A standard rotational grazing setup on a New Zealand beef or dairy farm involves a lot of walking. Someone has to put up a single-wire electric fence across a paddock, peg it, energise it, move the cattle through, take the wire down twelve hours or twenty-four hours later, and put it up again somewhere else. On a small block it’s a couple of hours of work a day. On a 1,000-cow dairy operation it’s a structural part of a labour roster.
Halter is a New Zealand-built piece of ag-tech that replaces that wire with a GPS collar around each cow’s neck and an app on the farmer’s phone. Instead of a physical fence, the boundary is drawn on a screen. The cow learns it through audio and a mild pulse. The farmer moves the herd by tapping a new shape on a map.
It is genuinely unusual technology, and it has made New Zealand farming a slightly more interesting place to look at from the outside. This post explains what it does, what the trade-offs look like, and where farms like ours sit in relation to it.
What Halter actually is
Halter Limited is a Hamilton-founded company that builds GPS-enabled cow collars and the cloud and mobile-app stack behind them. The product is sold primarily to dairy and beef farmers in New Zealand and, more recently, Australia and the United States. Each animal in a herd wears a solar-charged collar containing a GPS receiver, an accelerometer, a small speaker, and a low-current pulse element. The collars talk to a base station on the farm, which talks to the Halter cloud, which talks to the farmer’s phone.
What this stack lets a farmer do, in plain terms:
- Draw a virtual paddock on a map and have the herd held inside it without a physical fence.
- Move the herd from one paddock to another by changing the drawing — the collars play audio and, if needed, give a brief pulse to guide animals across the gap.
- Get an alert when a cow’s behaviour changes in a way that suggests heat, calving, or illness, because the accelerometer and movement data picks up patterns that look different from baseline.
- See where every animal is in real time, including which cows are lying down, how much they’re walking, and whether any have separated from the mob.
The animal-control mechanism is the part that gets the headlines, but the data side is doing a lot of quiet work in the background.
How the boundary actually works
The training period is the bit most people don’t see. When a herd is first fitted with collars, animals are introduced to the system across roughly a week, with a real fence still in place behind the virtual one. The collar plays a rising audio tone as the cow approaches the boundary. If the cow keeps walking, a brief electrical pulse follows — comparable to the contact from a standard electric fence. The cow learns very quickly that the sound predicts the pulse, and after the training period the audio alone is enough to turn most animals around.
In day-to-day use, the pulse rarely fires. Halter has published figures suggesting a typical mature cow gets a pulse roughly once every several weeks once trained, and that the sound cue does almost all the work. The energy delivered is lower than a mains electric fence pulse and is regulated under New Zealand animal welfare standards.
The system can also “shift” cattle proactively. If a paddock needs to be rotated at 4am, the collars can guide the herd toward the new shape without a person being in the paddock. Cows are creatures of strong habit; over weeks of use they begin to walk to the gate of their own accord when the pre-shift audio plays.
Why dairy farms picked it up first
Dairy in New Zealand is structurally a good fit for the collar.
Cows are already individually identified and handled twice a day. The economic value of each animal is high enough to justify the per-collar cost. Pasture management is the central operational discipline of the entire farm, so any tool that improves pasture allocation — even by a few percent — has a measurable effect on milk solids per hectare. And dairy herds are usually grazed in tighter rotations than beef, meaning the labour saved per week is large.
The first wave of Halter adoption ran heavily through Waikato, Taranaki, Canterbury, and parts of Southland. The pattern was usually a mid-to-large dairy operation looking at a labour problem and concluding that a collar system was cheaper than another full-time staff member, or that it would let a smaller team manage more land without a drop in pasture quality.
Beef adoption has come later and slower. The economics are tighter — beef per-head margins are lower than dairy per-cow returns — and beef herds aren’t handled as frequently, so the calving and heat-detection data is less load-bearing. But for intensively rotationally grazed beef farms, especially on hill country, the labour and time savings still stack up.
Where it fits
The clearest fits, based on what’s public:
- Large rotational dairy. This is the core market. Daily shifts, big herds, pasture-allocation discipline that matters. The system pays for itself fastest here.
- Beef on hill country with rotational grazing. Steep terrain makes electric fencing slow to put up and take down. Virtual boundaries don’t care about contour.
