Asia-Pacific Data Center Market Analysis by Mordor Intelligence
The Asia-Pacific data center market size is expected to grow from USD 35.77 billion in 2026 to USD 43.41 billion in 2027 and is forecast to reach USD 116.4 billion by 2032 at 21.8% CAGR over 2027-2032.
AI workloads and standalone 5G core networks expand the computing requirements that data-center operators must accommodate. In Japan, grid-connection delays favor suppliers able to deliver energized capacity rather than an undeveloped site. Colocation operators can respond with equipped capacity, while dedicated-campus developers must coordinate new facility delivery with computing requirements. Malaysia's conditional incentives also connect efficient design with potential tax benefits. Grid connection timing decides when Japanese capacity can enter service, while Malaysian green-technology conditions help decide whether tax relief is available at all.
Key Report Takeaways
- By data center size, large facilities held a 30.4% Asia-Pacific data center market share in 2026, while medium facilities are forecast to grow fastest at 23.3% CAGR over 2027-2032.
- Across tier types, Tier 3 held a 63.12% share of the Asia-Pacific data center market size in 2026 and is forecast to grow fastest at 23.4% CAGR over 2027-2032.
- Among data center types, colocation led with a 70.36% share in 2026, whereas hyperscale/self-built facilities are forecast to grow fastest at 24.3% CAGR over 2027-2032.
- Within end users, IT and ITES accounted for the largest share at 27.05% in 2026, with BFSI forecast to grow fastest at 22.8% CAGR over 2027-2032.
- Geographically, China led with a 34.32% share in 2026, while India is forecast to grow fastest at 25.35% CAGR over 2027-2032.
Note: Market numbers in this report are based on Mordor Intelligence's proprietary estimation framework, which combines nine months of actuals with Q4 projections for base year 2026.
600+
data center reports by Mordor Intelligence
covering AI facilities, power, construction and more - global, regional or country-level, right down to individual data center components
Market Trends and Insights
Drivers Impact Analysis of Asia-Pacific Data Center Market*
| Driver | (~) % Impact on CAGR Forecast | Geographic Relevance | Impact Timeline |
|---|---|---|---|
| Accelerating roll-out of 5G core networks | +5% | India and Malaysia | Medium term (2-4 years) |
| Rising hyperscale self-build investments by Chinese tech majors | +7.5% | Datong in China | Medium term (2-4 years) |
| Government tax-holiday incentives for green data centers | +3.5% | Eligible locations in Malaysia | Medium term (2-4 years) |
| Surging AI/ML workload demand for on-prem GPU clusters | +8% | Ansan in South Korea | Short term (≤ 2 years) |
| Rapid integration of submarine-cable landing stations with edge facilities | +2% | Singapore's Bifrost landing site | Short term (≤ 2 years) |
| Build-to-suit models preferred by sovereign wealth funds | +1.5% | Asia-Pacific development locations | Long term (≥ 4 years) |
| Source: Mordor Intelligence | |||
Accelerating roll-out of 5G core networks
Standalone 5G rollouts create a facility opportunity through the computing infrastructure that supports mobile-network functions. GSMA Intelligence counted 17 standalone networks among 39 commercial 5G networks across 9 Asia-Pacific countries as of June 2025.[1] Reliance Jio Infocomm confirmed commercial standalone 5G operation in India as of June 30, 2026. Malaysia's YTL Communications (Yes) launched a standalone service during the fourth quarter of 2025. NTT DOCOMO launched commercial 5G Core services in Japan in February 2026 using AWS alongside its on-premises network function virtualization infrastructure. Hybrid core deployment makes cloud hosting one part of the infrastructure opportunity, with the division of functions determining how much demand reaches hosted facilities.
Rising hyperscale self-build investments by Chinese tech majors
Computing centers carrying ByteDance and JD.com branding show platform expansion turning into construction activity in Datong. Datong reported accelerated construction of the ByteDance-branded Volcano Cloud computing center Phase II during its 2025 plan-execution period. JD.com's intelligent computing center phases II and III were also under accelerated construction during that period. The same report said an upstream server-manufacturing project began construction and entered production in 2025. Concurrent center phases alongside server production connect the city's computing expansion with facilities and equipment supply, broadening the local development program beyond a single platform-branded site.
