Edition 22 ended on a hard question: if a durable placement constraint does not automatically hand its premium to whoever sits inside the border, is the winning asset the data centre — or the control point wrapped around it?
This edition takes up a control point that framework left implicit:
Time.
A megawatt isn't fully described by where it is or what it costs. It also has a date.
1 The deal
In Malaysia, TNB (Tenaga Nasional Berhad) says its Green Lane Pathway can connect data centres three times faster than the conventional process — reducing an implementation period of 36–48 months to as little as 12 months. [source]
That is potentially two to three years of difference before power reaches the site.
For an investor, those years are not simply a construction-schedule detail.
They can determine when a site is energised, when it reaches RFS (ready for service), when customers can deploy — and ultimately when the asset begins generating cash flow.
So there is another distinction worth making:
MW secured ≠ MW equally valuable.
Two projects may each eventually receive the same 100 MW. But if one turns that power into commercially usable capacity years earlier, they are not the same economic asset.
In a constrained market, power in the abstract is not enough.
What matters is power on a date you can underwrite.
And even the date is incomplete.
A promised energisation date backed by delivered grid infrastructure is not the same asset as one still dependent on future transmission works.
Nor does energisation itself mean the site is ready for customer operation.
Testing and commissioning still sit between power availability and customer-ready capacity.
The investment question therefore becomes:
When does the MW arrive? With what probability? When does it become commercially usable? What happens if that date slips? And who captures the value of arriving earlier?
2 The engineering read
“Power secured” is an incomplete description.
A megawatt travels a delivery chain before it can earn.
For time-to-cash-flow, the relevant sequence is:
→ grid infrastructure
→ energisation
→ testing & commissioning
→ RFS
→ customer deployment
→ revenue commencement
This is a time-to-revenue sequence, distinct from Edition 14's bankability conversion.
Each transition takes time. Each can become part of the critical path. And each carries a different probability of delivery.
That gives a useful distinction:
Power on time ≠ site delivered on time ≠ cash flow on time.
A utility may energise a site according to schedule while testing and commissioning still delay customer-ready capacity.
Once a delivery date enters a customer contract, that slippage can move beyond engineering.
It can become a cash-flow event.
I think about this through four clocks:
Grid clock — when can the power actually be delivered?
Engineering clock — when can infrastructure, testing and commissioning produce customer-ready capacity?
Contract clock — when has that capacity been promised to the customer, and what happens if delivery slips?
Cash-flow clock — when can service and revenue actually commence, and where do the economics of delay ultimately land?
The four clocks do not necessarily move together.
Australia shows why the first one alone needs careful underwriting.
At the end of March 2026, AEMO (Australian Energy Market Operator) reported 11 large data-centre projects above 5 MW, representing 5.4 GW of maximum demand, progressing through the transmission connection process.
AEMO reports that current experience indicates large data-centre connections are targeting approximately two years from application to energisation, although actual timing varies with project readiness and power-system conditions. [source]
But entering a connection process is not the same thing as having deliverable power.
Transgrid makes that distinction particularly clear.
In August 2026, it said Sydney's existing transmission network had insufficient capacity to connect any of the 20 GW of new large-load connection enquiries it had received. [source]
Targeted customer-funded network investments could unlock up to another 2 GW, while approximately 1.5 GW had already been allocated to signed data-centre connection agreements in Western Sydney.
So demand for connection is not deliverable capacity.
And even a stated delivery date is not enough.
Two projects may both claim RFS in the same year — one with a secured connection path and infrastructure progressing, another still dependent on future augmentation.
Same stated date. Very different distributions of delivery outcomes.
That changes the underwriting progression.
Instead of asking only:
How many MW?
I would ask:
Delivery-path underwriting pass
- What exactly has been secured?
- When is energisation expected?
- What remains on the critical path?
- What remains between energisation and RFS?
- With what probability will RFS occur on the committed date?
- And what happens economically if it doesn't?
Because ultimately:
The engineering schedule becomes the cash-flow schedule.
3 The capital allocation read
1. The market is already valuing readiness.
Cushman & Wakefield's Asia Pacific Data Centre Group says infrastructure-ready land is commanding a clear premium. [source]
In power-constrained markets, immediate utility access can shorten delivery timelines and reduce execution risk.
That does not tell us what one year of earlier power is worth.
Infrastructure-ready land bundles:
power + utilities + approvals + execution certainty + time.
So a standalone “price of time” cannot be extracted from the observed premium.
But the narrower conclusion is defensible:
Solving critical infrastructure earlier has economic value.
2. Capital is already being committed to preserve time.
Keppel has secured rights to lease a 123-hectare site near Morwell, Victoria, with up to 720 MW of gross power capacity, delivered in phases.
Under its Agreement for Lease, Keppel pays an annual access fee for early access to undertake pre-development works — including planning approvals and contracting power and water — before its private data-centre funds take up long-term leases. [source]
The fee is not a quoted “price of time”.
It also buys site access and optionality.
But the behaviour is revealing:
Capital is being committed today to preserve the ability to move earlier tomorrow.
A powerbank is therefore more than a stock of future MW.
It can preserve an option on when those MW become commercially usable.
3. But faster is only better if the head start survives the price paid for it.
Suppose two otherwise identical sites eventually support the same capacity.
Site A: cheaper land, power in four years.
Site B: more expensive land, power in one year.
