Regional · Forecast 2026-2035

Asia Pacific AI Data Center Market Size, Share & Forecast to 2035

Market at a glance

The Asia Pacific AI Data Center market was worth USD 10.24 billion in 2025 and is forecast to reach USD 64.77 billion by 2035, growing at a 20.25% CAGR. Greater China accounts for 40% of the market.

$10.24BMarket size, 2025
$64.77BForecast, 2035
20.25%CAGR, 2025-2035
40%Greater China share
Market size, USD million
1,020.0 2020 10,240.0 2025 64,770.0 2035F
Share of market by region
Greater China40%
Japan and South Korea19%
Southeast Asia18%
South Asia9%
Oceania14%

Source: DC Market Insights analysis

Key players NVIDIAAmazon Web ServicesMicrosoftGoogleAlibaba Cloud +13 more

Key findings

Driver

Hyperscaler Commitments Exceed USD 30 Billion Across Four Countries

Trend

Inference Grows from USD 3.89 billion to USD 37.56 billion

Challenge

Grid Connections Gate Up to 20% of 2035 Capacity

Opportunity

Southeast Asia and South Asia Add USD 21.20 billion

What is inside the report

  1. 01Introduction
  2. 02Research Methodology
  3. 03Executive Summary
  4. 04Market Scope and Definition: What Is an Asia Pacific AI Data Center?
  5. 05Asia Pacific AI Data Center Market Drivers
  6. 06Market Trends
  7. 07Market Challenges
  8. 08Market Opportunities
  9. 09Pricing Analysis: AI Data Center Price per MW
  10. 10Standards and Regulation
  • 23chapters
  • 2020-2024Historical period
  • 2026-2035Forecast period
  • 18Companies profiled

Full table of contents

Regional insights

  • Greater China40%
  • Japan and South Korea19%
  • Southeast Asia18%
  • South Asia9%
  • Oceania14%

Segmentation

Component
Electrical InfrastructureCooling InfrastructureBuilding Shell and Fit-OutAI Network FabricOperations Services and Software
Data Center Type
Hyperscale Self-BuildColocation and Wholesale AISovereign and Government AIEnterprise and Edge AI
Workload
AI TrainingAI Inference
Region
Greater ChinaJapan and South KoreaSoutheast AsiaSouth AsiaOceania

Companies profiled

  1. NVIDIA
  2. Amazon Web Services
  3. Microsoft
  4. Google
  5. Alibaba Cloud
  6. Tencent Cloud
  7. AirTrunk
  8. ST Telemedia Global Data Centres
  9. Equinix
  10. NTT DATA
  11. Digital Realty
  12. Princeton Digital Group
  13. YTL Power International
  14. AdaniConneX
  15. SoftBank Corp
  16. KT Cloud
  17. Schneider Electric
  18. Vertiv

Recent developments

August 2026

Cushman & Wakefield reported a record 26.5 GW APAC data centre pipeline for H1 2026, up 7.1 GW in six months. (Source: Cushman & Wakefield)

July 2026

NAVER, Brookfield and NVIDIA agreed to expand the GAK Sejong AI factory from 55 MW to 200 MW. (Source: NAVER)

May 2026

CBRE reported record APAC data centre investment of USD 11.6 billion in 2025. (Source: CBRE)

April 2026

the IndiaAI Mission reported more than 38,000 GPUs onboarded on its compute portal. (Source: Digital India, Government of India)

January 2026

Taiwan's Executive Yuan approved the Ten AI Initiatives Promotion Plan to expand AI computing capacity. (Source: National Development Council)

January 2026

the Samsung SDS-led consortium set a target of more than 15,000 GPUs by 2028 for Korea's National AI Computing Center. (Source: Samsung SDS)

Full analysis

The Asia Pacific AI Data Center Market covers the facility infrastructure built for artificial intelligence (AI) workloads in Asia Pacific (APAC): electrical systems, cooling, building shell and fit-out, the AI cluster network fabric, and the operations services and software that run these halls. Graphics processing unit (GPU) servers and storage are excluded. DC Market Insights values the market at USD 10.24 billion in 2025 and forecasts USD 64.77 billion by 2035, a compound annual growth rate (CAGR) of 20.25%. The build receipt for 2025 is 870 megawatts (MW) of new AI information technology (IT) capacity × USD 11.00 million of facility and network content per MW = USD 9.57 billion of new-build spend, plus 1,494 MW of installed AI capacity × USD 450 thousand per MW of service and refresh spend = USD 672.25 million. AI Inference is the fastest workload segment at 25.45% a year as deployed models move into production, and South Asia is the fastest region at 25.69% as India adds gigawatt-scale AI campuses.

Executive Summary: How Big Is the Asia Pacific AI Data Center Market?

