The world's largest operators have committed to returning more water to catchments than their facilities consume by 2030, and stormwater capture, retention and recharge are among the eligible ways to produce that replenishment. A second group of sites has a different problem. Air-cooled campuses that draw no mains water and no bore water are carrying their operational demand on harvested rainfall, which makes storage sizing and water quality a supply-security question rather than an offset one. The TNS™ is a source-control system that retains water where it falls, instruments what it retains, and can carry the site's standing fire-water reserve in the same storage. This page sets out where it fits, and just as plainly where it does not.
Google has committed to replenishing 120% of the freshwater it consumes across its offices and data centres by 2030, and reports 165 water-stewardship projects across 97 watersheds toward that target (Brandt, 2021; Fisher, 2026). Amazon Web Services, Microsoft and Meta have each made equivalent water-positive-by-2030 commitments (DatacenterDynamics, 2022). Between them, that is the industry standard rather than an outlier position.
Efficiency alone cannot make a water-consuming facility positive, so operators close the gap by funding catchment-scale projects that return volume to the source catchment (CEO Water Mandate, 2021). Nature-based and green-infrastructure works, rainwater harvesting, stormwater capture and groundwater recharge are explicitly eligible, and the volumes are quantified using Volumetric Water Benefit Accounting – the World Resources Institute-led method operators use to claim them (Reig et al., 2019).
That is the whole opening for source control. The obligation is durable, board-level and non-discretionary, and the thing being bought is measurable returned volume.
Stated plainly: the commitments above are those operators' own published commitments, cited to their own statements. Nothing on this page implies a relationship with, endorsement by, or engagement with Google, AWS, Microsoft or Meta.
Take the cooling water out and the water question does not go away. It moves.
Closed-loop and air-cooled designs are becoming the Australian default, and the largest campus yet proposed in this country is one of them. Removing evaporative cooling removes the make-up draw. It does not remove roof runoff, hardstand runoff, the fire-water reserve, the landscape demand or the discharge the site has to account for to its council.
On a site designed with no town water connection and no bore, harvested rainfall stops being a sustainability line item. It becomes the supply. A harvesting system carrying most of a facility's operational demand has to hold through a dry run, and the variables that decide whether it does are storage volume, catchment area and the quality of what arrives in the tank. Those are design decisions made once, at feasibility.
The TNS™ incorporating EnviroKerb® addresses all three. Storage is modular and can be added to as the campus stages. The kerb inlet traps the first flush before it reaches the tank, so the stored water stays fit for the non-potable uses it is being relied on for. The same lined storage carries the fire-water reserve required under AS 2419.1:2021.
The Water Services Association of Australia – the peak body for the urban water industry – set out five priorities for the sector in December 2025: early engagement with utilities at feasibility stage, public reporting of water and power use efficiency, minimum efficiency standards including extending the NABERS Data Centre tools to water, prioritising recycled water and circular-economy initiatives, and fair infrastructure cost recovery from operators (WSAA, 2025).
The direction of travel is unambiguous. On-site water performance, stormwater management and demonstrable circularity are moving from voluntary reporting toward conditions of approval and social licence. Because the Australian rules are still being written, a specification-ready position taken now sits inside the standard as it forms rather than chasing it afterwards.
The inland picture is different and now larger. Queensland's Western Downs is carrying the biggest campus proposed in Australia, on roughly 725 hectares near Dalby in the Condamine headwaters. There is no metropolitan supply to draw on out there and no utility relationship to negotiate, which puts the entire water conversation into the development application and the conditions attached to it.
Which makes it, hydrologically, a very large catchment with a very short response time – often in or near a stressed catchment, and generating a measurable heat footprint. The TNS™ is a source-control system: a permeable kerb-line inlet, modular storage and structural soil cells beneath the surface, across four variants, with sensor-based condition monitoring. Several site obligations land on the same asset.
