct-to-Chip Liquid Cooling is a data-centre cooling technology that removes heat directly from high-performance processors using liquid coolant.
Traditional data centres commonly rely on air conditioning and fans to circulate cool air around servers. However, modern AI computing systems generate considerable heat because GPUs perform large numbers of calculations simultaneously.
As GPU installations become denser, removing heat efficiently becomes increasingly difficult. Conventional air-cooling systems may require substantial airflow and cooling capacity to maintain suitable operating temperatures.
DCLC addresses this challenge by bringing a liquid-cooling mechanism directly to the processor.
The system generally works through specialised cold plates installed in contact with heat-generating components. Coolant circulates through channels in these plates, absorbing heat from the processors. The heated liquid then travels through a cooling loop where the heat is transferred away before the coolant returns to the system.
This approach can support higher computing density and more efficient heat management.
Importantly, the coolant is typically contained within a dedicated cooling circuit rather than being poured directly onto electronic components.
How Direct-to-Chip Liquid Cooling Works
The technology uses a relatively straightforward heat-transfer process.
| Stage | How the system works |
|---|---|
| 1. Heat generation | GPUs and other processors generate heat while performing intensive calculations |
| 2. Heat collection | Cold plates attached to processors absorb heat |
| 3. Coolant circulation | Liquid moves through channels inside the cooling plates |
| 4. Heat transfer | Heated coolant carries energy away from the server components |
| 5. Cooling and circulation | The system removes heat from the coolant and circulates it again |
The key advantage is that heat is collected close to where it is generated rather than relying entirely on cooling the surrounding air.
For AI infrastructure, this can make it easier to operate high-density GPU racks without requiring increasingly large conventional air-cooling systems.
However, actual efficiency improvements depend on the design of the cooling loop, the facility’s power usage, the heat-rejection system and the workload. NSE has not published a specific percentage reduction in energy consumption for this deployment.
NSE Plans to Expand GPU Capacity Beyond 80 This Quarter
The liquid-cooling deployment is part of NSE’s broader plan to increase its AI computing capabilities.
The exchange has already installed multiple GPUs and intends to expand the environment to more than 80 GPUs during the current quarter.
The infrastructure has been designed with a larger expansion target of up to 224 GPUs across high-density racks.
| Infrastructure metric | NSE’s stated plan |
|---|---|
| Existing GPU deployment | Multiple GPUs already installed |
| Near-term capacity target | More than 80 GPUs |
| Near-term timeline | Current quarter |
| Designed future capacity | Up to 224 GPUs |
| Cooling technology | Direct-to-Chip Liquid Cooling |
| Initial deployment location | NSE’s on-premises data centre |
| Potential next location | Co-location data centre |
The distinction between planned GPU capacity and installed capacity is important. NSE has announced its intention to exceed 80 GPUs during the current quarter, while 224 GPUs represents the infrastructure’s designed expansion capability rather than a statement that all 224 GPUs are already operational.
The larger architecture gives the exchange room to increase computing resources as its AI requirements grow.
Why Does a Stock Exchange Need AI Computing Power?
Stock exchanges operate complex technology systems that process enormous volumes of financial information.
Their infrastructure must support trading operations, market surveillance, risk management, regulatory compliance and data services. These activities generate large amounts of information that can require advanced computing resources to analyse.
Artificial intelligence can support several applications across financial-market operations.
1. Market surveillance
AI systems can help analyse trading patterns and identify unusual activity that may require further investigation.
By processing large datasets, these systems can assist surveillance teams in detecting potential irregularities and examining market behaviour.
2. Risk management
Financial markets involve multiple forms of risk, including price volatility, liquidity risks and operational disruptions.
AI-powered analytical systems can help process historical and real-time information to support risk monitoring and decision-making.
3. Data analysis
Exchanges handle extensive datasets covering trading activity, securities, market participants and financial instruments.
AI computing infrastructure can support advanced analytical applications that process this information more efficiently.
4. Operational efficiency
AI can also be used to automate selected internal processes, assist technology teams and improve the analysis of operational information.
The exact AI applications NSE intends to run on its expanded GPU infrastructure have not been fully disclosed. Its announcement primarily describes the deployment as preparation for growing AI workloads across the organisation.
Why NSE Is Keeping Its AI Infrastructure On-Premises
Another important aspect of the deployment is its location.
NSE has installed the new infrastructure within its own on-premises data centre rather than announcing a fully cloud-hosted arrangement.
The exchange says this approach is intended to strengthen data sovereignty while retaining control over AI data, computing resources and access.
Data sovereignty refers to the ability to maintain control over how data is stored, accessed and governed under applicable laws and regulatory requirements.
For a major financial-market institution, these considerations are particularly important because its technology systems handle sensitive operational and market-related information.
Keeping infrastructure on-premises can provide greater direct control over hardware, access management and security arrangements. However, it also means the organisation must manage the associated infrastructure, maintenance, power and cooling requirements.
NSE says its approach is designed to support its security governance framework and alignment with applicable regulatory and compliance requirements as AI adoption expands.
