For years, data center design was largely driven by availability, power capacity, cooling efficiency, security, and scalability.
Those priorities remain critical. But another factor is becoming increasingly important: Low Latency.
As AI, 5G, industrial automation, connected devices, and real-time digital services expand, the physical distance between where data is generated and where it is processed can directly influence application performance.
That is changing the infrastructure equation.
India’s data center market is already expanding rapidly. India’s live data center capacity reached around 1.7 GW in 2026, with another 1.3 GW under construction and 3.2 GW of projects reported as ready to break ground. Maharashtra alone accounts for about 55% of live capacity, according to BloombergNEF data reported by Express Computer.
At the same time, AI is creating demand for increasingly distributed computing environments. Recent industry reporting puts India’s colocation capacity at around 2 GW, with projections approaching 10 GW by 2031.
The question, therefore, is no longer simply how much data center capacity India needs.
It is also where that capacity should be located, how close compute should be to users and devices, and what infrastructure model can support increasingly time-sensitive workloads.
Why Response Time Matters More Than Before
Not every digital application requires the same response time.
A file backup can tolerate delay.
A video stream can buffer.
But an autonomous system, industrial control application, remote monitoring system, or real-time AI application may have much less tolerance for network delay.
The difference comes down to the nature of the workload.
| Application | Typical infrastructure requirement | Why response time matters |
| Cloud storage | Centralised cloud/data center | Delays are generally tolerable |
| Video streaming | Distributed content delivery | Faster content delivery improves experience |
| Online gaming | Distributed processing and networking | Delays can directly affect interaction |
| Smart manufacturing | Edge or local compute | Machines may require near-real-time decisions |
| Autonomous systems | Edge computing | Decisions may need to happen close to the device |
| AI inference | Distributed or edge compute | Faster processing can improve real-time responses |
| Defence applications | Edge/mobile infrastructure | Data may need to be processed where it is generated |
| 5G applications | Edge + telecom infrastructure | Network architecture increasingly supports distributed workloads |
This is why data center architecture is increasingly being evaluated alongside the application it supports.
The facility is no longer an isolated IT environment.
It is becoming part of the application’s performance architecture.
The Physics Behind Network Delay
There is a simple physical limitation behind the discussion.
Data has to travel.
When an application sends information to a distant centralized facility, the request travels through networks, routing equipment, and other infrastructure before processing occurs and the response returns.
The longer the path, the greater the opportunity for delay.
Reducing physical distance does not eliminate every source of latency. Network congestion, routing, processing, protocol overhead, and application architecture also matter.
But proximity can become an important design variable.
This is one reason edge computing has gained attention.
Instead of sending every workload back to a large centralized data center, certain workloads can be processed closer to the point where data is generated or consumed.
Centralised vs Edge Architecture
| Factor | Centralised Data Center | Edge Data Center |
| Location | Major data center hubs | Closer to users or data sources |
| Primary objective | Large-scale compute and storage | Localised processing |
| Deployment model | Large campuses/facilities | Distributed facilities or modular units |
| Typical workloads | Cloud, storage, large-scale compute | AI inference, IoT, 5G, industrial and real-time applications |
| Geographic footprint | Fewer, larger locations | Multiple distributed locations |
| Scalability | Large-scale capacity expansion | Incremental deployment closer to demand |
| Infrastructure challenge | Scale, power, cooling and resilience | Compact design, reliability, remote operation and maintainability |
The two models are not competitors in every scenario.
In many architectures, they work together.
Large data centers can handle intensive centralised workloads, while edge facilities can process selected data closer to the source.
AI Is Changing the Data Center Equation
Artificial intelligence is one of the strongest forces reshaping data center infrastructure.
India’s AI infrastructure investment is accelerating. In September 2026, a TCS subsidiary and partners announced plans for a 1 GW AI data center campus in Telangana with an investment of up to ₹700 billion.
Large-scale AI training will continue to require substantial centralised compute.
But AI workloads do not stop at training.
Inference is where models interact with applications, users, machines and devices.
That creates a different infrastructure question.
Does every AI request need to travel back to a distant hyperscale facility?
For some workloads, the answer may be no.
Consider an industrial vision system inspecting products on a manufacturing line.
A connected vehicle interpreting its surroundings.
A surveillance system identifying an event in real time.
A defence application processing information in a field environment.
In such cases, sending every piece of data to a distant facility before a decision is made may not be the most effective architecture.
Edge computing can allow selected workloads to be processed closer to where they originate.
This does not mean every AI workload should move to the edge.
It means data center architecture increasingly needs to distinguish between workloads that benefit from centralised scale and workloads that benefit from proximity.
5G Is Another Piece of the Puzzle
The growth of 5G adds another dimension.
5G enables higher bandwidth, greater device density and new classes of connected applications. But connectivity alone does not guarantee rapid application response.