The market is worth USD 10.24 billion in 2025. The Asia Pacific AI Data Center Market size was valued at USD 1.02 billion in 2020, reached USD 10.24 billion in 2025, and is anticipated to reach USD 64.77 billion by 2035, at a CAGR of 20.25% during the forecast period. Between 2020 and 2025 the market grew at 58.51% a year as the AI share of new APAC capacity rose from 8% to 30%.

Report Attribute Details
Historical Period 2020-2024
Base Year 2025
Forecast Period 2026-2035
Asia Pacific AI Data Center Market Size 2025 USD 10.24 billion
Asia Pacific AI Data Center Market CAGR (2025-2035) 20.25%
Asia Pacific AI Data Center Market Size 2035 USD 64.77 billion

DC Market Insights estimates that new AI capacity in APAC rises from 870 MW in 2025 to 2,860 MW in 2030 and 3,968 MW in 2035, lifting the market to USD 40.24 billion in 2030. Inference rises from 38% to 58% of spend, and Southeast Asia and South Asia together rise from 27% to 37% of the market. The host market for all workloads is covered in the Asia Pacific Data Center Market report.

Key takeaways:

  • USD 10.24 billion in 2025, USD 64.77 billion by 2035, 20.25% CAGR.
  • Electrical Infrastructure is the largest component at 38% (USD 3.89 billion).
  • Greater China leads with 40%; South Asia grows fastest at 25.69%.
  • An AI data center hall costs about USD 11.00 million per MW to build and equip in 2025, before servers.

Asia Pacific AI data center market size 2020-2035: USD 1.02 billion in 2020, USD 10.24 billion in 2025, USD 64.77 billion by 2035, 20.25% CAGR 2025-2035

Market Scope: What Is an Asia Pacific AI Data Center?

An Asia Pacific AI data center is a facility, or a hall inside one, built for GPU and accelerator clusters at 40 kilowatts (kW) to more than 130 kW per rack; its infrastructure was a USD 10.24 billion market in 2025. Chief Telecom’s LY2 site in Taipei, for example, is designed for 10 kW to 200 kW per cabinet (Source: Chief Telecom).

Included: medium-voltage and low-voltage switchgear, transformers, uninterruptible power supplies (UPS), busways and generators (Electrical Infrastructure); chillers, coolant distribution units (CDUs), direct-to-chip and immersion cooling (Cooling Infrastructure); the shell, structural fit-out and white-space construction (Building Shell and Fit-Out); the back-end GPU interconnect and front-end fabric switches, optics and cabling (AI Network Fabric); and commissioning, operations, maintenance, data center infrastructure management (DCIM) software and refresh (Operations Services and Software). Excluded: GPU and central processing unit (CPU) servers, storage, land, electricity bills and cloud rental revenue.

How the segments fit together: each of the three dimensions (component, data center type and workload) splits the same USD 10.24 billion total in 2025 and the same USD 64.77 billion in 2035. Readers who need only the thermal chain can compare this market with the Asia Pacific Data Center Cooling Market; leased AI halls are sized in the Asia Pacific Data Center Colocation Market.

Market Drivers: What Drives the Asia Pacific AI Data Center Market?

New AI capacity of 1,176 MW in 2026 drives the market. DC Market Insights estimates that APAC adds 3,360 MW of IT capacity in 2026, of which 35% is built for AI, and each AI MW carries USD 11.30 million of facility and network content.

Hyperscaler Commitments Exceed USD 30 Billion Across Four Countries

In October 2025 Google committed about USD 15 billion over 2026-2030 to its first AI hub in India, at Visakhapatnam, developed with AdaniConneX and Airtel (Source: Google Cloud). In June 2025 Amazon announced 20 billion Australian dollars (AUD 20 billion) of data center investment in Australia from 2025 to 2029 (Source: Amazon), and in April 2024 Microsoft committed USD 2.9 billion over two years to AI and cloud infrastructure in Japan (Source: Microsoft). Amazon Web Services (AWS) and SK Group are building an AI data center in Ulsan, South Korea, for operation by 2027, within AWS’s planned USD 5.88 billion investment in Korea by 2027 (Source: SK Telecom). DC Market Insights estimates that these programmes alone support about USD 5 billion a year of AI facility spend between 2026 and 2030, roughly a third of the USD 14.36 billion market in 2026.

Sovereign GPU Programmes Add 300,000 GPUs

In October 2025 NVIDIA Corporation (NVIDIA), the Korean government, Samsung, SK Group, Hyundai and NAVER Corporation (NAVER) announced the deployment of 260,000 NVIDIA GPUs in Korea (Source: NVIDIA). By April 2026 the IndiaAI Mission had onboarded more than 38,000 GPUs on its common compute portal, under a mission outlay of Rs 10,372 crore (Source: Digital India, Government of India). At about 1.2 kW to 1.5 kW per GPU including networking, 300,000 GPUs need 360 MW to 450 MW of IT load, or USD 4.0 billion to USD 5.0 billion of facility content at USD 11.30 million per MW. Sovereign and Government AI is therefore the fastest data center type at 23.45% a year.