Government tax-holiday incentives for green data centers
Conditional tax relief can improve the investment case for efficient data centers by linking fiscal benefits to qualifying infrastructure. Malaysia's Digital Ecosystem Acceleration (DESAC) tax-incentive scheme offers eligible data-center companies income-tax exemptions or special tax rates, subject to consideration by the Malaysian Investment Development Authority (MIDA).[2] Green-technology adoption aligned with sustainable-data-center guidelines is among its tier conditions, including renewable-energy generation or investment in energy-efficient equipment. At Tier 1, allowances may equal 100% of qualifying capital investment excluding land for new companies for 5 or 10 years, versus 60% for existing companies for 5 years. These distinctions make the facility's technical design and the applicant's status relevant to its potential tax treatment.
Surging AI/ML workload demand for on-prem GPU clusters
On-prem GPU clusters create demand for facilities that can accommodate both computing hardware and its supporting infrastructure. By December 29, 2025, Kakao had installed 255 nodes containing 2,040 B200 GPUs at Kakao Data Center Ansan in South Korea. This represented about 84% of its 2,424-GPU allocation under a government-led project. The facility combined stable power systems with hot-aisle containment that routed hot air directly to cooling units for high-density servers. For operators, the deployment connects institutional computing allocations to a requirement for suitable electrical and thermal infrastructure. Kakao planned to operate the full allocation for 5 years, giving the Ansan electrical and cooling infrastructure a multi-year service role beyond the initial hardware installation.
Restraints Impact Analysis of Asia-Pacific Data Center Market*
| Restraint | (~) % Impact on CAGR Forecast | Geographic Relevance | Impact Timeline |
|---|---|---|---|
| Prolonged grid connection approval cycles | -2.4% | Inner Tokyo in Japan | Long term (≥ 4 years) |
| Pronounced water scarcity risk in Northern China | -1.5% | Ulanqab in northern China | Medium term (2-4 years) |
| High real-estate cost inflation in tier-one Asia-Pacific cities | -1% | Japan and Australia | Short term (≤ 2 years) |
| Shortage of certified data-center engineers | -0.8% | Data-center locations in Singapore | Medium term (2-4 years) |
| Source: Mordor Intelligence | |||
Prolonged grid connection approval cycles
A development site cannot become usable data-center supply until its power connection supports operation. Japan's Agency for Natural Resources and Energy reported in October 2026 that connection periods for large loads, including data centers, had lengthened in some areas as grid headroom tightened.[3] In 2025, NTT Global Data Centers' Japan and APAC executive estimated that developers in concentrated areas such as inner Tokyo might wait 5 to 10 years for power. Inner Tokyo illustrates the localized delivery risk, since the reported waits concern constrained areas rather than every Japanese site.
Pronounced water scarcity risk in Northern China
Ulanqab's ban on groundwater as data-center cooling water narrows the feasible cooling designs there. Ulanqab prohibits data-center projects from using groundwater as cooling water under the Inner Mongolia Autonomous Region Groundwater Protection and Management Regulations. For a developer considering water-assisted cooling, that restriction makes access to an alternative permitted source part of the project's operating requirements. In 2025, 21Vianet's Ulanqab Cloud Computing Center used reclaimed water for cooling sprays only during the hottest days of summer. Peak summer cooling must therefore work without groundwater in Ulanqab, keeping alternative-source integration part of the design even when sprays run only briefly.
*Our forecasts treat driver/restraint impacts as directional, not additive. The impact forecasts reflect baseline growth, mix effects, and variable interactions.
Asia-Pacific Data Center Market Segment Analysis
By Data Center Size:
Facility Scale Shapes Delivery CommitmentsLarge facilities accounted for a 30.4% share of the Asia-Pacific data center market size in 2026. SoftBank Corp. and its subsidiary IDC Frontier Inc. broke ground on a large-scale AI data center in Tomakomai, Japan, in April 2025. The expected opening scale was 50 MW for FY2026, and SoftBank's September 2026 transition plan retained that opening schedule. SoftBank's AI computing infrastructure and external users were the planned customers of the facility.
