Fast Approvals, Full Cost Recovery: What New South Wales’ Data-Centre Rules Could Teach Kenya

Last Updated 1 hour ago by Kenya Engineer

New South Wales has responded to the data-centre boom with a policy bargain that is unusually clear. Developers that meet defined environmental, infrastructure and community expectations can receive a faster and more predictable assessment. In return, they must show how their projects will use resources efficiently and avoid shifting new costs onto households and ordinary businesses.

The framework announced on 17 August arrives as the Australian state reports 19 data-centre projects worth A$50.3 billion in its State Significant Development pipeline, in addition to more than 60 facilities operating or under construction. The state says investment in the sector grew by an average of about 75 percent a year over the three years to December 2025.

The scale explains both the welcome and the caution. Data centres create construction activity, digital capacity and demand for renewable energy. They also require large, highly reliable electricity connections, cooling, fibre, land, backup power and sometimes substantial water infrastructure.

The bargain: a clock for government and tests for developers

Projects that address the NSW Data Centre Guidelines can receive a commitment that their development application will spend no more than 75 days in state-government assessment. The framework retains environmental review and community consultation; it is a service standard, not an exemption from planning law.

The guidelines organise project expectations around six principles: world-class environmental and efficiency performance; no net cost to consumers and communities; additional water and energy supply; improved local infrastructure and amenity; investment across the future-industry supply chain; and training and skills.

This is important because planning delay and weak regulation are not opposites. A government can define its evidence requirements early, coordinate agencies and decide efficiently, while still imposing demanding conditions. The faster pathway gives investors a reason to prepare better applications rather than rewarding those that provide the least information.

Specific performance measures make the principles credible

The NSW guidelines move beyond general sustainability language. They set design ranges that combine Power Usage Effectiveness and Water Usage Effectiveness, recognising that a cooling design may save electricity by using more water or save water by consuming more electricity.

One pathway calls for a design PUE no higher than 1.25 with specified water-use limits; another allows a PUE of up to 1.3 with a tighter WUE threshold. The metrics are to be based on mature operation at full IT load and measured using recognised standards. Water-intensive cooling is expected to use recycled water for all cooling operations or have an agreed transition to rainfall-independent supply, while facilities using potable water temporarily should be able to reduce consumption during drought restrictions.

The framework also requires attention to noise and vibration, fire and explosion risks associated with diesel and lithium-ion systems, and air pollution from backup generators. These are not peripheral matters. A cluster of standby generators can have the installed capacity of a power plant even if it operates only during tests and outages.

On the electricity side, qualifying projects should demonstrate the ability to reduce grid-supplied demand by 25 percent of forecast average load for up to two hours. The preferred order is to shift flexible computing and ancillary loads first, then use on-site or nearby renewable generation and storage. Routine peak support using diesel is not encouraged.

Cost recovery is the centre of the policy

The most transferable feature of the NSW model is not the 75-day number. It is the principle that an exceptional new load should finance the additional infrastructure and service level it requires.

A large data centre may trigger new transmission or distribution works, substations, water mains, recycled-water treatment, roads and fibre ducts. If connection charges recover only part of these costs, the balance can appear in future utility tariffs or public budgets. The framework therefore envisages capital contributions, prepayments, guarantees and utility agreements that protect the wider customer base.

Some of this architecture is still being built. NSW has introduced enabling legislation and opened consultation on electricity-network connection and cost recovery until 14 September. Its pricing regulator, IPART, will separately review whether water pricing can recover the full cost of servicing data centres and manage scarcity risks. The announcement should therefore be described as a policy framework and reform programme, not as proof that every cost-allocation question has already been settled.

The framework also acknowledges speculative demand. Its guidelines cite estimates that only a minority of connection applications may proceed. Requiring credible schedules and financial commitments can prevent phantom projects from occupying network-planning capacity that other users need.

Why Kenya should pay attention now

Kenya already has the ingredients of a regional data-centre hub: international fibre connectivity, a renewable-heavy generation mix, a strong digital-services market and growing demand for local cloud and AI capacity. Konza is operationalising a Tier III-certified national data centre, and the planned Nxtra facility at Tatu City is expected to reach 44 megawatts at full scale.

Those advantages do not remove physical constraints. Large loads can require dedicated substations, transmission reinforcement and long-lead transformers. Cooling designs must respond to local temperature and water availability. Backup generation can create fuel, noise and emissions risks. Skilled engineers and technicians are required not only for construction but for continuous operations and maintenance.

Kenya’s draft AI and Emerging Technologies Policy recognises data-centre energy and water use, but the final framework will need an implementation mechanism. Without defined metrics and agency responsibilities, sustainability can become a paragraph in a policy while connection and planning decisions are made project by project.

Kenya should not copy the Australian thresholds blindly

A useful lesson is not the same as a template. NSW operates within a mature Australian electricity market, regulated utility-pricing system and different climate. Kenya must calibrate efficiency thresholds to local temperature, humidity, water stress, grid conditions and commercially available cooling technologies.

PUE alone would be particularly dangerous. A designer can sometimes improve energy efficiency by increasing evaporative cooling and water use. WUE, source-water quality, drought operating mode and the energy embodied in water treatment must be assessed together.

Kenya also needs to distinguish facilities by size and impact. An enterprise server room should not face the same process as a hyperscale campus, while a very large project should not be approved through ordinary building-control procedures without coordinated grid, water, environmental and emergency review. The threshold should be established transparently by the responsible agencies after technical consultation.

A possible Kenyan compact

Kenya could build its own bargain around seven elements. First, create a pre-application review bringing together the ICT ministry, county planning authority, NEMA, EPRA, the relevant electricity network operators and the local water service provider. The purpose would be to identify fatal site constraints before a developer completes an expensive design.

Second, require a capacity ledger that separates ultimate campus size, requested grid capacity, approved connection, construction phases, energised critical IT load and actual demand. Third, use connection deposits, guarantees or milestone rules to prevent speculative projects from blocking network capacity.

Fourth, require a combined energy and water plan: PUE and WUE targets, cooling technology, drought mode, source of additional water, demand response, storage, renewable procurement and backup-generation testing. Fifth, make the developer fund directly attributable network and utility upgrades under transparent agreements.

Sixth, publish operational performance in a consistent form, including energy and water intensity, outages, generator hours and material environmental incidents, while protecting legitimate security information. Seventh, give compliant projects a published decision timetable and a named team responsible for coordinating agencies.

Local benefits should be engineered into the programme through apprenticeship targets, technician training, Kenyan professional participation, local maintenance capability and investment in fibre, power or water assets that continue serving the surrounding area where technically and financially appropriate.

A larger role for engineers in digital policy

The data-centre debate is often led by investment, technology and privacy specialists. The NSW framework demonstrates why electrical, mechanical, civil, water, environmental, fire, acoustic and communications engineers must be involved from site selection onward.

For Kenya, the prize is not simply more buildings filled with servers. It is a reliable digital-infrastructure sector that strengthens rather than strains the power system, does not compete irresponsibly for drinking water, develops local technical capability and pays for the exceptional services it consumes.

Faster approvals can be part of that strategy. But speed becomes credible only when government knows exactly what evidence it requires and developers know exactly what performance they must deliver. That is the most valuable lesson in the New South Wales model.