What Is PUE and Why Does It Matter for a Data Centre?

What is PUE and why does it matter for a data centre? Power Usage Effectiveness, or PUE, is the ratio of the total power a facility draws against the power its servers and IT equipment actually use, and it is the standard industry measure of how much electricity goes to cooling, lighting and power conversion rather than computing. A lower PUE means less overhead and a more efficient site, and it directly affects how big your electrical supply, cables and switchboard need to be.

⚡ Quick Answers: Power Usage Effectiveness (PUE)

• What does PUE actually measure? It divides the total power a facility uses by the power its IT equipment uses, showing how much extra is spent on cooling and overhead.
• What counts as a good PUE? Closer to 1.0 is better. Older air cooled halls are often quoted around 1.5 or higher, while modern liquid cooled designs are often quoted between about 1.1 and 1.2.
• Who created the PUE metric? It was developed by members of The Green Grid, an industry group focused on data centre energy efficiency.
• Has PUE improved much in recent years? Uptime Institute reports that the global average PUE has stayed relatively stable since 2018, though its 2026 survey found minor improvements, held back by older legacy facilities.
• Does PUE change the size of my electrical supply? Yes. A higher PUE means the mains, switchboard and network application all need to cover more than just the IT load.
• Can an electrician measure my facility's PUE? We can log real demand at the IT load and at the main switchboard, which gives you the inputs needed to calculate it accurately.

1. The PUE formula in plain English

PUE is calculated as total facility power divided by IT equipment energy. Total facility power is everything the site draws, measured at the utility meter: servers, storage, networking gear, chillers, pumps, fans and lighting. IT equipment energy is measured separately, usually at the power distribution units feeding the racks, and covers only the computing, storage and networking load itself.

A PUE of 1.0 would mean every watt entering the building goes straight to computing, with nothing spent on cooling or overhead. In practice that never happens, because heat removal and power conversion always use some electricity. The further the number sits above 1.0, the more of your bill is going to overhead rather than the workload you actually need to run. The inverse of PUE, called data centre infrastructure efficiency or DCiE, expresses the same relationship as a percentage, where a PUE of 2.0 equals a DCiE of 50 per cent.

2. Where the overhead actually goes

Almost every watt a server draws becomes heat, and that heat has to be removed or the equipment throttles itself, fails early or shuts down. Chillers, computer room air handlers, pumps and cooling towers run continuously to move that heat out of the building, and cooling is usually the largest single contributor to a facility's PUE overhead. Power conversion adds a smaller but constant loss too, since electricity is stepped down and converted several times between the grid and a processor, through transformers, switchgear and uninterruptible power supplies, and each step loses a little as heat.

Lighting, building services and anything else running off the same meter also count toward total facility power, even though they have nothing to do with computing. This is why a facility with efficient cooling but old, inefficient lighting and building management systems can still report a disappointing PUE.

3. What counts as a good PUE in 2026

There is no single number that applies to every site, because climate, cooling design and facility age all change the result. Older air cooled halls are often quoted at around 1.5 or higher. Modern facilities using contained hot and cold aisles, free air cooling where the climate allows it, or liquid cooling for high density racks, are often quoted between about 1.1 and 1.2. The best hyperscale operators have reported figures close to 1.1 for years, achieved through scale, custom cooling design and continuous tuning that most smaller sites cannot justify.

For a Sydney server room or comms room, the realistic comparison is against similar sized facilities, not against a hyperscale campus built for a single purpose. A small room with a handful of racks and a basic split system will usually run a higher PUE than a purpose built data hall, simply because it lacks the scale to spread fixed overheads like lighting and standby equipment across a larger IT load.

4. Why PUE has barely moved industry wide since 2018

Uptime Institute's own research states that the global average PUE has remained relatively stable since 2018, even though individual operators keep investing in efficiency. Its 2026 Global Data Center Survey, the sixteenth annual edition, reported only minor improvements in average PUE levels this year, and it attributes the slow industry wide progress to legacy infrastructure, meaning that older buildings with outdated cooling and power systems are dragging the average down even as new builds get more efficient.

That matters for anyone operating an existing server room or comms room in Sydney. If your facility was fitted out more than five or ten years ago, its PUE is very likely worse than a new build, and the gap tends to widen as equipment ages and cooling systems lose efficiency. Measuring your own number is the only way to know where you actually sit.

5. Why bigger sites usually report a better PUE

Scale helps efficiency in a few concrete ways. A large data hall can justify investment in high efficiency chillers, variable speed drives and sophisticated controls that would never pay back in a small server room. Fixed overheads such as standby lighting, security systems and building management also get spread across a much larger IT load, which pulls the ratio closer to 1.0. Hyperscale and large colocation operators also tend to run newer facilities, built specifically around current best practice in cooling and power distribution, rather than a room retrofitted inside an older commercial building.