Medium facilities are the fastest-growing segment, forecast at 23.3% CAGR over 2027-2032. In July 2025, Yunnan's state-assets authority reported that the Yunnan Communications Intelligent Computing Center, built by Yunnan Communications Investment Group, met the national medium data-center standard. The example links a medium-class facility with an intelligent-computing application, showing how a specialized workload can fit a project outside the largest size categories. At 100P of intelligent computing capacity, representing 10% of Yunnan's provincial intelligent computing capacity, this medium-standard center demonstrates meaningful regional scale within one facility.
By Tier Type:
Deployment Scale Meets Resilience RequirementsTier 3 was the largest segment, with a 63.12% share of the regional market in 2026, and is the fastest-growing segment at a forecast 23.4% CAGR over 2027-2032. Yotta's national data center in Guwahati, India, was operating with 200 racks and a 2 MW IT load in February 2026. Yotta constructed and operated the Tier III facility on behalf of India's National Informatics Centre. Built-in scalability to double capacity gives the National Informatics Centre an expansion route within the commissioned site, linking further government computing requirements to an existing facility.
The Tier 1 and 2 segment spans basic capacity in Tier 1 and redundant power and cooling components in Tier 2. Tier 1 requires preventive-maintenance shutdowns, while Tier 2 permits component removal without a shutdown but can be affected by unexpected shutdowns. Tier 4 uses independent, physically isolated redundant systems to protect IT operations against equipment or distribution-path failures. Unexpected shutdowns affect Tier 1 and 2 facilities, whereas Tier 4's protection becomes more vulnerable when a failure coincides with maintenance shutdown of its redundant systems.
By Data Center Type:
Equipped Capacity and Dedicated DevelopmentColocation represented a 70.36% Asia-Pacific data center market share in 2026. APAC colocation supply was estimated at 11.9 GW in 2025E, measured as live IT capacity fitted with mechanical and electrical systems for operation. Because the 2025E estimate counts live IT capacity already fitted for operation, it measures an equipped supply base rather than projects still in development.
Hyperscale/self-built facilities are the fastest-growing segment, forecast at 24.3% CAGR over 2027-2032. Google broke ground on its gigawatt-scale Visakhapatnam AI hub in April 2026 with development partners AdaniConneX and Nxtra by Airtel. The planned hub comprises 3 data center campuses, and those partners lead construction for Google's planned AI capacity. Mint reported in October 2026 that the Adani Infra-owned project vehicle could continue work at its own risk pending litigation, so delivery timing carries legal risk.
By End User:
Platform Expansion Meets Institutional CommitmentsIT and ITES was the largest segment, with a 27.05% share of the regional market in 2026. In 2026, Chinese technology firms were driving Hong Kong data-center demand, with Alibaba Group Holding and Tencent Holdings increasing capacity in the city. Demand from these firms helped push Hong Kong lease prices up by nearly double in 2026, giving operators with available capacity pricing leverage.
BFSI is the fastest-growing segment, forecast at 22.8% CAGR over 2027-2032. Indian Overseas Bank selected Sify Infinit Spaces Limited in July 2025 for colocation services at a bank site near its disaster-recovery data center. The selection specified a 10-year co-hosting model.
Geography Analysis
China Asia-Pacific Data Center Market
China held a 34.32% Asia-Pacific data center market share in 2026. Its nationwide computing infrastructure included 15.56 million in-use standard racks in June 2026. Construction of a wind-powered underwater demonstration data center in Shanghai with 24 MW of total power capacity was completed in October 2025. Beijing Highlander reported stable commercial operation at the project in H1 2026 and first-batch computing-resource delivery to customers including China Telecom and Shanghai INESA. Lingang's customer deliveries establish a commercial outlet for this demonstration design within China's large computing base, linking an unconventional power configuration to actual service provision.
India Asia-Pacific Data Center Market
India is the fastest-growing geography, forecast at 25.35% CAGR over 2027-2032. India's government reported that national installed data-center power capacity increased from around 375 MW in 2020 to 1.57 GW by August 2026. In September 2026, Microsoft identified HDFC Bank as signed up for its live India South Central region in Hyderabad, connecting newly available infrastructure with a named financial-services buyer.