Electricity Demand Rose 17% in 2025

Global data centre electricity demand rose 17% in 2025 (Source: International Energy Agency (IEA)), and the APAC development pipeline reached a record 26.5 gigawatts (GW) in the first half (H1) of 2026, with about 1.4 GW of operational capacity delivered in the half (Source: Cushman & Wakefield). DC Market Insights models APAC new IT capacity rising from 2,900 MW in 2025 to 5,200 MW in 2030, and the AI share of it from 30% to 55%, which takes the market from USD 10.24 billion to USD 40.24 billion.

The three trends in the Asia Pacific AI Data Center Market are the shift of spend from training to inference (38% of 2025 spend, 58% by 2035), liquid cooling in every new AI hall, and AI capacity moving to power-rich secondary markets.

Inference Grows from USD 3.89 billion to USD 37.56 billion

AI Inference grows at 25.45% a year against 15.66% for AI Training, because deployed models need capacity close to users in Tokyo, Mumbai, Jakarta and Sydney, while training clusters concentrate on a few gigawatt campuses. Inference overtakes training in 2031 in the DC Market Insights model.

Cooling Rises from 17% to 21% of Spend

Cooling Infrastructure grows at 22.82%, from USD 1.74 billion to USD 13.60 billion. The SK Group and AWS Ulsan design uses a hybrid cooling system combining air and liquid cooling (Source: SK Telecom), the pattern behind CDUs and cold-plate loops being added on top of chillers. Detail by cooling type is in the Asia Pacific Data Center Cooling Market report.

Southeast Asia and India Take 37% by 2035

Malaysia had 1,039 MW of data center capacity under construction in H1 2026, the most of any APAC country, and the Johor pipeline reached 3,088 MW (Source: Cushman & Wakefield). Power availability increasingly determines where new capacity can be delivered (Source: CBRE Group (CBRE)), which moves AI halls away from Singapore, Hong Kong and Tokyo.

Market Challenges: What Holds the Asia Pacific AI Data Center Market Back?

Grid access and efficiency caps are the two challenges, and together they explain the USD 17.42 billion gap between the base and low 2035 cases.

Grid Connections Gate Up to 20% of 2035 Capacity

Korea’s Special Act on Activation of Distributed Energy requires a power system impact assessment for large electricity users (Source: Korea Legislation Research Institute), and Singapore releases new capacity in tranches such as the at least 300 MW in its Green Data Centre Roadmap (Source: Infocomm Media Development Authority (IMDA)). In the low case, 20% fewer MW are connected by 2035 and the market reaches USD 47.34 billion instead of USD 64.77 billion.

Efficiency Caps Raise Content per MW by USD 0.5 Million

Taiwan caps power usage effectiveness (PUE) at 1.3 for hyperscale and 1.4 for colocation sites of 5 MW or more (Source: Ministry of Economic Affairs, Taiwan), and Singapore aims for all data centers to run at PUE 1.3 or lower (Source: IMDA). Meeting these caps with AI racks requires liquid cooling and higher-efficiency electrical trains, which DC Market Insights estimates adds about USD 0.5 million per MW, part of the rise in content from USD 11.00 million in 2025 to USD 12.50 million in 2030.

Market Opportunities: Where Are the New Revenue Pools in the Asia Pacific AI Data Center Market?

Southeast Asia and South Asia add USD 21.20 billion of annual spend between 2025 and 2035, and installed-base services reach USD 13.18 billion by 2035.

Southeast Asia and South Asia Add USD 21.20 billion

Southeast Asia grows from USD 1.84 billion to USD 14.90 billion and South Asia from USD 921.80 million to USD 9.07 billion. Google’s Visakhapatnam hub, described by the Andhra Pradesh Chief Minister as India’s first truly gigawatt-scale data center (Source: Google Cloud), is the single largest project behind the South Asia line.

Installed-Base Services Reach USD 13.18 billion

Installed AI capacity rises from 1,494 MW in 2025 to 23,968 MW in 2035, and service, DCIM and refresh spend per MW from USD 450 thousand to USD 550 thousand a year, so the installed-base pool grows from USD 672.25 million to USD 13.18 billion. Monitoring software for these halls is sized in the Asia Pacific Data Center Infrastructure Management (DCIM) Market report.

Pricing: What Is the Asia Pacific AI Data Center Price per MW?

An Asia Pacific AI data center cost about USD 11.00 million per MW of IT load to build and equip in 2025, before servers, in DC Market Insights estimates, rising to USD 13.00 million per MW by 2035 as liquid cooling and denser electrical trains spread.