Site B has the operational advantage.
But that does not automatically make it the better investment.
The missing variable is:
What did the investor pay for those three years?
A simple DCF (discounted cash flow) illustrates the point.
At a 10% required return, an identical cash flow received two years later has a present-value factor of:
or approximately 17.4% lower present value today, all else equal.
That is an illustration of timing — not a claim that a data centre delivered two years later is automatically worth 17.4% less.
The underwriting test is therefore:
>
premium paid for earlier readiness?
Those benefits may include:
earlier potential cash flow + avoided carrying time + reduced execution exposure + commercial optionality.
If the seller has already captured the value of earlier readiness in the entry price, the buyer has acquired a scarce asset — not necessarily an excess return.
4. The scarce fast slot can be a control point.
Transgrid provides an unusually clear example.
Approximately 1.5 GW of capacity has been allocated to signed data-centre connection agreements in Western Sydney.
Additional large-scale connections require transmission augmentation, with relevant costs borne by the proponents creating the demand.
And because Transgrid expects demand for available capacity to continue exceeding supply, allocation is based on the order in which customers sign connection agreements and commit to funding the required transmission augmentation.
The scarce resource is therefore not only electricity.
It can also be a position in the delivery sequence.
Earlier access may require capital commitment before the MW is usable.
And the investor only keeps the resulting advantage if enough of that fast position remains under its control after paying for it.
Scarcity can exist.
Time can have value.
But the return still depends on who controls the bottleneck — and what they paid for that control.
5. Delivery risk doesn't disappear — it gets allocated.
Once a delivery date enters a customer contract, a slip can acquire a price.
One US data-centre services agreement filed by Hyperscale Data in 2026 provides a concrete example.
Under specified delay conditions, service commencement moves to actual delivery. Qualifying delay beyond 90 days gives the customer credits equal to 25% of recurring service charges for the delayed phase; beyond 180 days, the credit rises to 50% for the continuing delay. [source]
Those percentages are contract-specific.
They are not an industry benchmark.
But they show the mechanism:
→ delivery / RFS slippage
→ later service commencement
→ cash-flow consequence
The exposure can then migrate.
Linesight notes that long-lead-equipment delays can threaten tenant RFS dates and identifies liquidated-damages provisions in vendor contracts as one way of protecting critical milestones. [source]
Marsh documents an Asian powered-core-and-shell data-centre project where the owner sought to transfer tenant-facing delay exposure to the contractor. The resulting daily LD (liquidated damages) exposure and cap exceeded the contractor's balance-sheet tolerance and internal financial limits, creating demand for insurance risk transfer. [source]
But allocation is not elimination.
Extensions of time may excuse certain delays.
Liability caps can limit downstream recovery.
Insurance responds according to its own terms.
And what the owner owes its customer may differ from what it can recover elsewhere.
That residual mismatch is retained delivery risk.
So the question is not only whether the project reaches RFS on time.
Who is left holding the schedule risk when the critical path fails?
6. A lead-time advantage is not automatically a moat.
I see at least three states.
Developer advantage — secure the infrastructure path early, de-risk the project and monetise the readiness through development, partnership or sale.
Temporary window — earlier access is valuable today, but the advantage narrows as utilities add capacity, connection processes accelerate or competitors secure alternatives.
Durable owner advantage — the owner retains control over an interconnection, site or infrastructure position that remains scarce and difficult to replicate.
TNB's Green Lane itself demonstrates that lead times can be compressed.
But the counterforce matters too.
New transmission can take years. Demand can grow faster than network capacity. And solving one constraint can simply move the critical path into substations, equipment, approvals, commissioning or another infrastructure layer.
So speed-to-power should not be assumed to be a moat.
Its durability has to be underwritten.
House view
I would therefore think about the economic value of a megawatt as:
MW,
delivery date,
delivery probability,
commercial readiness,
customer time sensitivity,
retained delivery risk
)
This is an analytical framework, not an industry metric.
The first five variables describe the capacity and its commercial usefulness.
The last asks how much downside still sits with the owner if the delivery path fails.
A promised energisation date is not the same as a committed RFS date.
And a committed RFS date is only as valuable as the probability of completing the infrastructure, testing and commissioning required to meet it.
Two projects can therefore have:
the same MW,
the same target year,
and very different economic risk.
Edition 22 established:
Location determines which workloads a megawatt can serve.
Edition 23 adds:
Time determines when it can serve them — and the delivery path determines how confidently that date can be underwritten.
Investment Lens
4 What this means for the broader market
A MW arriving in 2027 and a MW arriving in 2030 are not necessarily the same economic asset.
But even that comparison is incomplete.
A megawatt has more than a capacity, location and date. It has a delivery path — and someone ultimately owns the risk that the path fails.
That changes the underwriting arc:
→ By when?
→ With what delivery probability?
→ When does the site become commercially usable?
→ What happens if the committed date slips?
→ Who retains that exposure?
→ At what premium was earlier readiness acquired?
→ Who ultimately captures the benefit?
Earlier access can create value.
But whether it becomes a durable owner return depends on whether the capacity arrives when promised, becomes commercially usable when required, survives the price paid to obtain it, and leaves enough of the resulting advantage with the owner.
A desirable infrastructure characteristic does not automatically become an investor return.
Scarcity is not enough.
Location is not enough.
And speed is not enough.
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?