Local cloud access expanded with SAKURA internet's June 2025 container-based GPU-cloud launch in Ishikari, Japan, and AWS's September 2025 New Zealand region launch. Microsoft's Indonesia region opened in May 2025 with in-country data residency, while FPT Corporation's Fornix HCM02 opened in Ho Chi Minh City, Vietnam, in August 2025 for AI and cloud services. Singapore primarily allocates capacity through competitive calls assessing power and water efficiency, while Malaysia's Bridge Data Centres uses reclaimed cooling water from a Johor plant opened in August 2025. Australia's energy adviser recommended preliminary data-center obligations to ministers in July 2026, needing further work and consultation, while Thailand approved TikTok System (Thailand)'s planned storage-processing expansion across Bangkok, Samut Prakan and Chachoengsao in May 2026. FPT Corporation operates its own AI and cloud data center in Vietnam, a different channel from the Microsoft and AWS cloud regions in Indonesia and New Zealand.
Regulatory Framework and Outlook
The comparison assesses China and India as selected jurisdictions from the perspective of a data-center developer that also provides customer-data services. China's Ministry of Industry and Information Technology (MIIT) sets contract and data-safeguard duties for internet data center (IDC) operators. Compliance exposure therefore depends on how an operator handles customer information as well as how it delivers the physical facility.
Regulatory Table of Asia-Pacific Data Center Market
| Decision factor | China | India |
|---|---|---|
| Primary authority | Ministry of Industry and Information Technology (MIIT) | Indian Computer Emergency Response Team (CERT-In) |
| Key instruments or direction | MIIT 2025 IDC customer-data security notice | CERT-In Directions, 2022; phased Digital Personal Data Protection (DPDP) Rules, 2025 |
| Commercial effect | Contractual security duties; customer notice and authorization before high-risk operations affecting customer-data security or external customer-data provision | Specified incident reporting within 6 hours; ICT logs for 180 days within India; specified customer records for at least 5 years after registration cancellation or withdrawal |
| Current policy attractiveness | Medium | Medium |
| Future direction of policy attractiveness | No change | Hardening |
| Source: Mordor Intelligence | ||
China's defined contractual-security duties support a workable compliance framework, while planned standards and capability evaluation extend the same customer-protection functions, supporting continuity in core service requirements. India's reporting, logging and customer-record duties create continuing operational work, which moderates its attractiveness for hosting providers. Rules 3, 5-16 and 22-23 of the DPDP Rules come into force in May 2027. This increases the commercial relevance of valid processor contracts for hosting providers handling personal data on a Data Fiduciary's behalf, while the Act retains responsibility for covered processing with the fiduciary.
Value Chain Stages
The value chain compares power supply, cooling systems, connectivity and facility operation through the different services each supplier contributes. LG Electronics was announced in August 2025 as the supplier of cooling equipment and a building-management system for an AI data center then under construction in Jakarta. Network infrastructure follows a separate delivery path, illustrated by KDDI and the Southeast Asia-Japan Cable 2 (SJC2) consortium starting operations in July 2025 after NEC completed construction. Supplier earnings can arise at different stages of provision, creating a distinction between equipment or network delivery and the recurring services built on that infrastructure.
Value Chain Analysis Table of Asia-Pacific Data Center Market
| Primary activity | Key participants | Value capture | Bottleneck | Resilience response |
|---|---|---|---|---|
| Power procurement and supply | Electricity suppliers such as Sembcorp Power | Electricity sales under supply arrangements | Matching power availability with facility commissioning | Align capacity phases with contracted power availability |
| Cooling and controls supply | Cooling-equipment and control-system suppliers such as LG Electronics | Equipment sales and systems supply | Coordinating cooling equipment with facility controls | Specify equipment and control interfaces together |
| Connectivity commissioning | Cable-system participants such as KDDI and cable constructors such as NEC | Connectivity services and cable-system supply | Dependence on completed routes and landing infrastructure | Qualify alternative network paths before service launch |
| Facility operation and hosting | Data-center developers and operators such as STACK Infrastructure and DayOne | Hosting and capacity-service fees | Commissioning dependencies across power, cooling and connectivity | Complete staged system acceptance before customer service begins |
| Source: Mordor Intelligence | ||||
In July 2025, DayOne signed a power purchase agreement with Sembcorp Power for a planned Singapore facility whose operations were to be fully covered by renewable electricity. The agreement was signed at groundbreaking for a 20 MW facility whose first phase was scheduled for service in 2026. Signing the agreement at groundbreaking makes electricity procurement a commercial commitment during construction, extending the energy supplier's contractual role into development before the operator can offer hosting services.