Year Facility and network content per new AI MW (USD million) Service and refresh per installed AI MW (USD thousand a year) New AI IT capacity (MW) AI share of new APAC capacity
2020 9.00 350 104 8%
2025 11.00 450 870 30%
2030 12.50 500 2,860 55%
2035 13.00 550 3,968 62%

Public anchors: Korean data center development cost including land is about KRW 17.6 billion (USD 13 million) per MW (Source: Financial News, citing Cushman & Wakefield), so the USD 11.00 million AI content excluding land is about 85% of that all-in figure. The NAVER, Brookfield and NVIDIA expansion of GAK Sejong from 55 MW to 200 MW (Source: NAVER) implies about USD 1.68 billion of facility content for the 145 MW increment at the 2027 rate of USD 11.60 million per MW.

Standards and Regulation: Which Rules Shape the Asia Pacific AI Data Center Market?

Three rule sets shape the USD 10.24 billion market: efficiency caps, grid-impact rules and national AI compute programmes.

  • Singapore Green Data Centre Roadmap (Infocomm Media Development Authority, IMDA): at least 300 MW of additional capacity and a PUE target of 1.3 or lower (Source: IMDA).
  • Taiwan Energy Utilization Manual rule (MOEA): PUE caps of 1.3 (hyperscale) and 1.4 (colocation) for sites of 5 MW or more (Source: Ministry of Economic Affairs, Taiwan).
  • Korea Special Act on Activation of Distributed Energy (2023): power system impact assessment for large users (Source: Korea Legislation Research Institute).
  • IndiaAI Mission: Rs 10,372 crore outlay and a common compute portal with more than 38,000 GPUs (Source: Digital India, Government of India).

Market Segmentation: Which Segment Leads the Asia Pacific AI Data Center Market, and What Is the Fastest Growing Asia Pacific AI Data Center Segment?

AI Inference is the fastest growing Asia Pacific AI data center segment at 25.45% a year to 2035, and Electrical Infrastructure is the largest component with 38% of 2025 spend (USD 3.89 billion).

By Component: Electrical Infrastructure Leads with 38.00%

Segment 2025 (USD million) Share 2025 2035 (USD million) CAGR 2025-2035
Electrical Infrastructure 3,892.05 38.00% 23,315.83 19.60%
Cooling Infrastructure 1,741.18 17.00% 13,600.90 22.82%
Building Shell and Fit-Out 2,765.41 27.00% 14,248.56 17.82%
AI Network Fabric 1,024.22 10.00% 7,124.28 21.40%
Operations Services and Software 819.38 8.00% 6,476.62 22.97%

Operations Services and Software grows fastest among components at 22.97% as the installed base compounds, and Cooling Infrastructure follows at 22.82%.

By Data Center Type: Hyperscale Self-Build Leads with 48.00%

Segment 2025 (USD million) Share 2025 2035 (USD million) CAGR 2025-2035
Hyperscale Self-Build 4,916.28 48.00% 28,497.13 19.21%
Colocation and Wholesale AI 3,482.36 34.00% 23,315.83 20.94%
Sovereign and Government AI 1,024.22 10.00% 8,419.61 23.45%
Enterprise and Edge AI 819.38 8.00% 4,533.63 18.66%

Sovereign and Government AI grows at 23.45%: the Samsung SDS-led consortium for Korea’s National AI Computing Center in Haenam targets more than 15,000 GPUs by 2028 (Source: Samsung SDS). Small sites are covered in the Asia Pacific Edge Data Center Market report.

By Workload: AI Training Leads with 62.00%

Segment 2025 (USD million) Share 2025 2035 (USD million) CAGR 2025-2035
AI Training 6,350.19 62.00% 27,201.80 15.66%
AI Inference 3,892.05 38.00% 37,564.40 25.45%

AI Training still buys USD 27.20 billion in 2035, but its share falls from 62% to 42%.

DC Exclusive: The APAC AI Facility Content per MW Model and Component × Data Center Type Matrix

Hyperscale self-builds buy 52.5% of 2025 AI cooling spend. This is a modelled estimate by DC Market Insights, built from 4 inputs (new APAC IT MW, AI share of new MW, facility and network content per AI MW, and service spend per installed AI MW, each for 16 years) and checked against the 25 facts in the source ledger. The matrix splits the 2025 market of USD 10.24 billion by component and data center type; rows and columns sum to the segment totals above.

Component (USD million, 2025) Hyperscale Self-Build Colocation and Wholesale AI Sovereign and Government AI Enterprise and Edge AI Total
Electrical Infrastructure 1,876.43 1,320.53 384.30 310.80 3,892.06
Cooling Infrastructure 913.95 555.48 161.65 110.09 1,741.17
Building Shell and Fit-Out 1,276.85 1,040.46 247.74 200.36 2,765.41
AI Network Fabric 526.93 282.53 123.33 91.43 1,024.22
Operations Services and Software 322.12 283.36 107.20 106.70 819.38
Total 4,916.28 3,482.36 1,024.22 819.38 10,242.24

Colocation and Wholesale AI buys USD 1.04 billion of shell and fit-out, the largest colocation cell after electrical, because wholesale operators build the shell before tenants specify cooling.

Regional Insights: Which Region Leads the Asia Pacific AI Data Center Market, and Which Grows Fastest?