| Site driver | How the TNS™ responds | Basis |
|---|---|---|
| Stormwater and flood resilience on a critical, highly impervious site | Captures, slows, retains and infiltrates runoff at source rather than conveying it off site; the storage can be lined for volume retention and sized to the design event. | Sized per the applicable local drainage manual and Australian Rainfall and Runoff; storage volume is a design output, not a product rating. |
| Water-positive replenishment targets | Returns volume to the local catchment through retention, infiltration and recharge, and displaces mains demand for landscape irrigation. | Replenishment means returning volume to the source catchment (CEO Water Mandate, 2021); VWBA is the recognised quantification method (Reig et al., 2019). |
| Mains water for landscape and amenity | Passively irrigates trees, verges and raingardens from retained stormwater in the root zone instead of potable supply. | The passive-irrigation mechanism is set out on How it works. |
| Site heat and amenity | Canopy and evapotranspiration cool the local microclimate; keeping trees watered is what makes canopy viable on a hardstand site. | Vegetated cover has been associated with land-surface temperature reductions of about 0.6–3.4°C in a systematic review (Yin et al., 2024). |
| Standing fire-water reserve | Lined retention storage can be configured to hold or top up the mandated reserve, so one tank farm does more than one job. | AS 2419.1:2021 sets the assured-supply duration; a harvested supply is a performance-solution pathway (Standards Australia, 2021). See below. |
| Measurement and disclosure | Real-time flow, level and condition sensing makes retained volume a measured quantity rather than a modelled one – which is the standard an assurance process asks for. | Monitoring by Matter (matter.city) – Matter SensAI. |
BGI makes no stormwater quality claim for the TNS™. Where a facility carries stormwater quality obligations, those are met by separate, appropriately certified measures – the position is set out in full on Compliance.
Conventional pit-and-pipe drainage is designed to carry runoff away from the site. Once that is the design premise, returning water to source looks expensive – so on-site retention is frequently dismissed before it is properly costed. The TNS™ starts from the opposite premise.
The manifold behind the permeable kerb, the modular storage tanks and the structural soil cells can be lined with an impermeable membrane. Lined, the captured volume is held rather than lost to the subgrade. Additional modular tankage can be added elsewhere on the site to increase the stored volume, so the storage can be sized for larger and longer-duration events – up to the 1% AEP storm where the design calls for it.
What that produces is not a drainage outcome, it is a water resource: a volume on site, available for non-potable use, and quantifiable as retained volume for water-balance and replenishment accounting. That is a materially different proposition from detain-and-discharge.
Retained volume, the number of nodes, tank sizing and the event the storage is designed to are all site-specific design outputs – calculated per micro-catchment for the site in question. They are not product ratings and BGI publishes none. Storage sized to the 1% AEP event describes what the system can be configured to do, not a property every installation carries.
Inlet capacity is in the same category. BGI publishes no permeability rate or inlet capacity for the kerb units – the number of units at a node is calculated from the micro-catchment runoff for the site.
On hardstand the runoff coefficient is close to one, and the detention requirement is what constrains the site layout. The detention basin built to hold flows back until they fit downstream comes out on most layouts, and the basin is usually the expensive one, because it takes developable area as well as construction cost.
A node holds the runoff from the micro-catchment it serves, so that volume never arrives at the bottom of the site. Across a precinct that shows up in the drainage design in three places. Pipes are sized on the flows that reach them. The existing downstream system can often take the discharge without an upgrade. And the basin comes out.
It isn't a replacement for conventional drainage. It's a smaller set of tools doing the same job in more places, which is why the network that remains is smaller.
Every number is site specific and comes out of the ARR model for the catchment being designed. Send a layout and a catchment plan and we'll run it.
Fire-hydrant supplies in Australia are governed by AS 2419.1:2021, under which the supply must sustain the required flow and pressure for at least four hours (Standards Australia, 2021). To illustrate the order of magnitude: two hydrants running concurrently at 10 L/s for four hours is about 288 kL, before hose reels and any sprinkler or deluge demand is added. A hyperscale campus can therefore need several hundred kilolitres to more than a megalitre of assured, standing, non-potable water – held in tanks that, by design, sit idle.
The reserve cannot be drawn down for other purposes, which is precisely why it is treated as dead capital. A layered tank design resolves that: the fire-reserve volume sits permanently at the base of the tank farm on a protected low-level offtake, while harvested stormwater fills the working volume above it for landscape irrigation, non-potable make-up and other site demand. Mains or bore top-up and continuous level monitoring keep the reserve assured at all times. One tank farm, three duties – fire reserve, stormwater retention, reuse supply.
On approvals: on-site and alternative supplies are already permissible, and authorities accept tank and bore arrangements. A harvested supply would be taken forward as a performance solution – fire authority and water utility engaged at feasibility stage, as the WSAA urges, with assured reliability demonstrated through top-up, dual booster pumps and continuous level and flow monitoring (Queensland Fire Department, 2025; WSAA, 2025). That is a design-and-approval workstream on each project. BGI does not represent that any such arrangement is pre-approved.
