데이터센터 전력, 입지, 파이프라인 인텔리전스 - 아시아태평양판

OVERVIEW

One Malaysian utility has contracted 8.3 GW to data centres and is delivering 1,054 MW, both as at March 2026. In metropolitan Korea, 3 per cent of requested capacity has cleared the statutory assessment that now gates every project above 10 MW, as at July 2026. Australia's system operator tracks 225 projects and reports that more than 40 per cent have withdrawn or gone backwards in a single year. Singapore has provisionally allocated approximately 280 MW across two Call-for-Application rounds - eight awarded proposals, seven distinct applicant groups. Indonesia carries roughly 41 per cent of installed-capacity margin above peak load alongside connection lead times reported at up to two years, because the binding constraint is substations rather than generation. Japan's grid is split at two frequencies with 2.1 GW of conversion between them, and no demand-connection lead time is published at all.

The markets that look largest on announced capacity are not the markets that energise fastest, and in four of the six the two rankings are close to inverted. Every mark in the report is set by five gates applied identically at national and hub level, and listed in full - geography by geography, tranche by tranche, with the evidence and citation behind each - in chapter 1's Table 1.5 and in Appendix F, the dataset underlying every mark.

The six country chapters follow the same fourteen-section structure so they can be compared, but the conclusions are not forced into a common shape. Singapore's constraint is administrative. Australia's is state-level and now federally contested. Japan's is a frequency boundary and an unpublished demand queue. Korea's is a statutory assessment with a three per cent pass rate. Malaysia's is a contracted book eight times its metered load. Indonesia's is a substation problem wearing a generation surplus as a disguise. Treating these as variants of one story is the most common analytical error in the region.

WHAT THIS EDITION COVERS

Fifteen chapters in three parts, a pre-commitment verification checklist and six appendices, across 283 pages. Part I compares the region: the six-country scorecard, the regional project register, power systems compared, demand and customer evidence, site and infrastructure intelligence, design and delivery constraints, economics, capital and transactions, and commercial and procurement intelligence. Part II covers Japan, South Korea, Singapore, Malaysia, Indonesia and Australia, each on the same fourteen-section structure - decision summary, market structure and demand, hubs and emerging locations, verified capacity and pipeline, power availability, cost and connection timing, site and permitting constraints, fibre, water, cooling and physical risks, construction and supply chain, capital and economics, commercial opportunities and counterparties, a twelve-to-thirty-six-month outlook, implications by reader type, a country scorecard, and sources and verification points. Part III is the final decision framework. Appendices: methodology and the deliverability test, glossary, verification agenda, index, source register, and the ratings dataset. Systematic coverage is limited to the six markets named; mainland China, Hong Kong, Taiwan and India are not covered, and Thailand, Vietnam, the Philippines and New Zealand appear only where they bear on the six.

WHO THIS IS FOR

Investment committees, developers, operators, energy suppliers, lenders and vendors weighing a 50–200 MW commitment in Asia-Pacific who need the binding constraint named market by market, with the evidence and the date behind every figure.

WHAT YOU GET

• One question asked of six markets: where 50–200 MW can be delivered, when it can be energised, what power will cost, and what will stop it
• A delivery table separating 50 MW from 100–200 MW per market, with the constraint that decides each one named
• Every mark set by five gates applied identically at national and hub level, with the full dataset behind them in Appendix F
• Every figure carries two evidence fields - nature and provenance - recorded separately, plus the date it was first stated
• A regional project register that records evidence rather than predicting outcomes; committed, early development and announced are what could be evidenced at the date stated
• No single regional pipeline number, because the evidence does not support one
• Six country chapters on an identical fourteen-section structure, so they can be compared without forcing their conclusions into one shape
• A pre-commitment verification checklist and a verification agenda listing what a buyer should confirm before relying on any of it

Table of Contents

Executive summary

Six markets, one question, and the answer in one paragraph.

The twelve findings

What the evidence supports, stated as findings.

How to read this report

Definitions, evidence fields and the deliverability test.

Part I The region

• Eight chapters comparing the six markets against each other.

Chapter 1 The six-country comparison scorecard

• Every mark, with the evidence and citation behind it.

Chapter 2 The regional project register

• A record of evidence, not a claim about outcomes.

Chapter 3 Power systems compared

• What each system can actually energise, and when.

Chapter 4 Demand and customer evidence

• Who is contracting, and what is metered.

Chapter 5 Site and infrastructure intelligence

• Land, substations, fibre and water across thirteen hubs.

Chapter 6 Design and delivery constraints

• What the build itself imposes on the timetable.

Chapter 7 Economics, capital and transactions

• What has been paid, by whom, on what terms.

Chapter 8 Commercial and procurement intelligence

• Counterparties and how capacity is actually procured.

Part II The six markets

• Japan, South Korea, Singapore, Malaysia, Indonesia and Australia, each on the same fourteen-section structure.

Chapter 9 Japan

• A frequency boundary, 2.1 GW of conversion, and no published demand-connection lead time.

Chapter 10 South Korea

• A statutory assessment above 10 MW with a three per cent pass rate.

Chapter 11 Singapore

• An administrative constraint: two Call-for-Application rounds, approximately 280 MW allocated.

Chapter 12 Malaysia

• 8.3 GW contracted against 1,054 MW drawing at one utility, as at March 2026.

Chapter 13 Indonesia

• A substation problem wearing a generation surplus as a disguise.

Chapter 14 Australia

• 225 projects tracked, more than 40 per cent withdrawn or backwards in a year.

Part III Decision

• The final decision framework and the pre-commitment verification checklist.

Chapter 15 The final decision framework

• The decision matrix an investment committee can defend.

REFERENCE

• Appendices
  • Methodology and the deliverability test, glossary, verification agenda, index, source register, and the ratings dataset.