None of this means a smaller site cannot improve. Contained aisles, better air sealing, more efficient cooling units and simply removing old, underused equipment can all bring a modest server room's PUE down meaningfully, even without a full rebuild.

6. What PUE means for your cables, mains and switchboard

PUE is not just a sustainability number. It directly sets how much total power your site needs to draw for a given amount of computing, which flows straight into how the electrical supply is sized. A facility with a PUE of 1.8 needs a meaningfully bigger supply, bigger consumer mains and a bigger switchboard than one with the same IT load running at a PUE of 1.2, because the network application has to cover the full facility draw, not just the servers.

This is where a load assessment earns its keep. Rather than guessing from nameplate ratings, our power logging service measures real demand at both the IT load and the total facility meter over days or weeks, which gives you an accurate current PUE and the real numbers needed for a network application. From there, sizing consumer mains, switchgear and a switchboard upgrade follows AS/NZS 3008 for cable capacity and the connection rules set by the Service and Installation Rules of NSW. Our guide to how much power a data centre uses in Australia walks through the load side of that calculation in detail.

7. Measuring PUE properly before you upgrade

A rough estimate of PUE is worse than no number at all if it leads to an undersized or oversized supply application. Proper measurement means metering the total facility power at the main switchboard and the IT load separately at the rack or PDU level, over a period long enough to capture normal operating variation, not just a single quiet afternoon. Cooling load in particular swings with outside temperature, so a PUE measured only in winter will understate what the site draws in a Sydney summer.

Once you have reliable numbers, they feed straight into planning. A rising PUE over time can flag ageing cooling plant before it becomes a reliability problem, while a falling PUE after an upgrade proves the investment worked. Either way, the same measured data supports your network application to Ausgrid, Endeavour Energy or Essential Energy when it is time to increase supply capacity, and you can check your network area in seconds with our postcode checker before you start. Our guide to data centre power supply upgrades in Sydney covers that process end to end, from load assessment through to a staged cutover.

Why it's Level 2 ASP work

PUE tells you how much total power your site draws, and once that number points to a bigger supply, the work shifts to the connection between the network and your installation. That is Level 2 ASP territory: preparing the network application, installing larger consumer mains and coordinating approval with the distributor. Our Level 2 electrician page explains what that accreditation covers, and our switchboard upgrades service covers the distribution board work that usually comes with it.

For the very largest campuses, where the load cannot be carried on a standard low voltage connection, the distributor's own high voltage design and large customer process takes over. Our Level 3 design and Level 1 network installation pages explain how that side of the network works, purely as background, since that scale of connection sits with the distributor's own engineers rather than with us.

How High Demand Electrical helps

We are a licensed local team, not a call centre, and for data centres, server rooms and comms rooms we focus squarely on the supply side of the PUE equation: measuring real demand, sizing the cables and mains that feed the site, preparing the network application and upgrading the switchboard behind it. We work across Sydney's West, South West and Sutherland Shire, and our data centre power upgrades page sets out the full scope of that work. For the thinking behind why facilities draw so much power in the first place, see our guide to why data centres use so much power. We will always say plainly if your facility's load points toward a high voltage connection, since that sits outside what a low voltage Level 2 ASP can deliver.

PUE and Data Centre Power FAQs (Sydney)

Is a lower PUE always better?

Generally yes, since it means less of your power goes to overhead. The realistic comparison is against similar sized facilities, because a small server room will rarely match a purpose built data hall.

What is a typical PUE for a Sydney server room?

Older, air cooled rooms are often quoted around 1.5 or higher, while newer, well designed spaces can sit closer to 1.2. The only way to know your own number is to measure it. See our power logging page.

Does PUE affect how big my power supply application needs to be?

Yes. The distributor assesses your total facility draw, not just the IT load, so a higher PUE means a larger supply, larger consumer mains and a larger switchboard. See our guide to data centre power supply upgrades.

Why hasn't industry average PUE improved much?

Uptime Institute attributes the slow industry wide progress to legacy infrastructure, meaning older facilities with outdated cooling and power systems hold the average back even as new builds get more efficient.

Who can measure PUE and plan a supply upgrade for my site?

Call High Demand Electrical on 0402 559 777. We log real demand at the IT load and the main switchboard, then size the mains and network application from measured numbers. See power supply upgrades.

Get a Power Supply Upgrade Quote: Call High Demand Electrical on 0402 559 777 to measure your facility's PUE and plan the right supply for your data centre, server room or comms room. Every job needs a site assessment before scope or pricing can be confirmed.