Divide Westpac IQ's $155 billion by the 400,000 jobs its report attaches to the number and you get about $387,500 of capital per job supported [2][3][1]; set the roughly $75 billion net GDP boost against the same capex and the pipeline pays back about 48 cents of national income per dollar spent [4][2]. Those are modelling outputs rather than signed leases, and the material carrying them does not say whether the $155 billion is Australian or US dollars, though it is careful to mark the global figures Deanne Stewart cited as US$750 billion of current allocations and a projected US$1 trillion next year [8][14]. Taken at face value, that makes the Australian pipeline roughly a fifth of the capital already allocated to the sector worldwide [5], which is either a serious claim on global capex or a number waiting for a denominator.
The part with a date attached is the load. Data centres go from 3% of Australian electricity to 13% by 2035-36, about 4.3 times the share [5][3], and the Australian Energy Market Operator's warning is about sequencing rather than volume: the demand arrives before the transmission does, and consumers cover the difference [6]. That is the mechanism to follow, because grid connection is queue time, and queue time is the one input an allocator earning more than 20 percent per annum cannot outbid [9].
So the siting decision moves. Lucinda Jerogin of CommBank says proposals are broadening into the Northern Territory and South Australia, where electricity and grid constraints are less severe [13], and Equinix's Tim Robinson says land cost is now pushing facilities outward from the city fringe over time [12]. What that forecloses is the metro build: capex sited by where the electrons already are rather than where the customers and the fibre sit, and jobs landing in states with thinner networks to reinforce.
If transmission gets built to the 2035-36 path, the $155 billion is a floor rather than a target, because AirTrunk (which sells capacity, so weigh it accordingly) says hyperscaler appetite over the last few months already runs ahead of what anyone expected [7]. Migration to the less constrained states solves power and hands the binding constraint to water and suitable sites, which Jerogin lists alongside grid connections in what decides which projects proceed [13]. And a moratorium out of the Tasmanian inquiry, or traction for the New South Wales freeze campaign, slows the pipeline without changing any of the economics [11].
On this evidence the $155 billion is a conditional ceiling, and Stewart's own framing concedes as much, since she put the unlock explicitly behind grid connections, zoning approvals, construction costs and suitable sites [1]; the counter, which she also makes, is that power and grid constraints are not putting investors off and that policy consistency matters more than either [10]. What would settle it is dated connection capacity. Until a network operator publishes some, 13% of the national grid is an assumption about somebody else's construction schedule.
