8 July 2026 · 8 min read · NZ Flow Group
When New Zealand councils talk about rolling out household water meters now, they're increasingly not talking about the mechanical dial type that's sat on properties for decades. They're talking about smart metering, devices and networks that report water use automatically, over technologies like NBIoT and LoRaWAN, without anyone needing to visit the property. Here's what's actually driving that shift, how the technology works, and a gap in the rollout we think ratepayers deserve to know about.
A traditional domestic mechanical meter has a spinning mechanism that counts water passing through it, certified to the international OIML R49 accuracy standard, and it has to be physically read by a person, usually every six months. Smart doesn't necessarily mean a different meter, it means the meter can report its own readings without someone visiting the property, and that shows up in New Zealand in two different ways.
The first, and the one a lot of current rollouts are actually using, is AMR, automatic meter reading. A small comms module is fitted to the meter and counts the pulses it generates, transmitting readings back to the council's system daily instead of waiting for a manual read. Whether that module goes onto the existing meter or a new one depends on what's already on the property. Many older mechanical meters have no pulse output at all, so there's nothing for a comms module to read, and in a lot of these cases the old meter is swapped out for a new mechanical meter that does have a pulse output, with the comms module bolted straight on to count it. Where the existing meter is newer and already has a pulse output, the module can simply be added without replacing anything. IAWAI, the joint water company formed by Waikato District and Hamilton City councils in partnership with Waikato-Tainui, is trialling the retrofit version of this in Te Aakau, upgrading Waikato District Council meters that were originally installed back in 2017, and the same kind of retrofit is common on apartment sub-meters around the country. The second broad approach is full AMI, advanced metering infrastructure, where the meter itself is a digital device from the start, with no mechanical dial at all, communicating continuously rather than being paired with a bolted-on module. All three, retrofit AMR, new-meter-plus-module AMR, and full AMI, count as smart metering in the sense that matters to a council or ratepayer, daily data instead of a twice-yearly read, but they're different combinations of hardware getting there.
Getting a reading off a meter and back to head office is its own engineering problem, and New Zealand rollouts are currently split between two low-power wireless technologies to solve it.
NBIoT runs over the same licensed cellular spectrum as a mobile phone network, through providers like Spark. Watercare has been using it for smart loggers on commercial meters in Auckland, taking readings roughly every 30 minutes. Because it rides on infrastructure the telcos already operate and maintain, a council doesn't need to build its own network, coverage is generally strong anywhere there's existing cell service, and licensed spectrum means less risk of interference from other devices. The trade-off is an ongoing per-meter connectivity fee paid to the network operator, and coverage is only as good as the nearest cell tower.
LoRaWAN is the other technology gaining ground here, and some regions are backing it specifically because it works differently: it runs on unlicensed spectrum through a council or contractor's own network of gateways rather than a telco's towers. Ventia's Vianet business has already rolled out more than 80,000 LoRaWAN-connected smart water meters across Australia and New Zealand, and Queenstown Lakes District Council's smart meter trial across Glenorchy, Luggate, and Hāwea runs on a low-power wide-area network of this kind. LoRaWAN's strengths are long range, up to tens of kilometres in open terrain, very low power draw that can stretch battery life out to a decade, and no ongoing per-device carrier fee once the gateway network is built. Its weaknesses are the flip side of the same coin: someone has to build and maintain that gateway network in the first place, it carries much less data per transmission than a cellular connection, and because it runs on unlicensed spectrum it's more exposed to interference from other LoRaWAN devices nearby, which matters more as a network scales up.
Neither is simply better. NBIoT tends to suit dense urban rollouts where cellular coverage is already excellent and a council would rather pay a per-device fee than build infrastructure. LoRaWAN tends to suit councils covering large, lower-density geographic areas, where relying on cell towers or running cables doesn't stack up financially, and where owning the network outright appeals more than an ongoing telco contract. Expect both to keep showing up in different regions depending on geography and who's doing the procuring.
Close to a quarter of all treated water in New Zealand is lost to network leaks before it reaches a tap, and in some networks that figure approaches 50%. Smart meters directly attack that problem. Because they report frequently rather than twice a year, they can flag a leak or an abnormal usage pattern within days instead of months, and some devices go further and measure pressure as well as flow, which can catch pipe problems before they turn into failures.
That capability lines up with a broader legislative push. Under Local Water Done Well, council water assets are shifting into ring-fenced, financially independent entities regulated by the Commerce Commission, with formal disclosure obligations covering investment, costs, and service quality. An entity that can't account in near real time for what's flowing through its network is exposed financially as well as operationally, and that's exactly the gap smart metering is built to close.