- Farms with a labour shortage. Rural staffing is hard. Anything that lets a smaller team cover more ground without compromising pasture management has structural value.
- Operations with a strong pasture-monitoring discipline already. Halter doesn’t think for the farmer — it executes a plan. Farms that are already careful about residuals, cover, and rotation cycles get the most out of it.
Where it doesn’t fit
The honest picture has limits too.
- Sheep. Halter’s collar is built for cattle. Sheep are smaller, fleeced, more easily handled in mobs by dog rather than individually, and the per-head economics don’t currently justify a collar. There are research projects looking at virtual fencing for sheep elsewhere in the world, but no commercial system at the same scale yet.
- Off-grid or low-coverage country. The collars need to talk to a base station, and the base station needs power and connectivity. A back-paddock with no cellular signal and no line-of-sight to a base unit isn’t a natural fit.
- Small herds. A ten-cow lifestyle block doesn’t benefit from the system the way a 400-cow rotation does. The fixed costs of the base station and subscription don’t amortise across the smaller animal count.
- Farmers who want maximum animal-handler contact. Some regenerative graziers value walking the herd themselves every day — reading body condition, behaviour, and pasture state by eye. The collar removes some of that walking. For some farms that’s a feature; for others it changes a part of the practice they actively want to keep.
- Farms running mostly extensive set-stocking. The labour saving comes from frequent shifts. A farm that moves cattle once every three weeks doesn’t have a labour problem the collar solves.
How it sits inside the regenerative conversation
Tech and regenerative farming aren’t opposed. The point of regenerative grazing — animal impact concentrated in short bursts, long recovery for the pasture, soil biology rebuilt over years — is more about timing and stocking density than about whether the fence is wire or audio. A collar that makes high-density rotational grazing easier is, in principle, a tool that supports the practice.
The honest critique people make of any farm tech is whether it changes the farmer’s relationship with the animals and the land. Most regenerative graziers walking their herd daily would say that the walking is part of how they learn the farm. A collar takes some of that off the calendar. Whether that’s a good trade depends on what the farmer chooses to put back into the freed-up hours — more time on soil monitoring, planting, infrastructure, family — or whether the time just gets reabsorbed into more head per acre.
There’s also a question of who owns the data. A herd on collars is a herd whose movement, health, and grazing behaviour is being recorded continuously. The farmer is the customer, but the long-term aggregation of that data is a real strategic asset for the company holding it. Worth thinking about, the same way you’d think about any cloud-hosted business system.
Where Mangaroa sits
A note on this farm specifically, because the question comes up.
Mangaroa is a 3,000-acre mixed regenerative operation running Coopworth ewes and Angus and Angus-cross beef cattle in Whitemans Valley, Upper Hutt. The grazing system is rotational, run with electric fencing and direct shepherding by the team and dogs. We are not currently a Halter farm.
The reasoning is the same trade-off as above. Sheep are most of the stock count and Halter doesn’t yet do sheep. The beef rotation is already managed manually as part of how the team reads pasture and animal condition daily. The labour saving on the cattle side alone wouldn’t tip the economics for an operation our size, and the daily walking is part of the practice the team values.
That doesn’t mean it’ll be true forever. As the technology evolves — particularly if a viable sheep collar lands — the calculation may shift. For now, the practice here is closer to the older shepherd-and-electric-fence pattern than to a phone-app rotation, and the meat reflects that: pasture-finished, drafted on condition rather than truck schedule, butchered by a small Upper Hutt team rather than a large works.
What to take from this if you’re a NZ meat customer
The tag on the meat in your fridge doesn’t tell you whether the cattle wore a collar. It probably doesn’t tell you most of how the herd was grazed, either. If that matters to you, the questions worth asking the farm or the butcher are smaller and more specific than tech vs no tech:
- How often is the mob shifted, and on what residual?
- Is pasture allocated by eye, by app, or by cover-meter walk?
- How long is the recovery period before the same paddock is grazed again?
- How was the animal drafted to the works — by spec, by truck space, or by individual condition?
Halter is one tool inside that picture. It’s a piece of New Zealand-built ag-tech worth understanding because it’s becoming part of how the country farms cattle. Whether a particular farm uses it is a much smaller question than how that farm grazes, finishes, and handles its animals — and that’s the part that ends up on the plate.