Competitive Landscape
Market Concentration
Competition is distributed across cloud platforms and independent facility operators, giving customers access to computing infrastructure through either cloud-service provision or independent hosting. Amazon Web Services, Inc. supplies cloud infrastructure, while Equinix, Inc. operates data centers and interconnection services. Huawei Cloud Computing Technologies Co., Ltd uses local cloud infrastructure to serve APAC carriers, financial organizations and internet companies. In August 2026, Digital Realty Trust Inc. operated 3 data centers in Singapore with approximately 84 MW of combined capacity. A facility entrant cannot replicate Huawei Cloud's service channel to carriers and financial organizations through hosting capacity alone, because those buyers receive cloud services rather than only facility space.
Beijing VNET Broadband Data Center Co., Ltd was a principal consolidated affiliated entity of VNET Group as of December 31, 2025. VNET Group's wholesale data center capacity in service in China rose 49.4% year over year to 1,007 MW as of June 30, 2026. VNET Group secured a 345 MW wholesale order from a cloud service provider in the second quarter of 2026. That customer commitment shows how a cloud platform can be a buyer of independent facility capacity as well as a competitor for computing-infrastructure customers. Concentrating delivery obligations in one customer relationship makes that cloud provider's requirements commercially significant even within VNET's expanding wholesale base.
Competitive expansion also depends on access to development opportunities where capacity allocation is managed. In August 2026, Digital Realty was selected for a provisional 50 MW allocation to develop a new data center on Jurong Island, Singapore. The planned facility is at Jurong Town Corporation's low-carbon data-center park, with selection made through Singapore's second Data Center Call for Application. Challengers seeking capacity at this park must compete for an allocation as well as for customers. Selection therefore gives Digital Realty a position in the park's development pipeline before it can offer the allocated capacity as an operating service.
Asia-Pacific Data Center Industry Leaders
-
Amazon Web Services, Inc
-
Equinix, Inc.
-
Huawei Cloud Computing Technologies Co., Ltd
-
Digital Realty Trust Inc.
-
Beijing VNET Broadband Data Center Co., Ltd
- *Disclaimer: Major Players sorted in no particular order
Asia-Pacific Data Center Market Companies Covered in this Report
- Amazon Web Services, Inc
- Equinix, Inc.
- Alibaba Cloud
- Chindata Group Holdings Ltd
- AirTrunk Operating Pty Ltd
- Space DC Pte Ltd
- NTT DATA, Inc.
- Huawei Cloud Computing Technologies Co., Ltd
- Global Data Solutions Co., Ltd. (GDS)
- Google LLC
- LG Uplus Corp
- Range Technology Development Co., Ltd
- NEXTDC Limited
- Digital Realty Trust Inc.
- Beijing VNET Broadband Data Center Co., Ltd
Recent Industry Developments in Asia-Pacific Data Center Market
- Sep 2026: Microsoft said its India South Central datacenter region in Hyderabad was live, with HDFC Bank among the organizations that had signed up.
- Aug 2026: Digital Realty announced its selection under Singapore's second data-center call for applications, with a provisional 50 MW allocation for a new facility at Jurong Town Corporation's low-carbon data-center park on Jurong Island.
- Apr 2026: AirTrunk announced its acquisition of Lumina CloudInfra, an India-based data-center developer with operations in Mumbai, Chennai and Hyderabad.