Fire water is the most reliable of the reuse duties, because it carries no cooling-grade quality constraint. It simply has to be there.
A hyperscale campus arrives at a council as the largest impervious-area application that council will ever see, and usually with no local precedent to assess it against. The design decisions that determine retention, discharge and on-site reuse are fixed early and become expensive to revisit once conventional drainage is locked in. We are available to councils for a technical discussion on what a source-control condition can reasonably require, at no cost and with no engagement implied.
The TNS™ specified into the development itself – car parks, access roads, perimeter and landscape buffers, detention and raingarden areas – to meet the stormwater, discharge, heat and amenity conditions of approval the facility carries anyway. The retained volume then serves landscape, non-potable and fire-reserve duty on the same site.
The TNS™ deployed across surrounding streets and the wider catchment as a fundable replenishment programme, designed and instrumented to be quantified under VWBA. This is the same category of work operators already fund internationally – urban rainwater harvesting, constructed wetlands, catchment restoration – and it is the pathway most directly aligned with the 2030 commitments.
The on-site system proves the technology in the operator's own asset, under its own monitoring. The off-site programme is what scales the returned volume to something material against a replenishment target. A facility that does both has the evidence and the volume in the same catchment.
Volumetric Water Benefit Accounting quantifies the volume of water an activity produces relative to what would have happened without it. BGI's role is to design and instrument an asset that can be quantified that way, and to hand over the evidence.
Identify the catchment the site sits in and the shared water challenges the work is intended to address – the framework's starting point, not an afterthought.
Define the TNS™ works and, where the programme is co-funded, agree in advance how benefit is allocated between funders.
For a data centre site the counterfactual is conventional pit-and-pipe: runoff conveyed away, no recharge, mains water for landscape and fire top-up.
Sensor-verified flow and level data replaces modelled estimates. Eligible categories are summed without double-counting – recharge against avoided abstraction, on-site against off-site in the same catchment.
BGI designs to the framework, instruments the asset so volumes are measured rather than modelled, and provides the monitoring evidence. BGI does not certify, audit or issue replenishment volumes. VWBA is a World Resources Institute and CEO Water Mandate framework; the claim is the operator's, made against its own baseline and subject to its own assurance process. Claims must also satisfy the framework's additionality principle – benefit beyond what regulatory compliance already required – which is a question about the specific site, not about the technology.
Method: Reig et al. (2019) and the CEO Water Mandate implementation guide (2021). Monitoring: Matter SensAI. Both are named on this page because the claim depends on them.
The evaporative losses of a hyperscale facility are far larger than the volume a stormwater system captures. With enough added storage the retained volume can offset a share of non-potable make-up, and it can more reliably carry the fire reserve. It is not a cooling-water efficiency measure and is not offered as one. On an air-cooled site there is no evaporative make-up demand to substitute for, and this limitation falls away.
Retained volume, node count and the design event are calculated per site. Nothing on this page is a per-unit yield figure, and BGI publishes none.
BGI makes no stormwater quality claim for the TNS™. Quality obligations are met by separate, appropriately certified measures.
It is a performance-solution pathway requiring the fire authority and the water utility, project by project. It is not pre-approved and is not represented as such.
This page describes BGI's own system against the conventional conveyance baseline. It makes no performance comparison with any named product, because BGI has run no side-by-side testing.
The capacity, demand and investment figures above are third-party projections, cited as published. They frame scale; they are not a bottom-up market model and not a forecast of BGI revenue.
The honest version of this proposition is large enough. It does not need the version we could not defend.
No. Evaporative cooling losses at a hyperscale facility are far larger than the volume a stormwater system captures, and BGI does not present the TNS™ as a cooling-water substitute. With sufficient added storage the retained volume can offset a share of non-potable make-up demand, and it can more reliably carry the site's standing fire-water reserve. It is a stormwater, retention, heat and replenishment asset.
It is a design-and-approval pathway, not a standard product. AS 2419.1:2021 requires an assured supply sustaining the required flow and pressure for at least four hours, and authorities already accept tank and bore supplies. A harvested supply would be taken forward as a performance solution with the fire authority and water utility engaged at feasibility stage. BGI does not represent that any such solution is pre-approved.
No. Volumetric Water Benefit Accounting is a World Resources Institute and CEO Water Mandate framework. BGI can design to it, instrument the asset so volumes are measured rather than modelled, and provide the evidence – but the claim is the operator's, quantified against its own catchment baseline and subject to its own assurance process.
