AI in Agriculture in Aotearoa New Zealand: A Living Whitepaper
Updated 9 March 2026
Introduction
AI adoption in New Zealand agriculture has moved beyond concept-stage discussion and into practical deployment, especially in dairy, pasture management, animal health, horticulture robotics, weather intelligence, and biosecurity. That shift matters in a sector operating inside New Zealand’s wider food-and-fibre economy, which MPI forecasts will reach NZ$62 billion in export revenue in the year to 30 June 2026. MPI’s current strategy also places explicit emphasis on growing the value of the food and fibre sector, reducing on-farm emissions, and protecting the country from pests and diseases—all areas where AI is increasingly being applied. (mpi.govt.nz)
At the national level, the broader AI environment has also become more supportive. The AI Forum’s March 2025 productivity survey reported that 82% of surveyed organisations were using AI, 93% saw worker-efficiency gains, and 56% reported positive financial impact. In parallel, the New Zealand Government’s July 2025 AI strategy explicitly favours an adoption-focused approach rather than attempting to compete in frontier-model development, and it singled out agriculture as a flagship application area, citing companies such as Halter and Aimer Farming. (aiforum.org.nz)
Executive Snapshot
- Dairy is the clear commercial leader for AI adoption in New Zealand agriculture, with real traction in virtual fencing, herd monitoring, pasture measurement, mastitis detection, and advisory copilots. (halterhq.com)
- Horticulture is advancing through precision use cases rather than broad farm copilots: weather downscaling, disease-risk modelling, orchard robotics, and machine vision for fruit detection are the strongest themes. (hortnz.co.nz)
- Public-good AI is rising alongside commercial AI, especially in emissions accounting, weed-risk assessment, biosecurity surveillance, and climate adaptation. (agresearch.co.nz)
- Adoption is real, but uneven. DairyNZ’s 2025 GenAI research found meaningful benefits for current users, but adoption remains concentrated among innovators and early adopters rather than the farming mainstream. (dairynz.co.nz)
- The biggest constraints are not model capability alone; they are trust, interoperability, proof of value, digital literacy, and access to high-quality farm data. (dairynz.co.nz)
Current News and Market Developments
1) Livestock AI has reached scale
Halter remains the most visible New Zealand example of scaled agricultural AI. In June 2025, the company announced a Series D raise of NZ$155 million at a NZ$1.55 billion / US$1 billion valuation. By July 2025 it said more than 1,000 farmers were using its system, and by January 2026 Halter said its virtual-fencing and animal-guidance platform was being used on approximately 700,000 animals across hundreds of farms. (halterhq.com)
That scale matters because Halter shows what AI adoption looks like when it is embedded inside a workflow farmers already value: pasture allocation, remote herding, heat detection, health alerts, and grazing control. It also shows how AI in New Zealand agriculture is being commercialised first through operational systems, not abstract analytics. (halterhq.com)
2) Pasture intelligence is becoming a competitive category
Aimer Farming is emerging as another important local case. In October 2025, Aimer announced a NZ$750,000 investment from Cultivate Ventures, on top of about NZ$1.4 million raised between November 2024 and June 2025, to expand its mobile, drone, and satellite roadmap. Its core platform uses smartphone-based computer vision to estimate pasture mass and generate feed wedges and grazing plans. (aimer-farming.com)
The strategic significance is twofold: first, Aimer lowers the hardware barrier to precision pasture management; second, it fits New Zealand’s pasture-based production model unusually well. That makes it one of the clearest examples of AI being adapted to local farm systems rather than imported unchanged from overseas row-crop contexts. (aimer-farming.com)
3) In-shed animal-health AI is showing measurable ROI
Bovonic’s QuadSense is one of the stronger near-term proof points for practical dairy AI. Bovonic says more than 4,000 units have been installed across more than 160 farms in New Zealand and Ireland, and its 2025 validation survey of 33 New Zealand farms found an average annual benefit of NZ$29,547, with payback in about six months. Reported gains included labour savings, lower bulk somatic cell count, reduced antibiotic use, and improved milk-quality outcomes. (bovonic.com)
This is important because it demonstrates that AI adoption is not only about “copilots” or chat tools. It is also showing up as focused, outcome-specific automation around costly biological problems, where return on investment is easier to see and measure. (bovonic.com)
4) Horticulture robotics is consolidating and globalising