Scope and Methodology
Segmentation Overview
By Data Center Size
| Large |
| Massive |
| Medium |
| Mega |
| Small |
By Tier Type
| Tier 1 and 2 |
| Tier 3 |
| Tier 4 |
By Data Center Type
| Hyperscale/Self-built | ||
| Enterprise/Edge | ||
| Colocation | Non-Utilized | |
| Utilized | Retail Colocation | |
| Wholesale Colocation | ||
By End User
| BFSI |
| IT and ITES |
| E-Commerce |
| Government |
| Manufacturing |
| Media and Entertainment |
| Telecom |
| Other End Users |
By Geography
| China |
| Japan |
| India |
| Indonesia |
| Australia |
| Singapore |
| New Zealand |
| Malaysia |
| Thailand |
| Rest of Asia Pacific |
| By Data Center Size | Large | ||
| Massive | |||
| Medium | |||
| Mega | |||
| Small | |||
| By Tier Type | Tier 1 and 2 | ||
| Tier 3 | |||
| Tier 4 | |||
| By Data Center Type | Hyperscale/Self-built | ||
| Enterprise/Edge | |||
| Colocation | Non-Utilized | ||
| Utilized | Retail Colocation | ||
| Wholesale Colocation | |||
| By End User | BFSI | ||
| IT and ITES | |||
| E-Commerce | |||
| Government | |||
| Manufacturing | |||
| Media and Entertainment | |||
| Telecom | |||
| Other End Users | |||
| By Geography | China | ||
| Japan | |||
| India | |||
| Indonesia | |||
| Australia | |||
| Singapore | |||
| New Zealand | |||
| Malaysia | |||
| Thailand | |||
| Rest of Asia Pacific | |||
Market Definition
- IT LOAD CAPACITY - The IT load capacity or installed capacity, refers to the amount of energy consumed by servers and network equipments placed in a rack installed. It is measured in megawatt (MW).
- ABSORPTION RATE - It denotes the extend to which the data center capacity has been leased out. For instance, a 100 MW DC has leased out 75 MW, then absorption rate would be 75%. It is also referred as utilization rate and leased-out capacity.
- RAISED FLOOR SPACE - It is an elevated space build over the floor. This gap between the original floor and the elevated floor is used to accommodate wiring, cooling, and other data center equipment. This arrangement assist in having proper wiring and cooling infrastructure. It is measured in square feet (ft^2).
- DATA CENTER SIZE - Data Center Size is segmented based on the raised floor space allocated to the data center facilities. Mega DC - # of Racks must be more than 9000 or RFS (raised floor space) must be more than 225001 Sq. ft; Massive DC - # of Racks must be in between 9000 and 3001 or RFS must be in between 225000 Sq. ft and 75001 Sq. ft; Large DC - # of Racks must be in between 3000 and 801 or RFS must be in between 75000 Sq. ft and 20001 Sq. ft; Medium DC # of Racks must be in between 800 and 201 or RFS must be in between 20000 Sq. ft and 5001 Sq. ft; Small DC - # of Racks must be less than 200 or RFS must be less than 5000 Sq. ft.
- TIER TYPE - According to Uptime Institute the data centers are classified into four tiers based on the proficiencies of redundant equipment of the data center infrastructure. In this segment the data center are segmented as Tier 1,Tier 2, Tier 3 and Tier 4.
- COLOCATION TYPE - The segment is segregated into 3 categories namely Retail, Wholesale and Hyperscale Colocation service. The categorization is done based on the amount of IT load leased out to potential customers. Retail colocation service has leased capacity less than 250 kW; Wholesale colocation services has leased capacity between 251 kW and 4 MW and Hyperscale colocation services has leased capacity more than 4 MW.
- END CONSUMERS - The Data Center Market operates on a B2B basis. BFSI, Government, Cloud Operators, Media and Entertainment, E-Commerce, Telecom and Manufacturing are the major end-consumers in the market studied. The scope only includes colocation service operators catering to the increasing digitalization of the end-user industries.
