NSW’s independent pricing regulator is set to examine how data centres should pay for the water infrastructure needed to service them, including whether charges should reflect peak demand or average consumption.
The NSW government has asked IPART to review the water pricing and regulatory framework for data centres, including how the costs of servicing large, high-intensity users should be recovered. The review will run alongside IPART’s existing review of developer charges for metropolitan water businesses, which covers the costs of connecting, extending or upgrading water, wastewater and recycled-water services.
The reviews are set to shift the debate beyond how much water data centres consume to whether the demand that determines the capacity the network must provide should also determine how much of the resulting infrastructure cost the data centre pays.
Sydney Water says data centres currently account for less than one percent of Greater Sydney’s total water demand, or around 3.5 billion litres a year. However, demand from the sector could increase Greater Sydney’s total water demand by up to 20 to 25 percent by 2035, Sydney Water said, although that demand does not need to be met through drinking-water supplies.
“The new NSW Data Centre Policy Framework sets world-class water efficiency standards,” a Sydney Water spokesperson told W.Media. “It will also include an IPART review which will consider how water pricing can reflect the full costs associated with servicing data centres, ensuring other water users on the network are protected and the impacts of drought and water scarcity are appropriately managed.”
It is understood that additional large-scale recycled-water infrastructure will be required to service future data centre demand, with treatment and network assets needing to be sized to accommodate peak water use.
That creates a potential tension with the position taken by industry body Data Centres Australia in its submission to IPART’s separate review of metropolitan developer charges.
Peak demand versus actual consumption
DCA argues that using peak daily demand as the basis for allocating infrastructure costs can overstate the actual burden imposed by data centres. Its position is that peak demand figures generally represent worst-case conditions, for example, like extreme weather combined with particular cooling-system configurations, rather than typical daily consumption.
The industry body argues that annualising such a peak figure by assuming it occurs every day can significantly overstate the volume drawn from the network. DCA also argues that data centres can manage peak demand through measures including on-site water storage and flow-control systems such as night-time filling.
The result, in DCA’s view, is that average annual consumption provides a more representative measure of the actual impact a facility has on the water network and should have greater weight in calculating developer charges.
However, a treatment plant or network connection that has enough capacity to supply average demand may not have sufficient capacity when a data centre reaches its maximum water requirement. The question for the regulator is therefore whether the peak capacity that has to be provided for a data centre should translate into a corresponding infrastructure charge, even where operational measures can reduce the frequency or duration of that peak.
Who pays for the first connection?
There is another issue brewing behind the debate: the first-mover problem. A large data centre arriving in an area may trigger investment in new treatment, recycled-water or network infrastructure that subsequently has capacity available to other developments.
DCA’s position is that the first project should not be penalised simply because it is the first customer to arrive.
Instead, it argues that if the initial customer funds infrastructure that subsequently benefits other developments, mechanisms should allow some of those costs to be recovered from later users.
That raises a broader question for IPART: whether developer charges should reflect the incremental cost imposed by an individual data centre, or whether a large project can legitimately be required to fund infrastructure that forms part of a wider network expansion.
IPART’s existing developer-charges review is examining how the costs of servicing new development should be recovered and how those costs should be allocated between developers and existing customers.
IPART has both pieces of the puzzle
The NSW Government’s new data-centre-specific review potentially brings these questions together. IPART’s draft terms of reference cover the pricing and regulation of water services to data centres, including infrastructure upgrades, additional supply capacity and the impact of large or high-intensity users. And the regulator has said its existing developer-charges review is relevant to some of the issues in the new terms of reference.
Undertaking the two reviews simultaneously, IPART says, will help ensure that data centre infrastructure costs are considered within the broader developer-charges framework and that relevant findings are aligned.
Public submissions on the new draft terms of reference are open until 2 October, after which IPART will commence the full review and consult further with stakeholders.
The reviews are timely. Sydney Water’s current data centre demand remains relatively small in the context of the overall Greater Sydney water system. But if the sector’s demand grows as projected, the question of how much new infrastructure is required, including how much of that cost should be recovered from data centres, becomes a real bone of contention.
The emerging debate is therefore moving on from whether data centres use too much water. It is now focused on whether the network built to meet their peak, not their average, should also be the basis on which they’re charged. IPART’s two reviews are where that question may finally get answered.