Investment Thesis
Demand is accelerating.Supply is constrained.
Australia will need several times its current data centre capacity within a decade. New supply is gated by land, grid connections and planning, not by capital. APDC exists to resolve those bottlenecks early, when doing so is cheapest and worth the most.
Australia is short of the one thing AI can't do without
Most Australian data centres were built for a different era of computing. They were designed around cloud workloads at a few kilowatts per rack; AI training and inference clusters ask for many times that. Little of the existing fleet can be retrofitted economically, and the new capacity now under construction is largely spoken for before it opens.
Sydney and Melbourne, where the industry has historically concentrated, are running out of grid headroom. The United States hit the same wall and responded by building regionally — over half of new hyperscale capacity there now lands outside the major metros. Australia is at the start of the same shift, and it favours whoever holds the right regional ground.
24×
Projected growth in AI agent token demand by 2030
Goldman Sachs Research, 2026
1.3%
US data centre vacancy rate
Blackstone SEC Filing, 2026
7GW+
Australian capacity required by 2035, up from 1.35GW today
2×
Market rents have more than doubled in four years
Blackstone SEC Filing, 2026
Australian data centre investment is forecast at $85bn–$135bn by 2035, with the sector potentially accounting for up to 11% of the country's electricity consumption, compared with around 1% today. Meanwhile, Microsoft, Amazon, Google and Meta are collectively committing hundreds of billions of dollars each year to data centres and AI infrastructure.
Three bottlenecks, none of them quick
Capital is not the constraint on new supply. The constraint is the small number of sites where land, power and approvals line up — and each of those takes years, not cheques, to assemble.
01
Land
Sites near fibre routes, substations and demand centres are finite. The obvious ones are gone; the good ones that remain are contested.
02
Power
Grid connection queues run in years, not months. A site with a secured connection is worth a multiple of the paddock next door.
03
Approvals
Planning and environmental processes cannot be bought forward with capital. Time invested early becomes embedded value later.
We do the slow, unglamorous work first
Our capital goes into the phase where value is actually created: turning raw land into a powered, approved project with a committed customer.
01
Capital follows milestones, not the other way around
Funding is staged against site control, a power pathway, planning approval and end-user commitment. Construction capital doesn't move until development risk has been retired.
02
Demand is contracted, not assumed
Projects are anchored to end-user commitments before major capital commits. We don't build on speculation.
03
Two ways to win on every project
A de-risked project can be sold to a builder-operator, or held into stabilised earnings. Which path we take depends on the market at the time, not a plan written years earlier.
04
No single project defines the outcome
A portfolio spread across markets and stages means one approval decision, one connection queue or one negotiation never carries the whole result.
Test the thesis against your own view
We welcome confidential conversations with investors weighing exposure to Australian digital infrastructure.