AI in Agriculture in Aotearoa New Zealand: A Living Whitepaper
Updated 2 April 2026
Introduction
AI in New Zealand agriculture is now best understood as a set of working production systems rather than a single technology trend. The strongest adoption remains in dairy, where AI is already embedded in virtual fencing, animal monitoring, mastitis detection, pasture measurement, and advisory tools. Horticulture is following a different path, with adoption clustering around robotics, disease-risk modelling, weather intelligence, optical grading, and post-harvest automation. Public-sector and research-led AI is also becoming more important, especially in emissions measurement, weed and pest risk analysis, and climate adaptation. (mpi.govt.nz)
This matters in a sector with unusually high economic leverage. MPI’s latest Situation and Outlook for Primary Industries says food and fibre export revenue is forecast to reach NZ$62 billion in the year to 30 June 2026, including NZ$27.4 billion from dairy and NZ$9.2 billion from horticulture. At the same time, the Government’s July 2025 AI strategy explicitly backs an adoption-led approach and points to agriculture as a core area where New Zealand can tailor AI to local conditions rather than compete in frontier-model development. (mpi.govt.nz)
Executive Snapshot
- Dairy remains the commercial centre of gravity. The most mature AI deployments are in livestock control, reproduction and health alerts, milking-shed monitoring, pasture estimation, and farmer-facing knowledge tools. (halterhq.com)
- Adoption is substantive but uneven. DairyNZ’s AI guidance and conference material show that farmers are seeing real value, but trust, integration, workflow fit, and data quality still determine who moves beyond experimentation. (dairynz.co.nz)
- Horticulture is narrower, but highly strategic. The leading use cases are orchard robotics, disease-risk prediction, downscaled weather forecasting, AI-assisted sorting, and biosecurity preparedness. (roboticsplus.co.nz)
- Public-good AI is becoming core sector infrastructure. AgResearch and the Bioeconomy Science Institute are using AI in livestock genomics, weed-risk assessment, weed control, and emissions platforms, while MBIE is now advancing a national AI research platform with bioeconomy applications at its centre. (agresearch.co.nz)
- The next competitive edge is interoperability. The sector has no shortage of farm data; the bottleneck is joining sensors, apps, software and advisory systems into workflows farmers can trust and act on. (dairynz.co.nz)
What Changed Since 9 March 2026
- A new national AI-platform contest has moved closer to decision point. MBIE says phase-two proposals for the Artificial Intelligence Research Platform were due 31 March 2026, with platform funding decisions scheduled for late April 2026 and investment decisions in May 2026. One shortlisted proposal, led by the Bioeconomy Science Institute, is explicitly focused on agriculture, aquaculture and forestry; another, from Waikato and Canterbury, proposes “Outdoor AI” with initial emphasis on horticulture and agriculture. (mbie.govt.nz)
- Aimer has shifted from fundraising story to scaling story. In late March 2026 the company appointed a new CRO and said 2026 priorities include expanding its customer base and rolling out satellite- and drone-based pasture assessment alongside its smartphone computer-vision platform. (aimer-farming.com)
- Wearables have become a more visible mainstream dairy topic. DairyNZ’s February–April 2026 coverage and March 2026 field event both point to stronger farmer interest in wearable data, integration with herd systems, and ROI-driven adoption decisions. (dairynz.co.nz)
Current News and Market Developments
1) Livestock AI has clearly reached scale
Halter remains the clearest scaled example of AI in New Zealand farming. Its current materials say the platform is used on approximately 700,000 animals as at January 2026, while other company pages say it is trusted by over 1,000 farmers and is operating across New Zealand, Australia and the United States. That makes Halter less a pilot technology than a live farm operating system for virtual fencing, animal movement, pasture allocation, reproduction workflows, and health monitoring. (halterhq.com)
The significance is strategic: Halter shows that agricultural AI diffuses fastest when it is embedded inside a workflow farmers already value every day—moving stock, managing grazing pressure, protecting waterways, and responding to reproductive or health signals. (halterhq.com)
2) Pasture intelligence is becoming a stronger category
Aimer Farming continues to stand out because it matches AI to New Zealand’s pasture-based production model rather than importing a row-crop playbook. Its platform uses smartphone-based computer vision and proprietary models to estimate and forecast pasture, generate feed wedges, and support grazing and supplement planning. In September 2025 the company said its Gen3 model improved measurement accuracy by 5–10%, with around 93% of pasture covers falling within 200 kg DM/ha of a platemeter value. In March 2026 it signalled the next step: scaling customer adoption and extending into satellite and drone-based measurement. (aimer-farming.com)
