Data Centre Capacity in Asia: How to Read MW Claims Without Confusing Planned, Built and Operational Supply
A data-centre development can be described as a 100MW project long before 100MW of computing infrastructure is operating on the site.
The land may have been acquired. Utility capacity may be proposed. The first building may be under construction. A multi-phase campus may ultimately be designed for the headline figure. Yet only part—or none—of that capacity may currently be operational.
That distinction is becoming more important as data-cententre development expands across Asia. Cushman & Wakefield reported that the Asia-Pacific development pipeline reached 26,455MW in the first half of 2026, but separated that figure into 4,764MW under construction and 21,691MW in planning. It also recorded approximately 1,372MW of new operational capacity delivered during the half-year.
Those categories describe very different stages of property development. For developers, investors, industrial-land owners and occupiers, the useful question is therefore not simply, “How many megawatts?” It is: what exactly does the megawatt figure represent?
MW Is Not a Complete Description of a Data Centre
Megawatts provide a convenient way to discuss the scale of data-centre infrastructure, but the number needs a definition.
Uptime Institute distinguishes between power-demand capacity—the maximum amount of power provisioned for a data centre—and IT-equipment capacity, which relates to the workload that computing equipment can support.
Property and industry datasets can also describe capacity using their own methodologies. One report may focus on operational IT load, another may discuss total facility capacity, while a project announcement may describe the eventual capacity of an entire campus.
This means two credible sources can publish different MW numbers for the same market without one necessarily being wrong. Before comparing them, readers need to establish whether the measurement boundaries are equivalent.
| Capacity description | What it usually tells you | What it does not prove |
|---|---|---|
| Planned capacity | Capacity intended for a proposed or future project | That construction has begun or capacity will definitely be delivered |
| Under construction | Capacity associated with projects currently being built | That the facility has been commissioned or is serving customers |
| Operational capacity | Capacity associated with facilities that have entered operation under the dataset’s methodology | That every MW is currently occupied or used by customers |
| Campus ultimate capacity | Potential scale after all intended phases are completed | That all phases are funded, approved, powered or under construction today |
| Power secured or allocated | Some form of electricity provision or allocation has been arranged | That buildings and IT infrastructure have been completed |
Johor Shows Why Development Stage Matters
Johor provides a useful example because its data-centre market has expanded quickly.
Cushman & Wakefield reported 1,110MW of operational capacity in Johor in H1 2026 and 602MW under construction. Its overall development pipeline for the market was reported at 3,088MW.
Describing Johor simply as having more than 3GW of operating data centres would therefore misrepresent the dataset. A significant portion of the figure relates to development that had not yet reached operational status.
The distinction matters for property analysis.
An operational facility represents delivered real estate and infrastructure. An under-construction project represents development execution risk. Planned supply introduces additional questions around approvals, financing, power, water, construction, customer commitments and timing.
A large pipeline is evidence of development interest. It is not the same thing as existing supply.
Singapore Demonstrates Another Problem: Definitions Can Differ
Singapore’s Infocomm Media Development Authority stated in its Green Data Centre Roadmap that the country had more than 1.4GW of data-centre capacity and intended to provide at least another 300MW in the near term.
Commercial property datasets may report different totals depending on their definitions, markets covered, facility thresholds and measurement methodology.
Readers should therefore resist comparing two headline figures until they have checked:
- the reporting period;
- the geographical boundary;
- whether the measurement concerns power or IT capacity;
- which facilities are included;
- whether the figure is operational or projected;
- and whether planned phases have been included.
This is especially important when comparing markets. A 1GW figure built using one methodology should not automatically be ranked against a 900MW figure produced using another.
The Five-Question Capacity Test
A practical way to evaluate a data-centre capacity claim is to run it through five questions.
1. What is being measured?
Determine whether the figure represents IT load, available electrical capacity, total facility load, contracted capacity or an ultimate project design.
“100MW data centre” is incomplete without this context.
2. What development stage has been reached?
Ask whether the project is proposed, approved, planned, under construction, commissioned or operational.
A land transaction should not be reported as an operating data centre. Likewise, construction commencement does not establish that the entire campus capacity has been delivered.
3. What does the number cover?
A headline capacity may apply to one building, one development phase, an entire campus or a portfolio of facilities.
Large campuses are frequently delivered in stages. The final design capacity should therefore be separated from what exists at the reporting date.
4. Is the infrastructure behind the number credible?
Power is particularly important. Asia Property Reviews has previously examined why electrical infrastructure is becoming one of the defining constraints on data-centre property across Southeast Asia.