In February 2025, Robotics Plus announced its acquisition by Yamaha Motor to form the foundation of Yamaha Agriculture. The company’s Prospr platform combines autonomy, sensing, and AI-powered data workflows for orchard and vineyard operations such as spraying and weed control, while retaining headquarters and IP development in Tauranga. (roboticsplus.co.nz)
This is a notable signal for New Zealand agriculture because it shows local AI and robotics moving from pilot status into global platform ownership and distribution. It also reflects a pattern in horticulture: AI adoption is strongest where labour shortages, repetitive field operations, and crop-value density justify capital investment. (roboticsplus.co.nz)
5) Sector institutions are now deploying AI directly
DairyNZ’s December 2025 Precision Dairy Farming Conference drew more than 400 participants from 22 countries, including 90 New Zealand dairy farmers, with AI, automation, sensors, and farm-systems modelling all central themes. Around the same period, DairyNZ launched DAiSY, an AI-powered website assistant built on more than 1,100 pages and 880 tools/resources from its own knowledge base. (dairynz.co.nz)
That matters because it shows AI adoption moving from startups alone into the sector’s knowledge institutions. In practice, this tends to accelerate trust and uptake, because farmers are more willing to use AI when it is grounded in familiar sources and existing advisory channels. (dairynz.co.nz)
Research Overview
DairyNZ / Perrin Ag: GenAI is useful now, but still early
DairyNZ’s December 2025 report on GenAI for dairy farmers is one of the most important current research snapshots. It found that New Zealand dairy farmers already using GenAI are applying it for:
- decision support,
- task enhancement,
- communication support, and
- increasingly, contextual analysis of farm-specific data such as spreadsheets, photos, wearables, and feed information. (dairynz.co.nz)
The report also found that mainstream diffusion remains limited. Farmers in the innovator and early-adopter segments were building tailored chatbots, digital twins, and advanced workflows surprisingly quickly, but broader adoption was still low and constrained by proof-of-concept concerns, integration friction, and trust. Use cases that had crossed into the early majority were mainly simpler tasks such as drafting documents, transcribing meetings, and basic Q&A. (dairynz.co.nz)
Precision dairy research: the problem is data fragmentation, not data scarcity
A 2025 conference case study presented through the Precision Dairy Farming proceedings described interviews with Southland dairy farmers who routinely monitored 4 to 7 different applications spanning weather, soil moisture, pasture, and livestock. The central problem was not lack of data, but the cognitive burden of combining it. The study’s AI chatbot prototype attracted interest around five capabilities: bridging information silos, natural-language access to historical farm data, automated compliance reporting, anomaly detection across multiple data streams, and more adaptive farmer-facing interaction. (dairynz.co.nz)
This is one of the clearest signals of where the next wave of on-farm AI value is likely to come from in New Zealand: not just sensing, but sense-making across tools. (dairynz.co.nz)
AgResearch and the Bioeconomy Science Institute: AI is becoming research infrastructure
AgResearch’s 2025 annual reporting said it had more than 50 AI-focused projects under way, including AI analysis of historical CT scan data to unlock livestock-genomics insights relevant to methane, feed efficiency, animal welfare, and meat quality. Separately, AgResearch also described work with the University of Waikato’s AI Institute on explainable AI for animal-health and welfare research, including automated behavioural analysis. (agresearch.co.nz)
The newly formed Bioeconomy Science Institute, launched on 1 July 2025, is now the main public-research vehicle behind much of this applied work across agriculture, horticulture, forestry, aquaculture, and biotechnology. That institutional change makes AI less of a side project and more of a system capability across the primary sector. (agresearch.co.nz)
Horticulture and biosecurity research is gaining momentum
Horticulture New Zealand’s April 2025 coverage of the DeepWeather project described AI-assisted weather downscaling aimed at improving forecasting at much finer resolution for New Zealand’s microclimates. The intended agricultural benefits include better irrigation, spray, harvest, labour, and frost planning. (hortnz.co.nz)
Lincoln Agritech’s work in the STELLA Horizon Europe project is another important 2025 development. In Hawke’s Bay apple orchards, it is combining automated spore sampling, UAV and satellite imagery, environmental monitoring, and AI-powered risk models to detect bull’s-eye rot earlier—an issue with major export and reputation implications for a sector where New Zealand apple exports reached a record NZ$1 billion in 2025. (lincolnagritech.co.nz)
Case Studies
Halter: AI as a farm operating system