Another important signal is commercial behaviour around the company. Aimer, Bovonic and Herd-i jointly ran on-farm technology showcases in late 2025 and said more events were planned for 2026, suggesting the market is moving toward bundled demonstrations of complementary tools rather than one-off point solutions. (aimer-farming.com)
3) In-shed animal-health AI is showing the kind of ROI mainstream farmers notice
Bovonic’s QuadSense remains one of the stronger proof points for narrow, high-value dairy AI. Bovonic says more than 4,000 units have been installed across more than 160 farms in New Zealand and Ireland. Its 2025 validation survey of 33 New Zealand farms reported average annual benefits of NZ$29,547, with payback in roughly six months, alongside lower bulk SCC, labour savings, and reduced antibiotic use. (bovonic.com)
That is strategically important because it highlights how adoption is happening: many farmers are not buying “AI” in the abstract; they are buying faster mastitis detection, lower penalties, easier milkings, and fewer missed cases. (bovonic.com)
4) Dairy institutions are now active AI deployers, not just observers
DairyNZ has moved from studying AI to deploying it. Its DAiSY assistant now draws on more than 1,100 pages and 880 tools and resources from the DairyNZ knowledge base and responds using only DairyNZ website content. Separately, DairyNZ’s December 2025 International Precision Dairy Farming Conference brought together more than 400 people from 22 countries, including 90 New Zealand dairy farmers, with AI, sensors, automation and farm-systems modelling central to the programme. (dairynz.co.nz)
This matters because institution-led AI often accelerates trust. In agriculture, adoption usually rises when tools are anchored in familiar advisory brands, sector-owned data, and transparent source material. (dairynz.co.nz)
5) Horticulture AI is consolidating around robotics, forecasting and quality control
Robotics Plus remains the clearest New Zealand horticulture AI story at scale. Its February 2025 acquisition by Yamaha Motor formed the base of Yamaha Agriculture, with Prospr positioned as an autonomous platform for orchard and vineyard operations such as spraying and weed control. Robotics Plus said the company would retain headquarters and IP development in Tauranga while scaling globally. (roboticsplus.co.nz)
Beyond field robotics, horticulture is also moving on data and sensing. Horticulture New Zealand’s coverage of the DeepWeather project described a machine-learning-enhanced weather model designed to produce 1 km forecasts from lower-resolution weather modelling, with intended use cases including spray timing, frost response, irrigation and harvest planning. In post-harvest operations, February 2026 industry coverage also highlighted rising use of optical grading, automation and AI-enabled camera systems in packhouses as capacity and labour constraints intensify. (hortnz.co.nz)
Research Overview
DairyNZ / Perrin Ag: GenAI is useful now, but the bottlenecks are practical
DairyNZ’s AI guidance, based on its 2025 Perrin Ag work, says current farmer use is strongest in decision support, task enhancement, communications and analysis of farm-specific data. It also points to the next 3–5 years being shaped by deeper integration with farm software, more ready-to-use AI workflows, and more real-time decision support. (dairynz.co.nz)
The message is not “AI is coming”; it is that mainstream diffusion still depends on better integration, trusted outputs, and usable workflows. DairyNZ’s own conference summary states this directly: AI has promise, but adoption requires trust, integration and clear value. (dairynz.co.nz)
Precision dairy research: integration is the real problem to solve
A 2025 conference case study reported in the DairyNZ precision-dairy proceedings found Southland farmers were commonly monitoring 4 to 7 applications across weather, soil moisture, pasture and livestock. Interest in an AI chatbot centred on bridging silos, accessing historical farm data in natural language, automating compliance reporting, detecting anomalies across streams, and improving interaction with fragmented digital systems. (dairynz.co.nz)
This is one of the most important signals in the market: the next wave of value is likely to come less from adding new sensors and more from making existing data legible across systems. (dairynz.co.nz)
AgResearch and the Bioeconomy Science Institute: AI is becoming research infrastructure
AgResearch’s 2025 annual report says it had more than 50 AI-focused research projects under way. A flagship example was the use of AI to analyse historical CT-scan data, allowing researchers to revisit more than 20 years of livestock imaging and extract traits linked to methane emissions, feed efficiency, animal welfare and meat quality. (agresearch.co.nz)
AgResearch is also advancing AI beyond analytics into field systems. Its Map and Zap® platform uses AI for weed identification and targeted laser control, while a separate weed-risk programme has automated literature discovery and assessment methods to speed weediness analysis and extend the same approach into insect pests and animal parasites. (agresearch.co.nz)