Greater China leads the Asia Pacific AI Data Center Market with 40.00% of 2025 spend (USD 4.10 billion), and South Asia is the fastest growing region at 25.69% a year to 2035.

Region 2025 (USD million) Share 2025 2035 (USD million) CAGR 2025-2035
Greater China 4,096.90 40.00% 21,372.85 17.96%
Japan and South Korea 1,946.03 19.00% 11,010.25 18.92%
Southeast Asia 1,843.61 18.00% 14,896.23 23.24%
South Asia 921.80 9.00% 9,067.27 25.69%
Oceania 1,433.91 14.00% 8,419.61 19.36%

Greater China Holds 40.00% Share

The Greater China AI data center market was USD 4.10 billion in 2025 and reaches USD 21.37 billion by 2035 at 17.96%. China held 25% of global data centre electricity use in 2024 (Source: IEA); in Taiwan, the Foxconn, NVIDIA and government AI factory features 10,000 Blackwell GPUs (Source: NVIDIA).

Japan and South Korea Holds 19.00% Share

The Japan and South Korea AI data center market was USD 1.95 billion in 2025 and reaches USD 11.01 billion by 2035 at 18.92%. Microsoft’s USD 2.9 billion Japan programme (Source: Microsoft) and Korea’s 260,000-GPU plan (Source: NVIDIA) anchor the region.

Southeast Asia Holds 18.00% Share

The Southeast Asia AI data center market was USD 1.84 billion in 2025 and reaches USD 14.90 billion by 2035 at 23.24%. Malaysia had 1,039 MW under construction in H1 2026 (Source: Cushman & Wakefield).

South Asia Holds 9.00% Share

The South Asia AI data center market was USD 921.80 million in 2025 and reaches USD 9.07 billion by 2035 at 25.69%. Google’s USD 15 billion Visakhapatnam hub for 2026-2030 (Source: Google Cloud) and the 38,000-GPU IndiaAI portal (Source: Digital India) drive the fastest growth.

Oceania Holds 14.00% Share

The Oceania AI data center market was USD 1.43 billion in 2025 and reaches USD 8.42 billion by 2035 at 19.36%. Amazon’s AUD 20 billion Australian programme for 2025-2029 (Source: Amazon) supports Sydney and Melbourne AI campuses.

Country Analysis: Which Countries Lead Asia Pacific AI Data Center Spend?

China leads with USD 3.69 billion in 2025, and India is the fastest growing large country at 25.74% a year, reaching USD 8.10 billion by 2035.

Rank Country Region 2025 (USD million) Share 2025 2035 (USD million) CAGR 2025-2035
1 China Greater China 3,687.21 36.00% 19,106.03 17.88%
2 Australia Oceania 1,229.07 12.00% 7,448.11 19.74%
3 Japan Japan and South Korea 1,126.65 11.00% 6,476.62 19.11%
4 Malaysia Southeast Asia 870.59 8.50% 5,828.96 20.94%
5 South Korea Japan and South Korea 819.38 8.00% 4,533.63 18.66%
6 India South Asia 819.38 8.00% 8,095.77 25.74%
7 Singapore Southeast Asia 409.69 4.00% 2,266.82 18.66%
8 Thailand Southeast Asia 256.06 2.50% 2,590.65 26.04%
9 Indonesia Southeast Asia 256.06 2.50% 2,914.48 27.53%
10 Hong Kong Greater China 204.84 2.00% 971.49 16.84%
11 Taiwan Greater China 204.84 2.00% 1,295.32 20.25%
12 New Zealand Oceania 153.63 1.50% 647.66 15.47%

The 12 countries hold 98.0% of 2025 spend; Vietnam, the Philippines and other markets make up the rest.

Forecast Scenarios: How Could the 2035 Asia Pacific AI Data Center Market Forecast Change?

The 2035 Asia Pacific AI Data Center Market ranges from USD 47.34 billion to USD 80.40 billion, with a base case of USD 64.77 billion.

Scenario 2035 size CAGR 2025-2035 Assumption
Low USD 47.34 billion 16.54% Grid limits cut 2035 new capacity by 20%; AI share of new MW 55%
Base USD 64.77 billion 20.25% New APAC IT capacity 5,200 MW in 2030 and 6,400 MW in 2035; AI share 62% in 2035
High USD 80.40 billion 22.88% All announced sovereign and hyperscale programmes on time; capacity 15% higher; AI share 68%

Competitive Insights: Who Are the Leading Asia Pacific AI Data Center Companies, and What Is the Market Share by Company?

The leading Asia Pacific AI data center companies are Amazon Web Services, Microsoft, Google, Alibaba Cloud and AirTrunk; DC Market Insights estimates these five control about 41% of 2025 spend as owners or developers, based on its mapping of announced AI campuses and committed MW.