The network and compute infrastructure behind the application also matter.
This is where edge infrastructure becomes relevant.
A 5G-enabled industrial environment could generate data from cameras, sensors, machines and connected equipment. Processing selected data locally can reduce the need to send every event to a distant central facility.
The same principle can apply to:
- Smart manufacturing
- Connected infrastructure
- Remote monitoring
- Video analytics
- AR/VR applications
- Autonomous systems
- Private 5G networks
- Real-time industrial analytics
DC&T’s modular data center offering is designed around several of these use cases, including private 5G, AI/IoT workloads, smart manufacturing and remote edge deployments.
Why Data Center Location Is Becoming a Design Decision
Traditionally, data center location has been heavily influenced by power availability, connectivity, land, climate, regulatory conditions, and access to infrastructure.
Those factors remain fundamental.
But workload geography is becoming another consideration.
Imagine two scenarios.
Scenario A
A manufacturing facility generates large volumes of machine data. Every data stream travels hundreds of kilometres to a central facility for processing before the relevant result is returned.
Scenario B
The facility has local compute infrastructure that processes time-sensitive data on site while sending aggregated or non-critical information to a central cloud environment.
The second architecture may reduce unnecessary network traffic and improve response times for selected workloads.
This is the basic logic behind distributed computing.
What changes when compute moves closer?
- Network architecture
Connectivity becomes part of the data center design rather than simply an external service.
- Physical infrastructure
The facility may need to operate in locations that were never designed to host conventional large-scale data centers.
- Cooling
Compact infrastructure can still involve significant heat density, particularly with AI and high-performance computing.
- Power
Edge sites need reliable power despite potentially being located outside major data center clusters.
- Security
Distributed facilities increase the number of physical locations that need to be secured and monitored.
- Remote operations
A facility located close to a factory, telecom site, transport network or remote operational environment may need extensive remote monitoring and management.
- Scalability
Infrastructure should be capable of expanding as local compute requirements increase.
This makes edge deployment a design challenge rather than simply a smaller version of a conventional data center.
The Case for Modular Data Center Infrastructure
Distributed deployments create a practical challenge.
Building every facility from the ground up using a conventional construction model may not always be the most efficient approach.
This is where prefabricated and modular data center infrastructure can become relevant.
Factory-engineered modules can allow components to be assembled and tested in a controlled environment before being transported to site.
That can reduce the amount of integration work required in the field and support faster deployment.
For distributed infrastructure, this can be particularly useful because the locations may vary significantly.
A modular system can be designed around:
- IT load
- Available footprint
- Power requirements
- Cooling architecture
- Environmental conditions
- Security requirements
- Connectivity
- Deployment timeline
- Future expansion
DC&T Global’s modular data center approach combines factory-engineered infrastructure with integrated mechanical, electrical and IT systems. The company positions these solutions for applications ranging from private 5G and AI workloads to remote edge deployments and enterprise-scale expansion.
What Should a Low-Latency Data Center Design Consider?
Proximity is only one part of the equation.
A well-designed facility needs to consider the complete infrastructure stack.
Network connectivity
The facility needs reliable connectivity to users, devices, telecom networks and central cloud or data center infrastructure.
Power architecture
Distributed sites still require dependable power infrastructure. Depending on the application, redundancy and backup systems may be necessary.
Cooling
High-density compute can generate substantial heat. The cooling strategy must reflect the actual IT load and equipment configuration.
Physical security
Edge facilities can operate in less traditional environments. Physical access control, monitoring and security therefore become important design considerations.
Fire protection
Compact infrastructure does not reduce the need for appropriate detection and suppression systems.
Remote monitoring
A distributed facility may not have a large on-site operations team. Monitoring and control systems can therefore become essential for operational continuity.
Maintainability
Maintenance access, component replacement and remote diagnostics need to be considered before deployment.
Scalability
The initial deployment should not become a limitation when compute requirements increase.
Centralised, Edge or Hybrid?
The real question for many organisations is not whether edge infrastructure will replace conventional data centers.
It is whether the workload requires a centralised, distributed or hybrid architecture.
| Requirement | Centralised | Edge | Hybrid |
| Large-scale training | Strong | Limited | Strong |
| Local real-time processing | Limited | Strong | Strong |
| Large storage requirements | Strong | Moderate | Strong |
| Remote industrial applications | Limited | Strong | Strong |
| 5G applications | Moderate | Strong | Strong |
| AI inference | Moderate | Strong | Strong |
| Geographic scalability | Strong in hubs | Strong across locations | Strong |
| Centralised management | Strong | More complex | Strong with orchestration |
For many organisations, the hybrid model may be the most practical.
Large facilities provide scale.
Edge facilities provide proximity.
