Virtual fencing is a livestock management technique that uses GPS-enabled collars and software to create movable, invisible boundaries that guide or restrict animal movement without constructing physical fences. Animals approaching a virtual boundary receive escalating cues from the collar, typically an audio warning followed by a brief electrical stimulus if they continue, which conditions most animals to respond to the sound alone after a short training period. [2] [3]
Research on electronically mediated control of grazing animals dates back to the late 20th century, with prototypes and reviews describing non-physical boundaries enforced by aversive cues. [4] In Australia, the national science agency CSIRO began work on virtual fencing for cattle in the mid-2000s, research that later underpinned the eShepherd product. [5] In Norway, Nofence was established in 2011, and piloted collars with commercial goat herds before regulatory approvals for goats (2017) and later cattle and sheep. [6] In New Zealand, Halter began commercialising its virtual fencing and herding platform for pasture-based dairying in 2016 and later expanded into beef production and international markets. [7]
Most systems combine three elements: GPS location in a collar or neckband, wireless connectivity to transmit data and receive fence updates, and a farm or cloud application to draw, move and schedule boundaries. Signals to animals are delivered as a pre-warning sound and, if needed, a mild electrical pulse. Collars may connect via on-farm radio gateways with backhaul to the internet, or directly through cellular networks. [2] [3]
Virtual fencing is used to:
Peer-reviewed studies report that cattle typically learn to respond to the audio cue within days, with pulse exposure declining markedly after initial training. A 2024 Journal of Dairy Science study of Halter collars in an intensive dairy setting found effective containment and remote herding, with many cows receiving no pulses once trained. [18] In 2022 the United Kingdom’s Animal Welfare Committee concluded that virtual fencing can be used without detriment to welfare when safeguards are in place, including limits on stimulus, defined training protocols and monitoring. [19] [20] A 2023 literature review commissioned by the Australian Government found welfare risks are greatest during learning and can be mitigated by predictable cues, controllability and operator training, while noting the need for long-term data and evaluation for species beyond cattle. [21] The SPCA and other welfare organisations have questioned whether shock-based collars meet public expectations for animal welfare, calling for independent oversight and clearer standards. [22]
By removing or reducing the need for mesh or wire, virtual fencing may reduce wildlife entanglement and landscape fragmentation compared with permanent fences, while allowing managers to keep livestock out of waterways and wetlands in line with stock exclusion rules. [2] Field projects and catchment programmes have used virtual fencing to maintain high exclusion rates from riparian buffers and have monitored associated water quality improvements, including in Great Barrier Reef catchments. [23] On rangelands, land agencies have deployed towers to direct cattle after wildfires as part of adaptive post-fire grazing programmes. [24] [11]
Economic performance depends on terrain, labour, connectivity and the frequency of subdivision. Extension analyses suggest virtual fencing can be cost-effective where frequent moves or new subdivisions are required, offsetting capital and subscription costs. [25] Public vendor information indicates differing models; for example, a 2025 Nofence Ireland announcement set a cattle collar price of €339 with the first year’s subscription waived in that market, [6] while presentations hosted by USDA Climate Hubs have cited typical Vence gateway and per-collar costs for extensive rangeland deployments. [26]
Adoption has accelerated across pasture-based dairying and extensive beef systems.
Origin market for Halter, with uptake on dairy platforms and expansion into beef. In 2025 Halter raised a growth round valuing the company at US$1 billion for international expansion. [7]
States and territories regulate use. In 2025 Victoria announced draft regulations to permit virtual fencing for cattle by year end, [27] and New South Wales likewise announced plans to legalise virtual fencing for cattle across the state. [28] At federal level, an Australian Animal Welfare Guide for Virtual Fencing is being prepared via the AWTG. [29]
Virtual fencing is referenced in NRCS materials and used on private and public rangelands. The Bureau of Land Management and partners have installed towers to steer cattle for rangeland health and wildfire objectives. [2] [3] [24]
As of 2022, the United Kingdom had over 140 virtual fencing users, mostly for cattle, and uptake was expected to increase markedly. [19] A widely viewed example was shown in 2025 on Clarkson's Farm , where goats at Diddly Squat Farm wore Nofence collars to clear dense bramble and scrub in hard-to-fence areas. Virtual boundaries were drawn and adjusted in a smartphone app, allowing quick redeployment without temporary netting or permanent fencing, and helping popularise the approach among UK graziers. [30]
Virtual fencing is used under general animal-welfare law rather than a technology-specific code. In August 2025 the Government published its Cabinet response to a Petitions Committee report on holding an inquiry into virtual fencing, reflecting ongoing policy attention. [31] Separately, stock-exclusion regulations under the Resource Management Act 1991 require excluding specified stock from lakes, wide rivers and natural wetlands, which can be achieved by permanent fencing or by stock-management tools that prevent access, including virtual fencing. [32] [33]
Use is regulated by states and territories. In 2025 Victoria and New South Wales announced regulatory changes to enable commercial use for cattle, with a national welfare guide under development through the AWTG. [27] [28] [29]
The UK Animal Welfare Committee’s 2022 opinion sets out principles for acceptable use and safeguards. [19] [20] Some European countries restrict shock-based collars for livestock. Sweden has indicated that approval would require evidence of welfare outcomes equivalent to conventional electric fencing. [34] [35]
Fencing statutes are mainly state-based and predate virtual fencing, but federal conservation and land-management programmes reference the technology in prescribed grazing and rangeland projects. [2] [3] [36]
Empirical work has evaluated learning curves, pulse exposure and behavioural responses in dairy and beef cattle, finding high containment after acclimation and minimal ongoing stress under appropriate protocols. [18] [13] A 2023 Australian government review identified gaps in long-term welfare data, device reliability and evidence beyond cattle, [21] while a 2024 paper in Biological Conservation proposed ethical and policy frameworks for virtual fencing in biodiversity and land-management contexts. [37]
Concerns focus on the use of aversive stimuli, risks during training, device-failure modes and social acceptability in export markets. In New Zealand, welfare organisations such as the New Zealand SPCA have questioned whether shock-based collars meet public expectations and have proposed limits and monitoring requirements; in February 2024 the SPCA supported a parliamentary petition seeking an inquiry into whether shock-based virtual fencing for dairy cattle is acceptable to the public, consumers and export markets. [22] In August 2025 the New Zealand Government issued its response to the Petitions Committee. It noted the recommendation to include guidance on virtual fencing in the update of the Code of Welfare: Dairy Cattle, including use near milking sheds, and recorded that the draft Code had been returned to the National Animal Welfare Advisory Committee in 2024 to reconsider key areas. Evaluation materials outlined a proposed minimum standard that equipment be designed, constructed, maintained and used to minimise the likelihood of distress, pain or injury, with cattle that do not adapt to the audio cue managed by alternative means, and recommended best practice that aversive techniques not be used for training. [31]
{{cite web}}: |archive-url= is malformed: timestamp (help)CS1 maint: url-status (link){{cite journal}}: CS1 maint: DOI inactive as of October 2025 (link)