Institutionally, the change is larger than any one project. AgResearch now sits within the Bioeconomy Science Institute, formed on 1 July 2025 by combining AgResearch, Manaaki Whenua, Plant & Food Research and Scion into a single bioeconomy-focused organisation. (agresearch.co.nz)
Horticulture and biosecurity research is gaining sharper AI use cases
Lincoln Agritech’s STELLA project in Hawke’s Bay is one of the clearest horticulture examples. It is using automated spore sampling, UAV and satellite imagery, environmental monitoring and AI-powered risk models to improve early detection of bull’s-eye rot in apples. (lincolnagritech.co.nz)
At the same time, MPI, the Bioeconomy Science Institute and the Ag Emissions Centre launched the On-Farm Emissions Calculator in October 2025 as the first tool to apply the government-mandated on-farm emissions standard released in December 2024. The calculator is designed as a standardised platform that other agri-tech providers can embed, which makes it potentially important infrastructure for future AI-driven decision support. (agresearch.co.nz)
Case Studies
Halter: AI as a farm operating system
Halter is the clearest case of AI embedded in routine livestock management at scale: virtual fencing, remote shifting, pasture allocation, reproduction insights and health alerts, now across roughly 700,000 animals. (halterhq.com)
Aimer Farming: low-friction AI for pasture-based systems
Aimer’s importance lies in reducing hardware friction. Using a smartphone-first model, it brings computer vision into everyday grazing decisions and is now extending into drone and satellite assessment. (aimer-farming.com)
Bovonic QuadSense: narrow AI with clear economics
QuadSense shows why focused animal-health AI is commercially persuasive: measurable ROI, fast payback, and minimal workflow disruption. (bovonic.com)
Robotics Plus: orchard autonomy with global backing
Robotics Plus demonstrates that New Zealand horticulture AI can progress from local development to global platform ownership, with autonomy and AI tied directly to labour-saving orchard operations. (roboticsplus.co.nz)
Lincoln Agritech / STELLA: AI for export-risk management
STELLA is a strong example of AI being used to reduce invisible crop risk with direct export implications for premium fruit supply chains. (lincolnagritech.co.nz)
Core Trends
1) AI adoption is strongest where data is already digital
Dairy leads because the sector already has collars, herd databases, milking systems, reproduction records and advisory software that produce machine-readable data. (dairynz.co.nz)
2) The market is shifting from tools to workflows
The strongest products are no longer standalone “AI tools”; they are systems embedded in grazing, milking, advisory, emissions and orchard-management workflows. (dairynz.co.nz)
3) Horticulture AI is narrower but more capital-intensive
Compared with dairy, horticulture adoption is concentrated in higher-value operations where labour scarcity, crop quality and export risk justify robotics, AI-assisted grading and targeted disease forecasting. (roboticsplus.co.nz)
4) Public-good AI is moving into the operating layer of the sector
Standardised emissions tools, weed-risk automation, laser weed control, and bioeconomy-oriented AI research platforms suggest AI is becoming part of agricultural infrastructure, not just private software. (agresearch.co.nz)
Constraints and Risks
The barriers are now fairly clear:
- Trust and validation: farmers need systems that are explainable, source-grounded and reliable in real farm conditions. (dairynz.co.nz)
- Interoperability: data sits across many apps and devices, creating cognitive and operational friction. (dairynz.co.nz)
- SME adoption gap: MBIE says larger firms are moving faster, while many SMEs still have no plans to evaluate or invest in AI. (mbie.govt.nz)
- Proof of value: narrow, ROI-visible systems are spreading faster than general-purpose tools. (bovonic.com)
- Capability and commercialisation: New Zealand’s national response is still evolving, with the AI research platform decision not due until late April–May 2026. (mbie.govt.nz)
Conclusion
As of 2 April 2026, AI adoption in New Zealand agriculture is best described as selective, practical and increasingly infrastructure-like. Dairy remains the most commercially advanced segment, led by scaled livestock systems, pasture intelligence, wearables and focused animal-health tools. Horticulture is progressing through robotics, weather intelligence, disease prediction and post-harvest automation rather than broad-based farm copilots. (halterhq.com)
The most important shift since the earlier update is not a single product launch. It is that AI is now being reinforced simultaneously by commercial deployment, sector institutions, and national research-policy infrastructure. With MBIE’s AI platform decision approaching, DairyNZ moving into AI-enabled advisory delivery, and the Bioeconomy Science Institute expanding applied research, the sector is moving from isolated innovation toward a more durable AI stack for agriculture. (mbie.govt.nz)
The near-term winners are likely to be the same as in the last edition, but the case is now stronger: systems that combine embedded workflow fit, trusted domain grounding, interoperable data, and clear on-farm economics are the ones most likely to move from early adopters into the New Zealand agricultural mainstream. (dairynz.co.nz)