New Plymouth District Council is the clearest example in progress: a rollout of 26,000 smart meters that has already paid for itself in leak detection alone, saving the equivalent of 68 Olympic pools of treated water a year while only halfway through installation, and allowing the indefinite deferral of an anticipated $4 million pump station and pipeline upgrade.
Wellington's new water entity, Tiaki Wai, has said there appears to be a clear benefit to smart meters over mechanical ones as it plans a household rollout, after an initial cost estimate of $500 to 590 million forced a public rethink of both cost and timing. In Hamilton and Waikato District, the joint entity IAWAI began a residential pilot in February 2026 specifically to test smart meter installation, performance, and customer data access before committing to a wider rollout. Older fully mechanical rollouts, like Auckland's and Kāpiti's from decades ago, are the baseline these newer smart deployments are being benchmarked against, and are themselves increasingly being upgraded to smart, either by retrofitting a comms module onto the existing meter, or by swapping in a new pulse-enabled mechanical meter with a module attached where the original meter can't support one.
Because smart meters report daily rather than twice a year, a leak that would once have gone unnoticed for months can now show up within days.
Smart or mechanical, metering itself makes a measurable difference. New Zealanders use an average of 211 litres of water per person per day, well above the European average of 128, according to Te Waihanga's Valuing Water research. Within New Zealand, the split by charging model is stark: Auckland (146 litres per person per day), Tauranga (169), Nelson (177), and Kāpiti Coast (193), all metered and volumetrically charged, sit well below the national average. Wellington (220), Palmerston North (224), and Hamilton (232), none of which currently meter households, sit above it. Tauranga uses 27% less water per person than Hamilton despite similar density and climate, one of the cleanest comparisons available. Metering has deferred real capital spending too: Kāpiti's rollout pushed back a proposed $30 million dam by an estimated 40 years, and Tauranga's deferred the Waiāri Water Supply Scheme by more than 10 years, worth an estimated $53.3 million in 2009 terms.
Are smart meters more accurate than mechanical ones? With no moving parts to wear down, smart meters generally hold their accuracy better over time and need less maintenance. But accuracy at installation is not accuracy forever, on either meter type, and this is where we think the current rollout has a gap nobody is talking about. Irrigation and industrial water takes are legally required to be verified within the installation year and every five years after, to within 5% accuracy, by a qualified provider. Domestic water metering, smart or mechanical, currently has no equivalent requirement at all. Our position is straightforward: it should. We'd recommend every household smart meter be verified against a certified test meter annually, not because anyone is currently requiring it, but because nobody should have to simply trust a bill generated by a device no one has checked in years.
Who owns the usage data a smart meter generates? The water supplier owns and operates both the meter and the data platform behind it. Ratepayers are usually given access to their own usage data through an app or online portal, but the underlying infrastructure and data ownership sit with the council or water entity, not the household.
Does a ratepayer get a say in what meter is installed from a certified list of options? No. Meter technology, brand, and communication network are all specified by the council or water entity through its own procurement process, the same as any other network asset. There's no published certified list a ratepayer can choose from, and no mechanism to request an alternative.
The case for smart metering specifically, rather than metering in general, comes down to speed. A mechanical meter read twice a year can tell a council it lost water sometime in the last six months. A smart meter reporting daily can tell it within days, and in New Plymouth's case that speed has already turned into millions of litres saved and capital spending deferred before the rollout is even finished. That's a genuine operational upgrade, not just a billing mechanism. But speed of data is not the same thing as accuracy of data, and right now nobody outside the irrigation, industrial, and municipal water-take sectors is required to prove either. Councils are installing tens of thousands of devices that will generate real bills for decades, on brand new technology (to the sector). Is there a mechanism in place for the ratepayer to challenge the accuracy of the data? If not, should there be?
We don't think verification needs to be mandatory from the council's side to fix this. What we think should be non-negotiable is the ratepayer's side of the equation: the right to request an independent flow test on your own meter, and to formally challenge the council if that test shows a discrepancy. It's worth noting that water meters are specifically exempted from New Zealand's Weights and Measures Act, the law that gives consumers exactly that right over scales, pumps, and other trade measuring instruments, and placed instead under whichever authority happens to administer the utility. For petrol and packaged goods, that consumer right exists because getting short-changed on a fuel fill or a bag of coal was considered serious enough to legislate for. A water bill built on an unverified smart meter is no different "in principle" though, and as this technology moves from pilot to national rollout, we think ratepayers deserve the same right to check.
NZ Flow Group has spent over a decade supplying, installing, and verifying flow meters for irrigation, municipal, and industrial customers across Canterbury. If your organisation is weighing up smart meter technology, installation standards, or a verification programme, we're happy to talk through what good practice actually looks like.
We'll talk you through meter technology, installation standards, and what a proper verification programme should look like.