The barriers are not singular. They are systemic and compounding.
Investors are not avoiding data centres—quite the opposite—however, they are pricing in risk. The constraints are interconnected and addressing them in isolation won't be sufficient.
1. Coordination of risk
Data centres sit at the intersection of energy, planning, construction and technology, yet no single framework coordinates delivery across these domains. The result is a fragmented risk profile that is difficult for investors to price with confidence.
Unlike traditional infrastructure, data centres follow a developer-led model. As such, developers, not contractors, drive outcomes, bringing more extensive design control, technical expertise and supply chain relationships. This shifts both risk and execution capability into specialised hands, creating a higher-risk, higher-control delivery model that does not fit neatly within established procurement frameworks.
What needs to change: The priority is not solving individual issues in isolation but improving system-wide coordination. This requires better alignment across energy (generation and networks), data centre development and water infrastructure; streamlined and more consistent planning frameworks; and targeted investment to relieve key bottlenecks. Victoria's establishment of a dedicated taskforce to support proponents through the approvals process and New South Wales's move in a similar direction are positive steps. We need more of this coordination nationally.
2. Grid connection certainty
Grid certainty is foundational to data centre investment. Developers are reluctant to commit time or capital until they have reasonable confidence that sufficient capacity will be available, at the right location, in the right configuration, within a workable timeframe and at a supportable cost.
Australia's connection framework was not designed with multiple large load-side applicants in mind. Assessment timelines are lengthy. Grid capacity doesn't always align with the urban locations data centres require, and the connection process rules don't give visibility over the status of competing applications.
This means two things: grid connection is inherently complex and may take longer than developers anticipate; and developments may face unforeseen constraints, or greater competition for shared network infrastructure than originally anticipated.
The Australian Energy Regulator has acknowledged that current frameworks weren't built for this type of demand. Reform that establishes faster, more predictable timelines is needed, not only for data centres but across the energy transition. However, reform that trades speed for system security will ultimately serve no one well.
What needs to be done:
- Improve transparency on capacity, so developers can make informed decisions about connection timing and site selection based on the status of existing applications.
- Fast-track applications for data centres that bring their own generation and system support solutions.
- Consider conditional capacity reservation for projects that have demonstrated a genuine commitment to proceed—eg by satisfying planning or tenure milestones within agreed timeframes.
Case study: Keppel
720MW data centre campus, Gippsland
Allens advised Keppel, a Singapore-headquartered global asset manager operating more than 39 data centres across Asia-Pacific and Europe, on its proposed development of a 720MW data centre campus on a 123-hectare site near Victoria's former Hazelwood power station.
The site's location within the proposed Gippsland Renewable Energy Zone, one of the state's largest electricity nodes, provides direct access to power, water infrastructure and intercity dark fibre networks, with the potential for a dedicated transmission connection to neighbouring terminal stations. This is a clear example of energy-proximate site selection enabling large-scale digital infrastructure development.
3. New generation to meet growing demand
The first half of 2026 has seen a notable flattening of the electricity spot market. Favourable weather conditions have allowed wind and solar to maximise output, new battery capacity is smoothing evening peaks, and the uptake of home batteries is moderating midday price troughs. While welcome developments for consumers, this is dampening investor appetite for the new generation and firming capacity the grid will need when anticipated data centre demand materialises. If that demand arrives in proximity to the exit of one or more coal-fired power stations, the spot market will face a period of significant volatility.
The generation needed to meet data centre demand can't be built overnight. A data centre may take one to three years from conception to operation. A wind farm, or portfolio of wind farms, large enough to support that facility's demand may take seven to ten years. Rising wind energy costs have further stalled a number of projects, creating a mismatch between offtakers' and developers' price expectations. These factors make it unrealistic for data centres to bring their own generation from day one.
What needs to be done:
- Incentivise long-term offtake commitments that provide sufficient revenue certainty to support investment in new generation and firming capacity. Where a data centre, or data centre customer, can demonstrate it has executed a long-term offtake with a new renewable facility scheduled to commence within a reasonable period of the data centre beginning operation, that commitment could be used to streamline planning and connection pathways.
- Strengthen the Renewable Electricity Guarantee of Origin (REGO) framework. While currently voluntary, there is a strong case for requiring data centres, or major data centre customers, to participate, including through time-stamped matching of renewable generation to load. This would give operators a credible basis for clean energy claims, stimulate demand for REGOs and provide an additional revenue stream for renewable generators.
- Deploy government-backed contract-for-difference mechanisms to reduce the risk premium that generation investors currently require. This aligns with the Electricity Services Entry Mechanism framework objectives being pursued under the Nelson Review, and data centre demand may provide further impetus to accelerate that work.
The current period of market stability is unlikely to persist. The time to put the right contracting and investment frameworks in place is now, not when the market is already under pressure.
4. Planning and approvals
Planning and approval frameworks vary markedly across states and territories, making some jurisdictions materially harder to navigate than others. Current approval processes for data centres are often based on standard industrial or commercial development categories. In these cases, the planning controls against which data centre applications are assessed are poorly suited to an emerging digital infrastructure asset class, in terms of environmental assessment requirements, the calculation of infrastructure contribution charges, and the need to augment services infrastructure to meet energy and water demands.
The key issues are clear:
- Planning and zoning barriers Controls that are ill suited to data centre developments, slow and uncertain approval processes, and differing state approaches make it difficult to secure timely approvals.
- Power and water approval constraints Grid upgrades and water access often require off-site works on third-party land, necessitating engagement with third-party land owners, external agencies and utility authorities, and triggering complex approvals and environmental assessments that can seriously delay projects.
- Misaligned contribution regimes Existing infrastructure contribution frameworks are not suited to data centres, often overcharging assets that are capital intensive but not employment dense.
Although New South Wales data centre developments benefit from a state-assessed approval pathway, proponents must still navigate local planning controls, and often engage in protracted negotiations with councils and utilities authorities to secure the approvals needed to construct and connect to essential services, frequently causing considerable delay in construction.
What needs to be done:
- Reform planning frameworks to explicitly recognise data centres as essential infrastructure.
- Enable fast-track approvals in suitable zones.
- Streamline off-site utility upgrades.
- Encourage low-water cooling technologies.
- Adopt fit-for-purpose contribution models tailored to digital infrastructure.
5. Social licence, community engagement and ESG considerations
Overlaying all of the above is the need to ensure Australian data centre investment doesn't lose its social licence, as has occurred in parts of the US.
The risks are real and growing:
- Perceived environmental footprint Media reports that hyperscale facilities will consume enormous amounts of power and millions of litres of water annually are shaping community attitudes. The International Energy Agency's projections on data centre energy consumption have intensified this scrutiny.1
- Regulatory inquiry A Senate inquiry into AI and data centres—examining the effectiveness of existing regulatory frameworks, and the potential impacts on communities, industries and the environment—was established on 13 May 2025.
- Community pushback Local opposition to developments in Lane Cove (Sydney), Footscray (Melbourne) and Hazelmere (Perth) reflects a small but potentially growing resistance to data centre construction in urban and suburban areas.
If broad community sentiment turns, the result will be substantial development risk, through complaints, activism, and litigation in the form of development challenges and tortious claims. This will stymie investment at precisely the moment it's needed most.
Poorly handled ESG amplifies every other barrier. Handled well and proactively, though, it can protect social licence and underpin the major investment program ahead. Data centre investors, developers, operators and tenants should treat community engagement as an ongoing program, not a one-off exercise, and stress-test ESG management to ensure it aligns with disclosures and reporting obligations.
Footnotes
-
IAE, 2026.