  • NVIDIA
  • Amazon Web Services
  • Microsoft
  • Google
  • Alibaba Cloud
  • Tencent Cloud
  • AirTrunk
  • ST Telemedia Global Data Centres
  • Equinix
  • NTT DATA
  • Digital Realty
  • Princeton Digital Group
  • YTL Power International
  • AdaniConneX
  • SoftBank Corp.
  • KT Cloud
  • Schneider Electric
  • Vertiv

AirTrunk is the largest independent platform: Blackstone agreed to acquire it in September 2024 at an implied enterprise value of over AUD 24 billion, when it had more than 800 MW committed to customers (Source: Blackstone). Google’s Visakhapatnam hub is developed with AdaniConneX and Airtel (Source: Google Cloud), Amazon’s Australian programme runs to AUD 20 billion (Source: Amazon), and NAVER is expanding GAK Sejong with Brookfield and NVIDIA to 200 MW (Source: NAVER). Schneider Electric and Vertiv supply much of the electrical and cooling content of these halls.

Recent Developments: What Changed in the Asia Pacific AI Data Center Market in 2026?

Six dated events between January 2026 and August 2026 shaped the market.

  • In August 2026, Cushman & Wakefield reported a record 26.5 GW APAC data centre pipeline for H1 2026, up 7.1 GW in six months. (Source: Cushman & Wakefield)
  • In July 2026, NAVER, Brookfield and NVIDIA agreed to expand the GAK Sejong AI factory from 55 MW to 200 MW. (Source: NAVER)
  • In May 2026, CBRE reported record APAC data centre investment of USD 11.6 billion in 2025. (Source: CBRE)
  • In April 2026, the IndiaAI Mission reported more than 38,000 GPUs onboarded on its compute portal. (Source: Digital India, Government of India)
  • In January 2026, Taiwan’s Executive Yuan approved the Ten AI Initiatives Promotion Plan to expand AI computing capacity. (Source: National Development Council)
  • In January 2026, the Samsung SDS-led consortium set a target of more than 15,000 GPUs by 2028 for Korea’s National AI Computing Center. (Source: Samsung SDS)

More regional and segment studies are listed under DC Market Insights industries; for all workloads see the Asia Pacific Data Center Market.

Methodology: How Did DC Market Insights Build the Asia Pacific AI Data Center Market Model?

The model multiplies 870 MW by USD 11.00 million per MW for 2025. DC Market Insights built the Asia Pacific AI Data Center Market bottom-up: new APAC IT capacity (2,900 MW in 2025, 6,400 MW in 2035) × AI share (30% to 62%) × facility and network content per AI MW (USD 11.00 million to USD 13.00 million) gives new-build spend, and installed AI capacity (1,494 MW to 23,968 MW, with 2% a year retired) × service spend per MW gives installed-base spend. Segment, region and country splits are set for 2025 and 2035, interpolated and normalised so that every split sums to the total every year.

Three bound checks:

  • Ceiling 1: the 2024 market of USD 6.14 billion is 1.2% of global data centre investment of about USD 0.5 trillion in 2024 (Source: IEA), which includes servers and all workloads worldwide.
  • Ceiling 2: 2026 new AI capacity of 1,176 MW is 42% of the roughly 2,800 MW a year APAC delivery pace implied by 1.4 GW delivered in H1 2026 (Source: Cushman & Wakefield), below 100%.
  • Floor: 1,494 MW installed × USD 450 thousand per MW of service spend gives USD 672.25 million before any new build; AirTrunk alone had more than 800 MW committed (Source: Blackstone).

Full approach: DC Market Insights research methodology.

