NEXTDC S4 Data Centre Phase 2, Horsley ParkHave your say on the Horsley Park data centre before 10 November 2026

Noise mitigation plan deferred to after construction

The EPA twice advised that the construction noise mitigation strategy is generalised and must be made site-specific before approval is granted. The proponent responded that a detailed strategy will be developed only after a construction program is confirmed, meaning the assessment remains incomplete at the time of determination.

Low-frequency and tonal noise risks unresolved

The EPA noted that data centres are known to produce low-frequency and tonal noise, and required contingency mitigation measures to be proposed before operations begin. The EIS defers this to post-commissioning monitoring, leaving nearby residents without an assured remedy if such noise characteristics emerge.

Cumulative noise from existing and new facilities unassessed

Residents have raised concern about compounding noise from existing industrial warehouses and the already-approved Phase 1 facility. The EIS for Phase 2 must assess the combined operational noise of both phases together with existing background sources, not each in isolation.

Emergency generator noise may exceed night-time limits

The Phase 1 noise assessment found that minor exceedances of Project Noise Trigger Levels are possible during emergency generator operation. Phase 2 adds further generator capacity adjacent to the same residential receivers, and the combined emergency scenario has not been fully assessed.

Air quality assessment not prepared to EPA standards

The EPA found that the Air Quality Impact Assessment for the adjacent Phase 1 was not prepared in accordance with the NSW EPA's Approved Methods for the Modelling and Assessment of Air Pollutants in NSW, and contained inconsistencies in the number of generators modelled and in emission rates for NOx, CO and hydrocarbons.

Cumulative diesel emissions from both phases not quantified

The Phase 1 air quality assessment used a qualitative, probability-based approach for cumulative impacts and did not model a quantitative worst-case scenario with all generators from both phases operating simultaneously. The EPA specifically required quantitative cumulative modelling.

Generator testing hours may trigger scheduled activity threshold

The EPA noted that inconsistencies in the number of generators listed across documents could push annual testing hours above 200, which would trigger the 'Electricity Generation' scheduled activity under the Protection of the Environment Operations Act 1997 and require an Environment Protection Licence. This has not been resolved for Phase 2.

Diesel storage triggers Environment Protection Licence

Phase 1 alone involves storage of 4,472 tonnes of diesel in above-ground tanks, triggering the chemical storage threshold under Schedule 1 of the Protection of the Environment Operations Act 1997 and requiring an Environment Protection Licence. Phase 2 adds further diesel storage, and the combined volume and licensing obligations must be assessed.

Spill containment design not provided

The EPA required the proponent to provide detailed design drawings showing refuelling locations, bunding or spill mitigation, and spill kit locations. As of the EPA's second advice letter, no information had been provided on how the day generator diesel tanks would be refuelled, leaving spill risk unassessed.

Large diesel storage near bushfire-prone land

A portion of the site boundary adjoins Category 1 bushfire-prone vegetation in the C2 Environmental Conservation zone. The proximity of large above-ground diesel storage to this bushfire-prone land poses a fire escalation risk that requires explicit assessment.

Phase 2 water demand places further stress on local network

A GHD water servicing assessment for Phase 1 found that by 2028 the S4 campus will draw an average of 2.65 megalitres per day (30.7 litres per second) from the Cecil Park Remainder Water Supply Zone, served by a single DN 450 main on Burley Street. Phase 2 adds further demand on the same constrained network.

Combined campus water consumption not fully assessed for Phase 2

The water servicing assessment was prepared for the full S4 campus staged to 2028 under Phase 1. The EIS for Phase 2 must demonstrate that the additional demand from two further buildings can be met without degrading supply pressure or reliability for existing residential and rural users in the zone.

Recycled water supply not yet secured for Phase 2

A recycled water connection via the Rosehill scheme was approved separately in September 2026 to serve the S4 campus, but it is not clear whether that approval covers Phase 2 volumes or whether a firm supply agreement is in place. Approval should not proceed without a confirmed, legally secured water supply.

Phase 2 adds further large buildings adjacent to residences

Phase 2 proposes two additional data centre buildings with a gross floor area of approximately 35,992 square metres on a site adjoining the already-approved Phase 1 campus. Residents to the east are rural-residential and have raised repeated concerns about building height and privacy impacts.

Visual assessment must address combined Phase 1 and Phase 2 massing

The Phase 1 EIS addressed visual impacts for that approval in isolation. A cumulative visual impact assessment showing the combined built form of both phases, including rooftop plant and cooling equipment, has not been demonstrated to be included in the Phase 2 EIS.

Light spill onto rural-residential properties not adequately addressed

Residents immediately east of the site raised concerns about 24/7 operational lighting affecting privacy and amenity. The proponent's response to Phase 1 consultation stated it did not expect light spill impacts, but provided no quantitative modelling to support that position.

Burley Road safety risk from construction and shift traffic

Residents have reported that vehicles associated with the existing S4 construction travel through Burley Road, a partially unsealed local road with blind crests and residential driveways, at all hours. The road is not designed for heavy or frequent industrial traffic.

Cumulative traffic from Phase 1 and Phase 2 not assessed together

Phase 2 relies on the vehicle access arrangements established for Phase 1, but the combined construction and operational traffic from both phases operating concurrently has not been demonstrated to be assessed. The EIS must model peak combined traffic generation.

C2 conservation land directly adjoins the Phase 2 site

The site's south-eastern boundary abuts land zoned C2 Environmental Conservation, which is also identified as Category 1 bushfire-prone vegetation. Construction and operational activities for Phase 2 will generate edge effects — noise, light, dust and altered hydrology — on this conservation land.

Threatened species previously recorded near the site

A community submitter reported that a threatened frog species was previously recorded in and around the Horsley Park site before clearing. The Phase 2 EIS must demonstrate that any remaining habitat for threatened species in or adjacent to the Phase 2 footprint has been surveyed and assessed.

Cumulative biodiversity offset obligations not assessed

Phase 1 required a biodiversity statement as part of its EIS. Phase 2 covers a further site area of approximately 8.2 hectares on a site described as entirely disturbed. The cumulative biodiversity offset obligations across both phases, and whether they have been secured, must be assessed before Phase 2 is determined.

EPA greenhouse gas recommendations not enforceable under Phase 1

During Phase 1 assessment, the EPA confirmed that its greenhouse gas recommendations were not currently enforceable and would not translate into development conditions or mandatory reporting requirements. Phase 2 presents a further opportunity to correct this gap.

134.4 MW facility will generate substantial ongoing Scope 2 emissions

Phase 2 is designed for a maximum power consumption of 134.4 megawatts operating continuously. The EIS must quantify the annual Scope 2 greenhouse gas emissions from grid electricity consumption and demonstrate a credible, time-bound pathway to sourcing renewable energy.