There is something idyllic in the notion that our data is sitting in a cloud waiting to beam down on demand, whenever needed. Of course, the reality is that whenever data moves from a device to a cloud-enabled service, it lands in a data centre, and the capacity of those data centres is a key enabler to the use of AI in Australia and elsewhere.

Below, we unpack what the Australian Government's proposed energy and network policies mean for the cost of building and running data centres and the questions they leave unanswered for developers, operators and their energy partners.

Matching of data consumption with in-jurisdiction renewable energy and firming capacity

On 15 July 2026, announcing a new AI framework for Australia, the Prime Minister stated:

'The new standards will set out clear rules for large data centres - including a legal obligation to underwrite their own new power supply, pay their full share of connection costs so energy bills are not impacted, reduce power when needed to strengthen the grid, and be as water efficient as possible'.

Less than two weeks later, the Energy and Climate Change Ministerial Council (ECMC) (comprising the Australian Federal, State and Territory Energy Ministers) stated that it:

'…supported, with [Queensland and the Northern Territory] opposing, the Commonwealth's commitment to legislate nationally consistent standards to ensure the next generation of large scale data centres underwrite new renewable power supply, and consumers face no price impacts from data centres entering the grids... They agreed to progress regulatory arrangements to mandate that data centres offset their electricity demand by investing in additional renewable generation located in the jurisdiction where the data centre is located unless the jurisdiction opts out of this requirement …'.

The Commonwealth does not have primary jurisdiction over electricity in Australia. The National Electricity Market (NEM) is instead governed by uniform legislation (the National Electricity Law) collectively adopted by the ACT, New South Wales, Queensland, South Australia, Tasmania and Victoria.

However, with not all NEM jurisdictions being in favour of the policy, the Commonwealth Government initially appeared determined to exercise all levers within its control to achieve a national approach. The position has since softened. At the National Cabinet meeting on 26 August 2026, it was noted that the 'Commonwealth will work with state and territory governments to develop consistent mandatory standards for data centre energy, water and land-use, and support skills and training opportunities'.

Two days later, the Federal Energy Minister clarified his Government's position:

'.. we are legislating nationally consistent standards across the board with no exceptions and no carve-outs. The only change is that where a state owned government electricity company asserts that they can do it cheaper than renewables, they can apply to the Commonwealth to do that' (ABC News, 28 August 2026).

The net position? No blanket carve-outs, but individual jurisdictions may apply to rely on cheaper existing generation – including fossil fuel supply – instead of investing in new renewables.

What jurisdictional matching means in practice

In jurisdictions where it is adopted, the requirement that data centres match their energy consumption with renewable energy located within the same jurisdiction is a significant policy shift. Data centres currently have full flexibility to procure power via wholesale or retail arrangements, choosing what level of consumption is met by renewables. The source of that power need not be co-located with the data centre (most are not), nor in the same State or Territory.

To implement jurisdictional matching, changes are proposed to be made to the Renewable Electricity Guarantee of Origin (REGO) scheme, administered by the Clean Energy Regulator under the Future Made in Australia Act 2024.

The REGO scheme is presently voluntary, but the Australian Energy Market Commission (AEMC) has advised the ECMC on regulatory options to require data centres to offset consumption with REGO certificates from new renewable generators.

Separately, the AEMC proposes a contract obligation requiring data centres to demonstrate their power demand is covered by firmed capacity through financial contracts and investment in storage and demand flexibility, with new market registration requirements flagged as the implementation mechanism.

This raises several questions:

• Will a data centre's contract obligation apply to its maximum demand at all times, or only in relation to forecast generation shortfalls on the system?
• How will penalties (financial or otherwise such as load shedding) apply for a failure to meet the contract obligation?
• Will the REGO offset obligations require that REGOs for renewable generation and a data centre's energy consumption be time-matched and, if so, on what granularity (hourly, peak/off-peak, weekly)?
• Will other obligations be imposed on market registered data centres such as forecasting and dispatch obligations for their electricity load?

Data centres are already procuring renewable energy to the extent it is available at an acceptable cost. Any increased costs from a jurisdictional matching requirement will, unless absorbed within existing margins, be passed on to data centre customers.

Early-stage project developers seeking to contract output and secure financing will need to understand how the REGO matching and firm contract requirements apply, so their project agreements appropriately allocate responsibility.

Power grid impact – network costs and system security

Electricity purchase costs currently make up 50-60% of an industrial customer's energy bill, with network charges accounting for the remaining 40-50%, a share expected to grow as the energy transition drives further network build-out.

Network pricing reforms: Data centre energy demand is driving the need for new network investment. The Federal Energy Minister has formally commenced a process to amend the National Electricity Rules to rewrite how shared network expansions are funded and how large users are charged for existing shared infrastructure. The aim: data centres and other large users underwrite network infrastructure costs so they are not recovered from other consumers.

State-level action

Some States are not waiting for a national consensus:

• The New South Wales Parliament introduced a Bill on 5 August requiring proponents of data centres and other 'large load infrastructure' to pay network infrastructure costs, ensuring other NSW consumers do not bear those costs. The Bill (if implemented) will enable NSW to derogate from standard electricity network pricing and access rules under the National Electricity Rules. The NSW Government has released a detailed consultation paper on 17 August 2026 proposing a framework that seeks to ensure that large data centres internalise costs and system impacts by funding necessary network and renewable energy infrastructure while preserving power system security and avoiding cost transfers to other consumers.
• In South Australia, a Royal Commission on AI has been established, with a final report due by 1 July 2027. The Commission is expected to consider how AI is transforming critical industries and the governance frameworks needed for responsible adoption. Its findings are likely to influence both South Australian and national approaches to AI regulation.
• In Victoria, the State Government has released a Sustainable Data Centre Action Plan built around five pillars: coordinated infrastructure, land and connectivity; energy innovation and grid readiness; sustainable water security and management; a skilled, future-ready workforce; and strengthened leadership and coordination.