Table of contents

  • 23 chapters
  • 1 figures
  • 9 data tables
  1. 01Introduction
    • 1.1Market Definition and Scope
    • 1.2Historical Period, Base Year and Forecast Period
    • 1.3Report Attributes
  2. 02Research Methodology
    • 2.1Build Receipt: New AI IT Capacity × Facility and Network Content per MW
    • 2.2Installed-Base Service and Refresh Spend
    • 2.3Segment, Region and Country Splits
    • 2.4Bound Checks: Two Ceilings and a Floor
  3. 03Executive Summary
    • 3.1Market Snapshot, 2025, 2030 and 2035
    • 3.2Key Takeaways
  4. 04Market Scope and Definition: What Is an Asia Pacific AI Data Center?
    • 4.1Included: Electrical, Cooling, Building Shell and Fit-Out, AI Network Fabric, Operations Services and Software
    • 4.2Excluded: GPU and CPU Servers, Storage, Land, Electricity Bills and Cloud Rental Revenue
    • 4.3How the Segments Fit Together
  5. 05Asia Pacific AI Data Center Market Drivers
    • 5.1Hyperscaler Commitments Exceed USD 30 Billion Across Four Countries
    • 5.2Sovereign GPU Programmes Add 300,000 GPUs
    • 5.3Electricity Demand Rose 17% in 2025
  6. 06Market Trends
    • 6.1Inference Grows from USD 3.89 billion to USD 37.56 billion
    • 6.2Cooling Rises from 17% to 21% of Spend
    • 6.3Southeast Asia and India Take 37% by 2035
  7. 07Market Challenges
    • 7.1Grid Connections Gate Up to 20% of 2035 Capacity
    • 7.2Efficiency Caps Raise Content per MW by USD 0.5 Million
  8. 08Market Opportunities
    • 8.1Southeast Asia and South Asia Add USD 21.20 billion
    • 8.2Installed-Base Services Reach USD 13.18 billion
  9. 09Pricing Analysis: AI Data Center Price per MW
    • 9.1Facility and Network Content per New AI MW, 2020-2035
    • 9.2Service and Refresh Spend per Installed AI MW
    • 9.3Public Cost Anchors: Korean Development Cost and GAK Sejong Expansion
  10. 10Standards and Regulation
    • 10.1Singapore Green Data Centre Roadmap (IMDA)
    • 10.2Taiwan Energy Utilization Manual Rule (MOEA)
    • 10.3Korea Special Act on Activation of Distributed Energy (2023)
    • 10.4IndiaAI Mission
  11. 11Asia Pacific AI Data Center Market, By Component
    • 11.1Electrical Infrastructure
    • 11.2Cooling Infrastructure
    • 11.3Building Shell and Fit-Out
    • 11.4AI Network Fabric
    • 11.5Operations Services and Software
  12. 12Asia Pacific AI Data Center Market, By Data Center Type
    • 12.1Hyperscale Self-Build
    • 12.2Colocation and Wholesale AI
    • 12.3Sovereign and Government AI
    • 12.4Enterprise and Edge AI
  13. 13Asia Pacific AI Data Center Market, By Workload
    • 13.1AI Training
    • 13.2AI Inference
  14. 14DC Exclusive: APAC AI Facility Content per MW Model and Component × Data Center Type Matrix
    • 14.1APAC AI Facility Content per MW Model
    • 14.2Component × Data Center Type Matrix, 2025
  15. 15Asia Pacific AI Data Center Market, By Region
    • 15.1Greater China
    • 15.2Japan and South Korea
    • 15.3Southeast Asia
    • 15.4South Asia
    • 15.5Oceania
  16. 16Country Analysis
    • 16.1China
    • 16.2Australia
    • 16.3Japan
    • 16.4Malaysia
    • 16.5South Korea
    • 16.6India
    • 16.7Singapore
    • 16.8Thailand
    • 16.9Indonesia
    • 16.10Hong Kong
    • 16.11Taiwan
    • 16.12New Zealand
  17. 17Forecast Scenarios 2035
    • 17.1Low Scenario
    • 17.2Base Scenario
    • 17.3High Scenario
  18. 18Competitive Landscape
    • 18.1Market Share of the Five Leading Companies, 2025
    • 18.2AirTrunk: Largest Independent Platform
    • 18.3Hyperscaler and Sovereign Campus Partnerships
    • 18.4Electrical and Cooling Equipment Suppliers
  19. 19Company Profiles
    • 19.1NVIDIA
    • 19.2Amazon Web Services
    • 19.3Microsoft
    • 19.4Google
    • 19.5Alibaba Cloud
    • 19.6Tencent Cloud
    • 19.7AirTrunk
    • 19.8ST Telemedia Global Data Centres
    • 19.9Equinix
    • 19.10NTT DATA
    • 19.11Digital Realty
    • 19.12Princeton Digital Group
    • 19.13YTL Power International
    • 19.14AdaniConneX
    • 19.15SoftBank Corp.
    • 19.16KT Cloud
    • 19.17Schneider Electric
    • 19.18Vertiv
  20. 20Recent Developments
    • 20.1APAC Pipeline Reaches a Record 26.5 GW (August 2026)
    • 20.2GAK Sejong Expansion to 200 MW (July 2026)
    • 20.3Record APAC Data Centre Investment of USD 11.6 Billion (May 2026)
    • 20.4IndiaAI Mission Passes 38,000 GPUs (April 2026)
    • 20.5Taiwan Ten AI Initiatives Promotion Plan (January 2026)
    • 20.6Korea National AI Computing Center Targets 15,000 GPUs (January 2026)
  21. 21Analyst Recommendations
  22. 22Appendix
    • 22.1Abbreviations: AI, APAC, GPU, CPU, MW, GW, kW, IT, CAGR, UPS, CDU, DCIM, PUE
  23. 23Glossary
  1. Fig. 1Asia Pacific AI Data Center Market Size, 2020-2035 (USD Million)
  1. Table 1Report Attributes
  2. Table 2Facility and Network Content per AI MW, 2020-2035
  3. Table 3Market by Component, 2025 and 2035 (USD Million)
  4. Table 4Market by Data Center Type, 2025 and 2035 (USD Million)
  5. Table 5Market by Workload, 2025 and 2035 (USD Million)
  6. Table 6Component × Data Center Type Matrix, 2025 (USD Million)
  7. Table 7Market by Region, 2025 and 2035 (USD Million)
  8. Table 8Top 12 Countries, 2025 and 2035 (USD Million)
  9. Table 9Forecast Scenarios, 2035

About this report

Amit JainDeepti Agrawal

Written by Amit Jain and reviewed by Deepti Agrawal, Senior Editor, Research. Figures are built top-down and bottom-up and reconciled before publication, with 2025 as the base year.