| Keyword | Definition |
|---|---|
| Rack Unit | Generally referred as U or RU, it is the unit of measurement for the server unit housed in the racks in the data center. 1U is equal to 1.75 inches. |
| Rack Density | It defines the amount of power consumed by the equipment and server housed in a rack. It is measured in kilowatt (kW). This factor plays a critical role in data center design and, cooling and power planning. |
| IT Load Capacity | The IT load capacity or installed capacity, refers to the amount of energy consumed by servers and network equipment placed in a rack installed. It is measured in megawatt (MW). |
| Absorption Rate | It denotes how much of the data center capacity has been leased out. For instance, if a 100 MW DC has leased out 75 MW, then the absorption rate would be 75%. It is also referred to as utilization rate and leased-out capacity. |
| Raised Floor Space | It is an elevated space built over the floor. This gap between the original floor and the elevated floor is used to accommodate wiring, cooling, and other data center equipment. This arrangement assists in having proper wiring and cooling infrastructure. It is measured in square feet/meter. |
| Computer Room Air Conditioner (CRAC) | It is a device used to monitor and maintain the temperature, air circulation, and humidity inside the server room in the data center. |
| Aisle | It is the open space between the rows of racks. This open space is critical for maintaining the optimal temperature (20-25 °C) in the server room. There are primarily two aisles inside the server room, a hot aisle and a cold aisle. |
| Cold Aisle | It is the aisle wherein the front of the rack faces the aisle. Here, chilled air is directed into the aisle so that it can enter the front of the racks and maintain the temperature. |
| Hot Aisle | It is the aisle where the back of the racks faces the aisle. Here, the heat dissipated from the equipment’s in the rack is directed to the outlet vent of the CRAC. |
| Critical Load | It includes the servers and other computer equipment whose uptime is critical for data center operation. |
| Power Usage Effectiveness (PUE) | It is a metric which defines the efficiency of a data center. It is calculated by: (𝑇𝑜𝑡𝑎𝑙 𝐷𝑎𝑡𝑎 𝐶𝑒𝑛𝑡𝑒𝑟 𝐸𝑛𝑒𝑟𝑔𝑦 𝐶𝑜𝑛𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛)/(𝑇𝑜𝑡𝑎𝑙 𝐼𝑇 𝐸𝑞𝑢𝑖𝑝𝑚𝑒𝑛𝑡 𝐸𝑛𝑒𝑟𝑔𝑦 𝐶𝑜𝑛𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛). Further, a data center with a PUE of 1.2-1.5 is considered highly efficient, whereas, a data center with a PUE >2 is considered highly inefficient. |
| Redundancy | It is defined as a system design wherein additional component (UPS, generators, CRAC) is added so that in case of power outage, equipment failure, the IT equipment should not be affected. |
| Uninterruptible Power Supply (UPS) | It is a device that is connected in series with the utility power supply, storing energy in batteries such that the supply from UPS is continuous to IT equipment even during utility power is snapped. The UPS primarily supports the IT equipment only. |
| Generators | Just like UPS, generators are placed in the data center to ensure an uninterrupted power supply, avoiding downtime. Data center facilities have diesel generators and commonly, 48-hour diesel is stored in the facility to prevent disruption. |
| N | It denotes the tools and equipment required for a data center to function at full load. Only "N" indicates that there is no backup to the equipment in the event of any failure. |
| N+1 | Referred to as 'Need plus one', it denotes the additional equipment setup available to avoid downtime in case of failure. A data center is considered N+1 when there is one additional unit for every 4 components. For instance, if a data center has 4 UPS systems, then for to achieve N+1, an additional UPS system would be required. |
| 2N | It refers to fully redundant design wherein two independent power distribution system is deployed. Therefore, in the event of a complete failure of one distribution system, the other system will still supply power to the data center. |
| In-Row Cooling | It is the cooling design system installed between racks in a row where it draws warm air from the hot aisle and supplies cool air to the cold aisle, thereby maintaining the temperature. |
| Tier 1 | Tier classification determines the preparedness of a data center facility to sustain data center operation. A data center is classified as Tier 1 data center when it has a non-redundant (N) power component (UPS, generators), cooling components, and power distribution system (from utility power grids). The Tier 1 data center has an uptime of 99.67% and an annual downtime of <28.8 hours. |
| Tier 2 | A data center is classified as Tier 2 data center when it has a redundant power and cooling components (N+1) and a single non-redundant distribution system. Redundant components include extra generators, UPS, chillers, heat rejection equipment, and fuel tanks. The Tier 2 data center has an uptime of 99.74% and an annual downtime of <22 hours. |
