New Zealand’s farmers have a better productivity record than almost anyone else in the economy. On the next technology wave, they are behind. AWS research reported by RNZ puts AI adoption across agriculture, forestry and fishing just under 40 percent, against 50 percent nationally, with most adopters stuck at the most basic stage.
That gap matters more on a sheep and beef station than in a law office. Labour is scarce, compliance keeps stacking up and margins are hostage to commodity prices. Those are exactly the pressures automation should be relieving.
This is not a sector that fears technology
In 2025, Stats NZ reported that primary industries labour productivity rose 9.8 percent in the year to March 2024, while the wider measured sector went backwards by 0.7 percent. Over the long run, the same release showed primary sector workers were producing 72 percent more per hour than in 1996, while services and goods-producing industries slipped.
Three decades of outperformance came from adopting new methods faster than everyone else. That makes the AI lag a skills and awareness problem, not a cultural one. It is also why it is fixable.
The valuable AI is the kind nobody has tried
AWS says physical AI, meaning robotics, autonomous systems and computer vision, offers the biggest gains for crop management, livestock monitoring and processing. Only 30 percent of agricultural businesses have heard of it, just 7 percent of those are using or piloting it, and 9 percent feel ready for next-generation tools.
AWS NZ country manager Manuel Bohnet told RNZ that “only 10 percent have really moved to an advanced usage and this is where you also really get most of the value out of it.” His diagnosis is skills: confidence follows understanding.
The payback is already showing up in the paddock
The returns are not theoretical. Lone Star Farm stock manager Carey Pawson-Edwards told RNZ its five South Island farms run more than 30 apps, and that Halter collars on cows with calves at foot are delivering weaning weights 10 kilos heavier. EID tags also unlock the Lumina Lamb premium programme, where traceability is a condition of entry. Data, in other words, is now a price lever, not just a management tool.
The NZ Herald reports the average Halter farmer saves 27 to 28 hours a week, spray drones finish in an hour what took six hours by tractor on hill country, and Ballance’s SpreadWise uses computer vision to map where fertiliser should not go. AgriTechNZ chief executive Brendan O’Connell says AI “has stopped being a future promise” for farmers.
The Herald’s sharper point is that cost and connectivity are no longer the obstacles. Starlink fixed the second. The real barriers are knowing which tools deliver value, and farm data scattered across systems that do not talk to each other.
Dairy is cleaning up its data first
In March 2026, Farmers Weekly reported that dairy farmers using generative AI were still few but growing, with some building custom tools for farm-specific problems. The underlying Perrin Ag report for DairyNZ, published in December 2025, fed a clear priority: get farm and sector data in order before chasing broader adoption. GenAI is already being embedded in DairyNZ’s DAiSY, Beef + Lamb NZ, FAR and Xero platforms, which lowers the bar for farmers who will never open ChatGPT.
The real risk is who keeps the $15 billion
This is where slow adoption stops being an efficiency story. MPI estimates at least $15 billion in potential value sits in the primary sector value chain, on top of 20 to 30 percent gains available inside the farm gate. MPI chief insight officer Jarred Mair argues new technology means productivity and sustainability no longer have to be traded off.
The warning attached is blunt. MPI says that value risks being captured by overseas financiers, on-sellers and payment platforms if New Zealand does not move quickly. Intermediaries who deploy AI faster will set the terms. Farmers who wait will be price-takers in a new layer of the supply chain.
Strategy beats enthusiasm
The national data shows what separates winners from dabblers. AWS’s September survey found 84 percent of AI adopters report higher revenue, but only 15 percent of businesses have a formal AI strategy. Those that do are more than twice as likely to reach advanced use, 41 percent against 19 percent.
New Zealand SMEs were already starting from behind. In 2025, MBIE cited survey data showing 68 percent of SMEs had no plans to evaluate AI, against 38 percent in Australia. The same MBIE report noted Halter’s USD1 billion valuation, proof the country can build world-class farm AI even as many farms are slow to use it.
What happens next
The fixes are already in motion. AWS is running a free agritech school, and sector bodies are wiring AI into tools farmers already use. The job for farm owners and processors is to treat AI like any other capital decision: pick the narrow applications with a proven return, integrate the data, and write down a plan. The sector earned its productivity lead by moving early. Leaving AI to the office economy would be the first time in 30 years it chose to wait.
Sources
- RNZ: Agriculture industry slow to take up new AI tech (2026-10-08)
- NZ Herald: From drones to data – How AI is transforming New Zealand farming (2026-07-25)
- Farmers Weekly: Farmers using artificial intelligence to save time (2026-03-05)
- Farmers Weekly: Technology opens new path to doubling primary exports (2026-08-26)
- DairyNZ / Perrin Ag: Opportunities of artificial intelligence (GenAI) for New Zealand dairy (2025-12)
- Stats NZ: Productivity statistics 1978-2024 (2025-05-16)
- AWS: Kiwi businesses are turning AI ambition into impact one year in (2026-09-22)
Join the discussion
Add useful context, ask a good question, or challenge an idea — keep it specific and respectful.