Read our methodology

How we built this

  • Historical period2020-2024
  • Base year2025
  • Forecast period2026-2035
  • Sizing approachTop-down + bottom-up

Need more than the report?

Advisory on this market

Due diligence, site selection and market entry work, by the analysts who wrote this report.

See consulting services

Frequently asked questions

What is the Asia Pacific AI data center market size?

USD 10.24 billion in 2025. DC Market Insights values the Asia Pacific AI Data Center Market at USD 10.24 billion in 2025, up from USD 1.02 billion in 2020, and forecasts USD 64.77 billion by 2035.

What is the Asia Pacific AI data center market growth rate?

20.25% a year from 2025 to 2035. The market grows from USD 10.24 billion to USD 64.77 billion as new AI IT capacity in Asia Pacific rises from 870 MW to 3,968 MW a year.

What is an Asia Pacific AI data center?

A facility built for GPU clusters at 40 kW to over 130 kW per rack. Its electrical, cooling, shell, network fabric and operations spend was a USD 10.24 billion market in 2025, servers excluded.

What is the Asia Pacific AI data center price per MW?

About USD 11.00 million per MW of IT load in 2025, before servers, in DC Market Insights estimates, rising to USD 13.00 million per MW by 2035 as liquid cooling spreads.

What is the fastest growing Asia Pacific AI data center segment?

AI Inference at 25.45% a year, from USD 3.89 billion in 2025 to USD 37.56 billion in 2035. Sovereign and Government AI is the fastest data center type at 23.45%.

Who are the leading Asia Pacific AI data center companies?

About 41% of 2025 spend sits with Amazon Web Services, Microsoft, Google, Alibaba Cloud and AirTrunk, in DC Market Insights estimates. NTT DATA, ST Telemedia and Equinix follow.

Which region leads the Asia Pacific AI data center market?

Greater China with 40.00% (USD 4.10 billion) in 2025. South Asia grows fastest at 25.69% a year, led by India at 25.74%.

What are the Asia Pacific AI data center market trends?

58% of spend goes to inference by 2035, up from 38%; cooling rises from 17% to 21% of spend; and Southeast Asia and India take 37% of the market by 2035.

How big is the Asia Pacific AI data center market compared with the wider data center market?

USD 10.24 billion in 2025. This report covers only facility infrastructure built for AI workloads: electrical, cooling, shell and fit-out, network fabric, and operations. GPU servers and storage are excluded. The host market for all workloads is covered in the Asia Pacific Data Center Market report.

How much new AI capacity does the APAC data center market add each year?

870 MW in 2025. DC Market Insights estimates new AI IT capacity in APAC rises to 2,860 MW in 2030 and 3,968 MW in 2035, lifting the market to USD 40.24 billion in 2030.

What is the Asia Pacific AI data center market size in 2035?

USD 64.77 billion by 2035. The market grows from USD 10.24 billion in 2025 at a CAGR of 20.25%. In the low case, with 20% fewer MW connected by 2035, it reaches USD 47.34 billion instead.

Get this report

How do I get the full report?

Request a quote; we reply by email within one business day.

How and when will I receive the report?

By email, within 24-48 business hours of payment.

How can I pay?

Once you accept the quote, we email you a secure payment link.

Can I see a sample first?

Yes. Request a free sample and we will email it to you.

Can the report be tailored to my needs?

Yes. We can add countries, segments or data cuts. Talk to an analyst about the scope you need.

Can I get an invoice or pay against a purchase order?

Write to sales@dcmarketinsights.com and our team will help.

Refund and returns policy

Request a quote

Asia Pacific AI Data Center Market

Tell us what you need. We reply by email with scope, delivery date and a quote, usually within one business day.

Contact Form

Your data is safe with us

  • Your details stay private. We never sell or share them.
  • Used only to reply to your message
  • No spam: we only send what you asked for
  • Encrypted (SSL) form, handled by our own team

Privacy Policy

Free sample

Asia Pacific AI Data Center Market

See the structure, tables and charts before you decide.

Smallform of Sample request

Your data is safe with us

  • Your details stay private. We never sell or share them.
  • Used only to send your sample and answer your questions
  • No spam: we only send what you asked for
  • Encrypted (SSL) form, handled by our own team

Privacy Policy

Talk to an analyst

Ask the team that built this report

Different scope, country or data cut? A senior analyst replies within one business day.

Contact Form

Your data is safe with us

  • Your message goes straight to our research team
  • Your details stay private. We never sell or share them.

Privacy Policy