Halter shows the strongest example of AI embedded in day-to-day livestock management: virtual fencing, remote animal movement, fertility and health monitoring, and pasture decision support, all delivered through collars, towers, and mobile software at large scale. The company’s funding round and January 2026 deployment figures suggest this is no longer an early pilot story. (halterhq.com)
Aimer Farming: low-friction precision pasture management
Aimer’s model is notable because it uses everyday smartphone hardware and computer vision to turn pasture measurement into a more accessible workflow. That makes AI usable in a grazing system where time, hardware cost, and repeatability are major adoption barriers. (aimer-farming.com)
Bovonic QuadSense: narrow AI with clear payback
QuadSense is a strong example of targeted AI adoption with measurable benefits. Its value proposition is simple—earlier mastitis detection in the shed—and that narrow use case appears to be helping it achieve faster commercial traction than more general-purpose AI products. (bovonic.com)
Robotics Plus: orchard AI plus autonomy
Robotics Plus represents the horticulture side of the market: AI linked to machinery, navigation, spraying, and input optimisation rather than text-based advisory tools. Its Yamaha transaction suggests international buyers see New Zealand specialty-crop robotics as commercially significant. (roboticsplus.co.nz)
Lincoln Agritech / STELLA: AI for export risk reduction
The STELLA work is a good example of AI addressing a New Zealand-specific crop problem with direct export consequences. It also shows how international R&D partnerships are extending local capability in precision disease surveillance and decision support. (lincolnagritech.co.nz)
Core Trends
1) AI adoption is strongest where data is already digital
Animal health, reproduction, and feed are leading use cases partly because those areas already generate machine-readable data through wearables, milk systems, herd databases, and software platforms. DairyNZ’s GenAI work explicitly found the heaviest current use in those domains. (dairynz.co.nz)
2) The market is shifting from tools to workflows
The clearest direction of travel is from standalone AI interactions toward embedded workflows inside farm software, sector portals, and operational systems. DairyNZ’s report points to this directly, arguing that embedded GenAI and off-the-shelf workflows may diffuse more easily than standalone large-language-model use. (dairynz.co.nz)
3) Horticulture AI is narrower but high-value
Compared with dairy, horticulture in New Zealand appears less focused on general GenAI adoption and more focused on high-value applications: weather, disease detection, robotics, crop mapping, and precision operations. That fits the economics of orchard and vineyard systems, where labour shortages and quality risk can justify specialised systems. (hortnz.co.nz)
4) Public-good AI is becoming strategically important
AI is increasingly being used not only to improve profitability, but to support emissions measurement, biosecurity, weed surveillance, and climate resilience. Examples include the On-Farm Emissions Calculator, AI-assisted weed-risk assessment across more than 20,000 exotic species, and AI-enhanced weather forecasting. (agresearch.co.nz)
Constraints and Risks
The main barriers are now relatively clear:
- trust and validation — especially because GenAI can hallucinate or overstate confidence; (dairynz.co.nz)
- interoperability — farmers often have many data sources but weak cross-system integration; (dairynz.co.nz)
- digital capability — value depends on prompt quality, output checking, and workflow design; (dairynz.co.nz)
- proof of value — mainstream farmers want visible benefit before investing time and changing routine; (dairynz.co.nz)
- data quality — AI readiness depends on complete, accurate, usable farm records. (dairynz.co.nz)
The Government’s current AI strategy may reduce some uncertainty by maintaining a light-touch, principles-based policy posture, but that does not remove the practical farm-level adoption barriers. (beehive.govt.nz)
Conclusion
AI in New Zealand agriculture is no longer a placeholder topic. It is already established in specific, economically meaningful workflows—especially in dairy—and is steadily expanding across horticulture, weather intelligence, biosecurity, and public research. The strongest current evidence points to a sector moving from experimentation to operational integration, led by companies like Halter, Aimer, Bovonic, and Robotics Plus, and reinforced by institutions such as DairyNZ, AgResearch, Lincoln Agritech, and MPI-linked research infrastructure. (halterhq.com)
The overall picture is best described as selective but substantive adoption. New Zealand agriculture is not yet in a phase of universal AI penetration, and GenAI use on farm remains early for much of the market. But where AI is tied to concrete workflows, trusted data, and visible ROI, adoption is already material. The near-term winners are likely to be systems that combine three things: embedded usability, domain trust, and interoperable farm data. (dairynz.co.nz)