100 MW ≠ 100 MW

Edition 23 closed with a question: if the value lies in arriving earlier, which control point actually captures it — the land, the interconnection, the utility allocation, or the data-centre owner? This edition takes that question forward.

A data-centre developer can announce 100 MW. Another can announce the same 100 MW. On a pipeline chart, they look identical. From an investor's perspective, they may be very different assets.

One may have secured land but still be waiting for power. Another may have advanced its approvals, secured a credible grid position and established a clearer path to ready-for-service. Both are called 100 MW. But 100 MW ≠ 100 MW.

That distinction is becoming more important as power becomes one of the binding constraints on data-centre growth in APAC. [source] There is a further dimension beyond location and time: what do you actually control that makes those MW deliverable?

In January 2026, Keppel secured rights to lease approximately 123 hectares near Melbourne for a data-centre campus with gross power capacity of up to 720 MW — under an Agreement for Lease where it pays an annual access fee for early access while obtaining planning approvals and contracting power and water. Its private data-centre funds can take up the long-term leases later. The Melbourne agreement expanded its APAC powerbank to over 1 GW. Keppel has said its growing powerbank has the potential to add approximately S$10 billion to its data-centre funds under management. [source]

The sequencing is what interests me. Keppel is securing the inputs required to make future capacity deliverable before committing the full capital required for the eventual facility. Part of the asset-formation process is moving upstream. Institutional capital is paying attention to that position too: when Stonepeak agreed to invest US$1.3 billion of preferred equity in Princeton Digital Group, the powerbank entered Stonepeak's stated investment rationale alongside execution and management. [source]

That does not tell us how much PDG's powerbank itself was worth. The transaction does not separately price secured power from operating assets, customers, management or future development potential. But it establishes something narrower and important: secured future power capacity has become part of the investment proposition of an APAC data-centre platform.

There is no universal power right. Different markets put the critical control point in different places.

In Singapore, EDB and IMDA provisionally allocated 200 MW of new capacity to four operators through DC-CFA2 — a position that only four of more than 20 proposals secured. [source] New South Wales works differently: Transgrid allocates network capacity for relevant large-load projects on the signing of a Network Connection Agreement (NCA), and even its conditional route requires applicants to demonstrate meaningful progress against planning criteria relating to permits, land ownership, construction, long-lead equipment and financing. [source] Owning land therefore does not automatically mean controlling a usable grid position. Malaysia is different again: TNB's Green Lane Pathway (GLP) and One-Stop Centre form a critical utility-delivery pathway. TNB says the fast-track process can reduce the conventional 36–48 month connection period to as little as 12 months. [source]

The Transgrid process surfaces something deeper than a simple gate. Read as an underwriting progression, the process moves roughly from application → readiness evidence → conditional agreement → capacity position → firmer delivery pathway. That interpretation is my own analytical lens; the underlying mechanism is Transgrid's published policy. [source] But what the progression illustrates is worth noting: the position is real at every stage; what it actually represents changes. A project can progressively strengthen its position as development conditions are cleared.

Instead of treating every announced megawatt equally, I would separate them by state.

This is not an industry accounting convention. It is an underwriting lens. And even Controlled MW are not binary: as the Transgrid progression illustrates, a project can move through stages within the Controlled state — each clearing a different condition, each removing a different uncertainty.

Land can become powered land. A power application can become a network position. A network position can become deliverable capacity. Deliverable capacity can become commissioned capacity. And commissioned capacity can become revenue. Each transition removes a different uncertainty — and each transition has a control point.

The strongest development position in a power-constrained market may not belong to the platform with the most announced capacity. It may belong to the one that controls the critical transitions — provided it has not overpaid for that control.