But power is not the only requirement. Fibre connectivity, water or cooling infrastructure where applicable, planning approvals, construction capability and suitable land also influence whether planned capacity can become operational capacity.
5. Can the claim be independently checked?
A credible capacity claim should be traceable to documentation.
Depending on the claim, evidence may include planning records, utility documentation, construction records, commissioning documentation, operator disclosures, engineering information and credible independent market research.
A marketing presentation repeating the same headline number is not independent verification of that number.
Do Not Confuse Capacity With Utilisation
Even operational capacity needs careful interpretation.
A data centre entering operation does not mean every available MW is immediately occupied by customers. Likewise, contracted capacity and physically installed computing equipment are different concepts.
For property investors, this distinction resembles the difference between a completed office building and a fully occupied office building. Development completion establishes that the asset exists; leasing and utilisation determine how much of that asset is commercially used.
Data-centre analysis should therefore separate at least three questions:
- Has the infrastructure been delivered?
- How much operational capacity exists?
- How much of that capacity has been committed or utilised?
Evidence for one does not automatically establish the others.
Why Developers Should Be Precise About Future Capacity
Headline capacity can be commercially useful. It communicates scale to potential customers, investors and surrounding industries.
But ambiguity creates problems when a project moves from marketing into due diligence.
Instead of saying:
“The development provides 300MW of data-centre capacity.”
a developer may need to distinguish between:
- capacity currently operational;
- capacity under active construction;
- capacity planned for later phases;
- and the ultimate designed capacity of the development.
That is not weaker communication. It is more informative communication.
Knight Frank’s Malaysia research in the first half of 2026 noted that infrastructure credibility, resource accountability, committed demand and utility readiness were becoming increasingly important as Malaysia’s data-centre market matured. The same principle applies to capacity claims: credibility improves when the number can be traced to a defined development stage and supporting evidence.
When a Data-Centre Milestone Becomes a Verifiable Achievement
Some data-centre developments may eventually produce genuinely exceptional and measurable milestones. These could involve physical scale, operational capacity, infrastructure deployment or another clearly defined benchmark.
The important distinction is between an impressive project description and a defined comparative achievement.
For organisations exploring record recognition in Asia, the Asia Record official application process requires a proposed achievement to be measurable and supported by evidence describing what was achieved and the conditions under which it was measured.
A property or infrastructure company considering an Asia Record application therefore needs more than a large headline number. The measurement boundary, geography, period, comparison basis and evidence all need to be clear.
This is also the practical answer for companies asking how to get an Asia Record or whether they should apply for Asia Record recognition. An exceptional development claim first needs to become a specific, measurable and verifiable claim.
Asia Record certification, business achievement recognition in Asia or status as an Asia Record holder should also remain separate from assessments of investment quality. Record recognition can document a particular measurable achievement; it does not establish that a data centre is a superior investment, suitable for every occupier or compliant with every regulatory and technical requirement.
The same distinction applies to company recognition in Asia, a corporate record in Asia, a business record in Asia or business awards in Asia generally: recognition of a defined corporate achievement should not be treated as evidence for unrelated property-performance claims.
A Due-Diligence Checklist for Data-Centre Capacity Claims
Before relying on a headline MW number, property professionals should check:
- What unit and definition of capacity is being used?
- Is the figure operational, under construction, planned or ultimate capacity?
- What was the reporting date?
- Does the capacity apply to one facility, one phase or the full campus?
- Is sufficient power merely proposed, allocated or actually available?
- Has the facility been commissioned?
- Are future phases dependent on additional approvals or infrastructure?
- Does the source explain its methodology?
- Are different market reports using comparable definitions?
- Can the underlying claim be verified using primary records or credible independent research?
Read the Development Stage Before the Headline Number
Asia’s data-centre market is producing increasingly large capacity numbers. Those numbers are useful, but only after their boundaries are understood.
A planned 500MW campus, a 500MW power allocation and 500MW of operational data-centre capacity describe three different situations. Treating them as interchangeable can distort market comparisons, project assessments and investment analysis.
The most useful data-centre reporting therefore does more than publish the biggest number. It tells readers exactly what the number represents, when it was measured and what evidence supports it.
For developers, that creates more credible project communication. For property investors and occupiers, it produces better due diligence. And when a development genuinely reaches an exceptional milestone, precise measurement is what turns a marketing claim into an achievement that can actually be verified.
