Beyond the Announcement

Southeast Asia's data-centre pipeline looks enormous.

But the more closely I looked at the headline numbers, the less useful the total became. Not because the announcements were necessarily wrong, but because the numbers were often answering different questions.

One project might describe the long-term capacity of an entire campus. Another might disclose only critical IT load. A third might announce the amount of power requested from a utility. Another might have secured renewable energy through a power purchase agreement. Somewhere else, construction may have started on only the first phase of a much larger plan.

All of those facts can be true. They are still not the same fact.

When they are placed in one pipeline number, a reader can easily come away believing that all of the capacity is equally close to operation. It is not.

The question behind this work

I have spent years working around data-centre projects in Singapore, Malaysia and India. I understood the broad development process, but I still found it surprisingly difficult to answer a simple question from public information:

How much announced capacity is demonstrably progressing toward operation?

Answering that question required more than collecting announcements. It required following each identifiable project through the evidence that can turn a plan into usable infrastructure.

That is why I built the Southeast Asia Data-Centre Conversion Monitor: a public-source project and evidence register covering Singapore, Malaysia and Johor, Thailand, Indonesia and Batam, Vietnam, and the Philippines.

The objective is not to create the largest possible project list. It is to build a universe that can withstand a reasonable challenge from a developer, utility executive, investor, operator or supplier asking: What exactly supports this number?

What the first review shows

As of 17 September 2026, the reviewed universe contains 65 project or programme records.

Across those records, the register identifies:

• 4,627 MW of verified announced or campus capacity;
• 1,929 MW IT explicitly described as IT load or critical IT capacity;
• 540 MW with project-specific evidence of physical power allocation;
• 154 MW tied to project-specific construction evidence;
• 30 MW supported by energization or ready-for-service power evidence;
• 6 MW with project-specific operating confirmation in this reviewed pipeline; and
• 70 MW supported by a publicly verifiable customer commitment.

These figures must not be added together. They overlap and answer different questions.

The 4,627 MW figure describes publicly announced project or campus ambition where the capacity is supported by a qualifying source. It does not mean that 4,627 MW has been allocated by utilities, is under construction or is ready to operate.

The 6 MW operational figure is also easy to misunderstand. It is not an estimate of Southeast Asia's total operating data-centre base. It is only the operational capacity confirmed within this project-focused pipeline review. The existing operating market is much larger and requires a different facility-level census.

A PPA is not a power connection

This is one of the most important distinctions in the register.

A power purchase agreement can be commercially significant. It may support renewable-energy procurement, price certainty, financing or sustainability commitments. But it does not, by itself, show that electricity can physically reach a particular data-centre site at the required capacity and time.

Physical readiness requires different evidence: a project-specific utility allocation or supply approval, an interconnection pathway, and, where necessary, the substations or transmission works required to deliver the load.

The register therefore keeps energy procurement, utility allocation, grid connection and grid infrastructure in separate fields. One cannot silently stand in for another.

Following the gates

For each project or phase, the research follows the available evidence across a series of execution gates:

Announcement. Land or site control. Government or investment approval. Planning and environmental approvals. Physical power allocation. Grid interconnection. Required substation or transmission works. Energy procurement. Water readiness. Financing. Customer commitment. EPC appointment. Groundbreaking. Construction. Energization. Ready for service. Operation. Utilization.

Projects do not always move through these gates in a neat sequence. Some activities happen in parallel, and much of the evidence can remain private. The register does not treat the absence of public evidence as proof that something has not happened.

It uses a more careful rule:

• Verified means directly supported by a strong source.
• Reported means supported by credible secondary reporting but not independently confirmed.
• Unknown means not publicly verified in the sources reviewed. It does not mean no.

There is no subjective readiness score and no probability of completion. A project's current stage advances only when the required evidence for that stage is present.

What this research can and cannot do

A public-source register cannot see every utility discussion, customer negotiation, financing arrangement or permitting exchange. It will always be less complete than the internal information held by the parties developing a project.

It can still impose useful discipline.

It can prevent an announced campus plan from being presented as operational IT load. It can stop a land purchase from being mistaken for a construction start. It can keep a renewable-energy contract separate from physical grid readiness. It can flag a target ready-for-service date that has passed without public operating confirmation. And it can preserve conflicting capacities or dates rather than quietly selecting the most convenient one.

That is the value of the monitor: not certainty, but traceability.

Where the series goes next

This is the introduction to a series that will explain the framework one question at a time.

The next pieces will examine why not all megawatts mean the same thing; why a PPA is not proof of physical power readiness; which execution gates matter between announcement and operation; how the constraints differ across the six covered markets; and where verified project progress creates opportunities for utilities, engineering firms, contractors, equipment suppliers and other partners.

The public series will share the framework, selected verified aggregates and conclusions that the evidence can support. The detailed project register, evidence IDs, source notes, conflicts and unresolved gates will remain controlled research.

That boundary is deliberate. The value is not in having a long list of projects. It is in being able to show what each claim means, what supports it and what remains unknown.

← Back to the research monitor