Highlights
- A global property group has pivoted much of its development pipeline toward data centres.
- Securing land, power and planning approvals has become its AI-era edge.
- Data centres now dominate a pipeline stretching to a substantial figure.
Not every artificial-intelligence beneficiary writes software or builds chips; some own the dirt beneath it. Goodman Group (ASX:GMG), a global industrial property and development group and a heavyweight of the ASX 200, has increasingly been cast as the landlord of the AI build-out. Data centres now make up the lion's share of its development pipeline, a pivot that has reframed a logistics-property name as a central player in the physical infrastructure powering artificial intelligence.
Owning the scarce inputs
The bottleneck in data centre delivery is rarely the building itself; it is the ingredients that go into siting one. Suitable land close to power and fibre, grid connections large enough to feed a hyperscale campus, and planning approvals that can take years to secure all have to be assembled before a single server is racked, and whoever controls that combination commands real leverage. A group with deep experience acquiring land, navigating approvals, connecting power and delivering built form is well placed to supply precisely what the largest technology customers need, when they need it.
That capability sits at the crux of the story. As demand for compute-ready space runs ahead of supply, value migrates toward those who can actually unlock new capacity, and land-and-power assembly is the hardest link to shortcut. The pivot leans deliberately into that scarcity, positioning the group not as a passive owner of buildings but as a developer able to conjure capacity where others cannot, a materially more valuable role.
The advantage is compounded by an existing global footprint. Industrial estates assembled over years for logistics often sit exactly where data centres now want to be, near population centres, transport corridors and power infrastructure, giving the group a bank of well-located land that would be almost impossible to replicate today. Repurposing and expanding that estate for compute is a faster path to capacity than starting cold, letting the group move at the pace the theme demands.
A pipeline reshaped around data centres
The shift is visible in the composition of the development pipeline, where data centres have come to dominate a book that has swelled to a substantial scale. That concentration marks a deliberate tilt toward the artificial-intelligence theme, converting a broad industrial-property engine into one increasingly geared to compute infrastructure. It is a considered wager on the durability of data centre demand, and on the group's ability to convert land and approvals into leased, income-producing capacity.
The reshaping also changes the group's earnings character. Data centre developments tend to be larger, longer-dated and more capital-hungry than conventional sheds, with customers who sign lengthy leases and demand exacting specifications around power and cooling. That lifts the value of each project but raises the stakes on execution, since a handful of large developments now carry weight once spread across many smaller ones. Managing that concentration, funding it, staging it and leasing it, becomes central to how the strategy plays out.
Coverage of ASX AI Stocks has widened to include the property and power layers, where owning scarce land and grid connections is as pivotal to the build-out as the servers themselves.
How the power constraint shapes the strategy
Power is emerging as the decisive variable in where and how fast the pipeline can be built. A hyperscale campus draws electricity on a scale that can rival a small town, and the queues to connect large new loads to the grid have lengthened as demand has surged. That reality hands an advantage to a developer that already controls land with existing or securable connections, and pushes the group toward sites where generation, transmission and approvals can be lined up together rather than fought for separately. Access to renewable supply matters too, since the largest technology customers carry their own clean-energy commitments and increasingly insist their capacity be powered accordingly. Solving the power equation ahead of rivals is therefore not a side issue but a core part of the edge, and it helps explain why land near substations and firm energy has become some of the most contested ground in the market.
Recurring rent versus development risk
The model blends development with landlordship, and the balance between the two defines its appeal. Building and then leasing data centre space can generate long-dated rental income underpinned by creditworthy tenants, an income stream prized for its stability and its insulation from the shorter cycles that buffet other property. Yet development at scale carries distinct risks: construction cost inflation, lengthy delivery timelines, and the standing assumption that demand will persist long enough to fill the halls at the rents underwritten. The group is, in effect, taking development risk today in exchange for recurring income tomorrow, and the quality of that trade depends on delivering on budget and on time into a market that stays as tight as it looks now.
What could unsettle the thesis
The pivot's success rests on several assumptions staying true at once. Data centre demand must remain firm, power must stay securable at a workable cost, planning regimes must keep granting approvals, and the group must execute a large, concentrated pipeline without stumbles. A cooling in artificial-intelligence capital spending, a tightening in power markets, community resistance to large developments, or delivery slippage would each test the tilt, and because the pipeline now leans so heavily on one theme, a setback would land with more force than against a diversified book. For now, controlling scarce land and energy looks like a durable edge few can match, but the very concentration that makes the story compelling also raises the stakes on the theme continuing to play out.