NSE May Extend Liquid Cooling to Its Co-Location Data Centre
The exchange has also indicated that it plans to extend Direct-to-Chip Liquid Cooling technology to its co-location data centre in the near future.
Co-location facilities allow market participants to place their trading servers close to an exchange’s trading infrastructure.
The physical proximity can help reduce network latency, which is particularly important for participants operating electronic trading systems that depend on fast communication.
NSE’s proposed expansion will depend on the requirements of market participants.
Extending liquid cooling to this environment could become relevant as computing density increases and participants seek infrastructure capable of supporting more demanding workloads.
However, the exchange has not announced a specific implementation date or detailed capacity target for the co-location deployment.
What NSE Officials Said About the Deployment
Viral Mody, Executive Director of NSE, said the exchange’s technology investments are guided by its objective of providing advanced technology to Indian market participants and developing resilient, scalable market infrastructure.
He described the deployment as part of NSE’s effort to improve operational efficiency, support evolving computing requirements and promote responsible resource utilisation.
Sampath Manickam, Chief Technology Officer of NSE, highlighted the growing demands that AI places on computing density, electricity and cooling infrastructure.
He said the DCLC deployment provides a scalable and energy-efficient foundation for high-density GPU computing and prepares the exchange for future AI workloads.
The statements underline the exchange’s intention to build computing infrastructure ahead of its anticipated technology requirements rather than waiting for demand to increase.
What Does NSE’s Move Mean for India’s AI Infrastructure?
NSE’s deployment reflects a wider shift in the infrastructure supporting artificial intelligence.
As organisations move beyond basic AI applications towards more advanced computing workloads, the need for GPUs, electricity, cooling systems and specialised data centres is increasing.
This is creating demand for technologies that can support higher computing density while managing heat and power consumption.
Financial institutions are among the organisations exploring these capabilities because they process large datasets and operate technology systems where reliability, security and performance are essential.
NSE’s decision is also significant because it brings advanced liquid-cooling infrastructure into India’s capital-market technology ecosystem.
The exchange’s plans could encourage further attention towards specialised cooling solutions among financial institutions and data-centre operators.
However, the broader impact will depend on how quickly organisations adopt similar systems and whether the technology delivers measurable improvements in operating efficiency and computing capacity.
The Bigger Picture
Artificial intelligence is changing the infrastructure requirements of businesses across industries. While much of the attention focuses on AI models and software applications, the computing systems behind them are becoming equally important. GPUs generate substantial heat during intensive workloads, making cooling a critical part of data-centre design.
NSE’s deployment demonstrates how infrastructure planning is evolving as financial institutions prepare for more demanding AI applications. Its plan to move beyond 80 GPUs, with a cooling architecture designed for up to 224, shows an effort to build capacity for future requirements. The next important measure will be how effectively the exchange uses this infrastructure and whether it can expand computing capabilities while maintaining reliability, security and operational efficiency.
Frequently Asked Questions
1. What is Direct-to-Chip Liquid Cooling?
Direct-to-Chip Liquid Cooling is a technology that uses circulating liquid coolant and specialised cold plates to remove heat directly from processors such as GPUs and CPUs.
2. Why has NSE deployed liquid cooling technology?
NSE has deployed the technology to support high-density GPU computing and its expanding AI workloads. It is designed to provide scalable cooling as the exchange increases its computing capacity.
3. How many GPUs does NSE plan to deploy?
NSE plans to expand its GPU capacity beyond 80 during the current quarter. Its cooling infrastructure has been designed to accommodate up to 224 GPUs in the future.
4. Is NSE’s 224-GPU infrastructure already operational?
No. The figure of 224 represents the infrastructure’s designed future capacity. NSE has said it plans to expand beyond 80 GPUs during the current quarter.
5. What is the difference between air cooling and direct-to-chip liquid cooling?
Air cooling uses fans and conditioned air to remove heat from servers. Direct-to-chip liquid cooling uses coolant circulating through cold plates attached to processors, collecting heat closer to its source.
6. Will NSE introduce liquid cooling at its co-location data centre?
NSE has said it plans to extend the technology to its co-location data centre in the near future, depending on the requirements of market participants. It has not announced a specific deployment date.
7. Does liquid cooling improve AI computing performance?
Liquid cooling can help maintain suitable operating temperatures and support higher-density computing systems. The actual performance and energy benefits depend on the infrastructure design and workload.
Looking Ahead
NSE’s next steps will involve expanding GPU capacity beyond 80 units during the current quarter and using its new cooling infrastructure to support increasingly demanding AI workloads. Its longer-term plan to accommodate up to 224 GPUs provides additional room for expansion, while the proposed co-location deployment could extend the technology to market participants.
The development also highlights an important change in India’s financial technology infrastructure. As exchanges and financial institutions adopt more AI-powered systems, investment in computing hardware, electricity management and advanced cooling will become increasingly important. NSE’s deployment is an early example of how India’s capital-market infrastructure is preparing for that shift.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