Together, they create a distributed infrastructure architecture where workloads can be placed according to their performance, data and operational requirements.
Where DC&T Global Fits
DC&T Global approaches data center infrastructure from both the EPC and modular deployment sides.
Its data center EPC offering covers engineering and design, strategic procurement, civil and MEP works, white-space buildouts, infrastructure integration, testing and commissioning. The company also supports hyperscale and enterprise data center builds.
For applications where deployment speed, distributed capacity or location flexibility matter, DC&T Global also provides prefabricated modular data center solutions.
Its offering includes:
- Brick-and-mortar data center builds
- Prefabricated modular data centers
- Edge data center infrastructure
- Private 5G-enabled modular deployments
- AI and IoT-ready infrastructure
- Integrated power and cooling systems
- Fire detection and suppression
- Structured cabling and IT infrastructure
- Testing and commissioning
- Deployment for remote and mission-critical environments
The company’s modular systems are factory-engineered and tested before deployment, with configurations tailored to workload, location and scale.
This becomes particularly relevant as data center requirements move from a single large facility toward a combination of centralised and distributed infrastructure.
The design challenge is no longer simply building more capacity.
It is building the right capacity, in the right location, for the right workload.
The Bigger Shift in Data Center Design
The growth of AI, 5G, IoT and real-time digital applications is changing the role of the data center.
The industry is still building large facilities.
India’s current pipeline demonstrates that clearly.
But alongside hyperscale campuses, there is growing relevance for infrastructure that can be deployed closer to users, enterprises, industrial environments and connected devices.
That creates a more distributed digital infrastructure landscape.
In this environment, response time becomes a design consideration alongside power, cooling, resilience, security and scalability.
And the most effective data center architecture may increasingly be the one that understands where the workload needs to happen, rather than simply where the largest facility can be built.
Conclusion
Data center design is entering a more distributed phase.
Hyperscale facilities will remain essential for large-scale cloud and AI workloads. But applications such as industrial automation, private 5G, AI inference, connected infrastructure and mission-critical systems can create a different requirement: compute that is closer to where data is generated and decisions are made.
That is why Low Latency is becoming a data center design priority.
The opportunity is not to choose between centralised and edge infrastructure.
It is to understand the workload, map its infrastructure requirements, and build an architecture that places compute where it can deliver the greatest operational value.
For data center developers and infrastructure partners, that means designing beyond the building itself.
It means designing for proximity, performance, resilience and the way computing is actually being used.
FAQs
1. What is Low Latency in data center infrastructure?
Low Latency refers to reducing the time required for data to travel between a user or device, the computing infrastructure processing it, and the resulting response. Data center location, network architecture, connectivity and workload placement can all influence response time.
2. Why are edge data centers important for low-latency applications?
Edge data centers place computing resources closer to users, devices or data sources. This can reduce the distance data needs to travel and support applications that require faster responses, including industrial automation, AI inference, private 5G, IoT and real-time analytics.
3. How does DC&T Global support low-latency data center deployments?
DC&T Global delivers data center infrastructure for latency-sensitive applications through prefabricated modular data centers, edge data center solutions, power and cooling infrastructure, connectivity, fire protection, and integrated testing and commissioning. Its approach enables organizations to deploy data center capacity closer to where data is generated or consumed, supporting applications that require responsive and reliable infrastructure.
4. Are edge data centers replacing hyperscale data centers?
No. Edge and hyperscale facilities generally serve different infrastructure requirements. Hyperscale facilities provide large-scale computing and storage, while edge infrastructure supports distributed workloads closer to users or data sources. A hybrid architecture can combine both.
5. What should organisations consider when designing an edge data center?
Key considerations include IT load, connectivity, power availability, cooling, physical security, fire protection, remote monitoring, maintainability, deployment location and future scalability. The design should also consider how the edge facility connects with central cloud or data center infrastructure.
6. Why are modular data centers suitable for edge deployments?
Modular data centers can be factory-engineered, integrated and tested before deployment. This can reduce on-site construction and integration requirements while supporting faster deployment across different locations. Modular architectures can also be configured around specific IT loads, footprints and application requirements.
Source
- Express Computer — India emerges as APAC’s top data centre market outside China — BloombergNEF-based capacity and pipeline figures for India.
- Financial Express — AI power demand and data localisation behind India’s data-centre boom — Current India capacity, growth projections and AI/data-localisation demand.
- Reuters — TCS unit plans $7.4 billion AI data centre campus in Telangana — Recent 1 GW AI data-centre investment announcement.
- DC&T Global — Data Center Build and EPC Services — DC&T’s data-center EPC scope, engineering, procurement, MEP, integration, testing and commissioning.
- DC&T Global — Prefabricated Modular Data Center Solutions — Modular, edge, private 5G, AI/IoT and deployment capabilities.