| Tier 3 | A data center having redundant power and cooling components and multiple power distribution systems is referred to as a Tier 3 data center. The facility is resistant to planned (facility maintenance) and unplanned (power outage, cooling failure) disruption. The Tier 3 data center has an uptime of 99.98% and an annual downtime of <1.6 hours. |
| Tier 4 | It is the most tolerant type of data center. A Tier 4 data center has multiple, independent redundant power and cooling components and multiple power distribution paths. All IT equipment are dual powered, making them fault tolerant in case of any disruption, thereby ensuring interrupted operation. The Tier 4 data center has an uptime of 99.74% and an annual downtime of <26.3 minutes. |
| Small Data Center | Data center that has floor space area of ≤ 5,000 Sq. ft or the number of racks that can be installed is ≤ 200 is classified as a small data center. |
| Medium Data Center | Data center which has floor space area between 5,001-20,000 Sq. ft, or the number of racks that can be installed is between 201-800, is classified as a medium data center. |
| Large Data Center | Data center which has floor space area between 20,001-75,000 Sq. ft, or the number of racks that can be installed is between 801-3,000, is classified as a large data center. |
| Massive Data Center | Data center which has floor space area between 75,001-225,000 Sq. ft, or the number of racks that can be installed is between 3001-9,000, is classified as a massive data center. |
| Mega Data Center | Data center that has a floor space area of ≥ 225,001 Sq. ft or the number of racks that can be installed is ≥ 9001 is classified as a mega data center. |
| Retail Colocation | It refers to those customers who have a capacity requirement of 250 kW or less. These services are majorly opted by small and medium enterprises (SMEs). |
| Wholesale Colocation | It refers to those customers who have a capacity requirement between 250 kW to 4 MW. These services are majorly opted by medium to large enterprises. |
| Hyperscale Colocation | It refers to those customers who have a capacity requirement greater than 4 MW. The hyperscale demand primarily originates from large-scale cloud players, IT companies, BFSI, and OTT players (like Netflix, Hulu, and HBO+). |
| Mobile Data Speed | It is the mobile internet speed a user experiences via their smartphones. This speed is primarily dependent on the carrier technology being used in the smartphone. The carrier technologies available in the market are 2G, 3G, 4G, and 5G, where 2G provides the slowest speed while 5G is the fastest. |
| Fiber Connectivity Network | It is a network of optical fiber cables deployed across the country, connecting rural and urban regions with high-speed internet connection. It is measured in kilometer (km). |
| Data Traffic per Smartphone | It is a measure of average data consumption by a smartphone user in a month. It is measured in gigabyte (GB). |
| Broadband Data Speed | It is the internet speed that is supplied over the fixed cable connection. Commonly, copper cable and optic fiber cable are used in both residential and commercial use. Here, optic cable fiber provides faster internet speed than copper cable. |
| Submarine Cable | A submarine cable is a fiber optic cable laid down at two or more landing points. Through this cable, communication and internet connectivity between countries across the globe is established. These cables can transmit 100-200 terabits per second (Tbps) from one point to another. |
| Carbon Footprint | It is the measure of carbon dioxide generated during the regular operation of a data center. Since, coal, and oil & gas are the primary source of power generation, consumption of this power contributes to carbon emissions. Data center operators are incorporating renewable energy sources to curb the carbon footprint emerging in their facilities. |
Research Methodology
Mordor Intelligence follows a four-step methodology in all our reports.
- Step-1: Identify Key Variables: In order to build a robust forecasting methodology, the variables and factors identified in Step-1 are tested against available historical market numbers. Through an iterative process, the variables required for market forecast are set and the model is built on the basis of these variables.
- Step-2: Build a Market Model: Market-size estimations for the forecast years are in nominal terms. Inflation is not a part of the pricing, and the average selling price (ASP) is kept constant throughout the forecast period for each country.
- Step-3: Validate and Finalize: In this important step, all market numbers, variables and analyst calls are validated through an extensive network of primary research experts from the market studied. The respondents are selected across levels and functions to generate a holistic picture of the market studied.
- Step-4: Research Outputs: Syndicated Reports, Custom Consulting Assignments, Databases & Subscription Platforms