No. The commitments described on this page are those operators' own public commitments, cited to their own published statements. Nothing here implies a relationship with, endorsement by, or engagement with any of them.
Everything that is not cooling. The site still sheds runoff off roofs, hardstand and access roads, still has to account for its discharge to the assessing council, still carries a standing fire-water reserve under AS 2419.1:2021, and on a harvested supply it still has to hold operational demand through a dry run. The TNS™ is source control for that runoff: it retains volume where it falls, intercepts the first flush before it reaches the tank, and the same lined storage can carry the fire reserve. On an air-cooled site the water question is supply security rather than offset, and it is the same asset that answers it.
It is calculated per site, and BGI publishes no indicative figure. The inputs are the local rainfall record – including the dry sequences in it, not the annual average – the catchment area contributing to the tanks, and the site's own demand profile across operations, landscape and any make-up duty. Those run through a continuous water balance over the record, and the storage volume is the output that holds the supply through the design dry sequence at the reliability the operator is prepared to accept. A number quoted without the rainfall record and the demand profile behind it is not an answer to this question.
It depends on what the water crosses. Roof runoff arrives comparatively clean; runoff off hardstand, access roads and laydown areas does not. The kerb-line inlet traps gross debris and the filter element beneath it takes the fines, so the first flush is intercepted before it reaches storage and the stored water stays fit for the non-potable duties it is being relied on for. BGI makes no stormwater quality claim and publishes no removal percentages – where the site carries a quality obligation, that is met by separate, appropriately certified measures.
It is a performance-solution pathway. The reserve volume sits permanently at the base of the tank on a protected low-level offtake, harvested water fills the working volume above it, and assured reliability is demonstrated through top-up, dual booster pumps and continuous level and flow monitoring. Approval rests with the fire authority, with the water utility and the assessing council engaged alongside it, and it has to start at feasibility rather than at conditions. BGI does not represent any such arrangement as pre-approved.
Early, or not at all. Both the WSAA priorities and the approval pathway for a dual-use supply point the same way: engagement at feasibility stage, while the stormwater strategy, the tank farm and the landscape scope are still being set. Once the site is designed around conveyance, the retention option has usually been costed out.
Every third-party figure above is traceable to the source below. Everything else is a design position or a site-specific design output, stated as such.
Brandt, K. (2021, September 9). Our commitment to water stewardship. The Keyword, Google. https://blog.google/company-news/outreach-and-initiatives/sustainability/replenishing-water/
CEO Water Mandate. (2021). Volumetric water benefit accounting (VWBA): A practical guide to implementing water replenishment targets. UN Global Compact and Pacific Institute. https://ceowatermandate.org/replenishment-guide/
DatacenterDynamics. (2022). AWS pledges to be water positive by 2030. https://www.datacenterdynamics.com/en/news/aws-pledges-to-be-water-positive-by-2030/
Fisher, K. (2026, June 4). Google commits to replenish more water than it uses at data centers by 2030. ESG Today. https://www.esgtoday.com/google-commits-to-replenish-more-water-than-it-uses-at-data-centers-by-2030/
Queensland Fire Department. (2025). Hydrant and water supply requirements for developments. https://www.fire.qld.gov.au/
Reig, P., Larson, W., Vionnet, S., & Bayart, J.-B. (2019). Volumetric water benefit accounting (VWBA): A method for implementing and valuing water stewardship activities. World Resources Institute. https://www.wri.org/research/volumetric-water-benefit-accounting-vwba-method-implementing-and-valuing-water-stewardship
Sadler, D. (2026, March 31). Australia's data centre boom collides with climate concerns. Information Age, ACS. https://ia.acs.org.au/article/2026/australia-s-data-centre-boom-collides-with-climate-concerns.html
Standards Australia. (2021). Fire hydrant installations: System design, installation and commissioning (AS 2419.1:2021).
Water Services Association of Australia. (2025). Data centres and water in Australia. WSAA. https://wsaa.asn.au/
Western Downs Digital Park. (2026). Project overview. WDDP Pty Ltd. https://www.westerndownsdigitalpark.com.au/
Yin, Y., Li, S., Xing, X., Zhou, X., Kang, Y., Hu, Q., & Li, Y. (2024). Cooling benefits of urban tree canopy: A systematic review. Sustainability, 16(12), 4955. https://doi.org/10.3390/su16124955
If you are shaping a data centre's stormwater strategy, its tank farm or its replenishment programme, the useful conversation happens while those are still open. Technical questions go to the people who did the design work.