Network access and system security reforms: Large electricity users on the transmission grid must already meet detailed system access standards, but those standards are less stringent for distribution-connected users – which includes most data centres in Australia. That is expected to change. Data centres can disconnect from the grid and switch to on-site back-up generation far faster than other users, and when this occurs at scale it can 'amplify frequency and voltage disturbances, trigger cascading events and risk system security' (Rainer Korte, AEMC Commissioner, 'Data | Power: Navigating policy, regulation and networks' 28 April 2026).

The AEMC is currently consulting on proposed reforms that would require data centres to 'ride through' system disturbances rather than disconnect and switch to back-up supply and exacerbate what may become a system security event.

These reforms align with the ECMC's goal that 'consumers face no price impacts from data centres entering the grids'. Increased costs will still translate into higher prices – but the Government’s position is that those costs are best borne by data centres and their customers, not the broader consumer base.

Data centre developers will want change-in-law clauses in their project agreements that allow pass-through of increased compliance costs. Grandfathering and transitional provisions will also be critical.

The ‘ride-through' requirements also have implications beyond the grid. Data centre operators typically make service availability commitments – uptime guarantees, latency tolerances, failover response times. A requirement to ride through disturbances rather than switch to back-up generation may, depending on the final standards, change the assumptions underpinning those commitments. Developers and operators will need to confirm they can still meet them.

Co-location of data centres with renewable generation, batteries and demand response

To date, Australian data centres have largely been located in urban and suburban areas. However, larger facilities are now targeting rural sites for data that does not need to be stored close to end-users.

Co-locating data centres with renewable generation or batteries – either behind the meter or in adjacent locations – favours rural sites. A co-located battery providing ‘ride-through' capability during system disturbances could keep a data centre grid-compliant while meeting its service availability commitments in a way a grid-only solution may not.

If grid connection and use costs can be shared between a data centre and a renewable energy or battery developer (or avoided with a partial off-grid solution), material savings are achievable. However, this requires significant upfront joint planning, suitable tenure and approvals, and agreement on structure, cost sharing and development responsibilities.

The proposed renewable matching and firming obligation is likely to make co-location not merely an attractive option but a practical pathway to compliance, reinforcing the case for early joint planning between data centre and energy developers.

Several practical questions remain:

• It is presently unclear whether the proposed renewable matching and firming obligation will apply to the data centre owner, the operator or someone else in the supply chain. In a co-location arrangement, this will influence which party (ie the data centre developer or its energy counterpart) is responsible for ensuring sufficient new generation and firming capacity is delivered and maintained.
• Power purchase agreements for generation from existing renewable energy facilities are unlikely to satisfy the co-location obligation. The requirement contemplates data centre developers underwriting new generation and firming capacity, which means that co-location arrangements will need to be structured around new-build assets rather than existing supply.
• Where battery storage is co-located to provide firming, the question of whether the battery can also participate in energy and other markets (eg to provide frequency control ancillary services or wholesale arbitrage) independently of the data centre load will need to be agreed between the parties. Such market participation may improve the economics of the co-located solution but may also introduce complexity around dispatch priority and the asset's primary compliance function.
• The accounting methodology for determining whether the matching obligation has been met, including the relevant measurement period and the size threshold for what constitutes a 'large data centre', remains to be settled. If the methodology changes after co-located assets have been sized and commissioned, then one or both parties may need to augment their capacity, and their development framework should anticipate that possibility.

Further co-location reforms may be forthcoming. In July, the AEMC recommended to the ECMC a streamlined connection process for data centres co-locating with existing renewable and firming, connecting with new firming, or providing demand flexibility. Consistent with this position, at the National Press Club in early August, the Federal Energy Minister stated:

'we are encouraging other measures to strengthen the grid, including offering data centres fast-tracked options where they can co-locate with existing generation, minimising the need for new transmission and distribution'.

Plugged in on paper, but does it translate to a connection?

While recent NSW, national and Commonwealth policies seek to address the growing electricity demands of data centres through measures relating to network cost allocation, connection processes and renewable energy expectations, important practical challenges remain. In particular, the ability to secure a timely and reliable connection will continue to be influenced by underlying transmission and distribution network constraints, including capacity, system strength, fault level and other technical requirements.

In addition, the increasing prevalence of co-located generation and battery assets raises more complex questions regarding connection point configuration, access arrangements, import and export capability, and the interface between private and regulated network infrastructure. The regulatory framework must therefore consider not only how the costs of large loads are allocated and how decarbonisation objectives are achieved, but also whether projects can be physically connected and operated within the electricity system in a timely and efficient manner.

Why this matters for deal structuring

Each of these reform threads point to the same underlying lesson: regulatory risk, contract structuring and technical planning cannot be sequenced or priced in isolation:

• A change-in-law clause only protects a developer if the underlying technical basis for compliance has been tested first; otherwise the clause is allocating a risk nobody has yet quantified.
• Allocating responsibility for the REGO matching obligation between a data centre developer and its co-located energy counterpart is not purely a drafting question – it turns on dispatch control and the terms of the connection agreement.

Developers who align their technical diligence, connection strategy and contract allocation early will be better placed to finance and price these projects than those who treat regulatory compliance as a separate workstream to be reconciled later.

A silver lining for whom?

Demand for power, storage and compute is surging. These anticipated reforms effectively make data centre developers the funders of Australia's next wave of renewable and battery projects – the price of a social licence to build.

Whatever shape the final rules take, one thing is clear: certainty matters more than the outcome itself. Developers who plan now for owner-liability, jurisdictional matching and firm contracting will be best placed when the rules land.