Nigeria Strawberry Pilot Reports 40% Yield Gain From Farm-Side AI
By AgriLinkage Technology | Published 20 October 2026 | Production intelligence
A completed strawberry-farming deployment in Nigeria has reported a 40% increase in yield, a 15% reduction in post-harvest waste and a rise in market-grade fruit from roughly 70% to 88%. The results were disclosed on 19 October by Nottingham Trent University after the SmartBerry knowledge-transfer project closed.
The numbers make SmartBerry more useful than another demonstration of sensors in a greenhouse: they link a relatively modest, farm-side technology package to saleable output and business turnover. But they should be read as project-reported results from one commercial partnership, not as a peer-reviewed estimate that can be transferred unchanged to other farms, cultivars or climates.
What the SmartBerry system changed
SmartBerry combined greenhouse and field sensors with artificial intelligence and edge computing. The sensor network measured temperature, humidity, soil moisture and light intensity inside and outside protected growing areas. Those measurements were used to support decisions about irrigation, plant stress and growing conditions instead of relying only on fixed watering schedules or visual inspection.
The practical distinction is where the computing happens. Edge systems analyse data close to the farm rather than sending every reading to a distant cloud service. That reduces dependence on continuous internet access, lowers the amount of data that must be transmitted and can shorten the time between a change in soil moisture and a usable irrigation signal.
The project was a partnership among Nottingham Trent University, Ashley's Strawberry Farms and Covenant University. The official UK Research and Innovation project record lists £245,725 of funding, a February 2024 to January 2026 funded period and a closed status. That record independently confirms the project, budget and timetable, although the operating results themselves are reported by the partners.
The reported production and commercial results
According to the university's completion account, strawberry yield increased by 40%, post-harvest waste fell by 15%, and the market-grade share rose by about 18 percentage points, from approximately 70% to 88%. In relative terms, that market-grade share increased by about 26%. The participating business's turnover was reported to have risen from around £18,000 to £32,000, an increase of roughly 78%.
Those measures capture different parts of the production chain. Yield indicates more fruit harvested. Market-grade share indicates that a larger portion met sale specifications. Lower post-harvest waste means more harvested fruit remained saleable. Turnover then reflects the combined commercial effect, but it can also move with prices, volumes, product mix and market expansion, so it cannot be attributed to the sensor system alone without fuller accounts.
The project also reported that Camarosa strawberries were produced in both optimal and less favourable seasons, in open-field and controlled-greenhouse settings. If repeatable, extending the fruiting window could be economically important because it changes when farms can supply buyers, not just how many kilograms they produce.
Why this is a credible signal, but not yet a universal benchmark
The project earned an Outstanding, or A, grade in its final report, according to Nottingham Trent University. The Innovate UK African Agriculture Knowledge Transfer Partnership programme was designed to move applied know-how into a business, not to run a blinded agronomic trial. Its eligibility rules required collaboration among a UK knowledge base, an African business and an African knowledge base, with projects lasting 18 to 24 months.
That design matters when interpreting the evidence. The Innovate UK competition framework emphasised strategic innovation and commercial exploitation. It did not require the replicated plots, untreated controls, statistical power or multi-location validation expected from a peer-reviewed field experiment.
Agrilinkage therefore treats the 40%, 15% and 88% figures as verified statements made by the project partners, not independently reproduced agronomic effects. Before adopting the same system, another grower would still need baseline yield and waste records, a comparable control area, water-use measurements, equipment and maintenance costs, labour-time data and at least one full production cycle under local conditions.
The decision value for growers and technology buyers
The most defensible lesson is not that artificial intelligence automatically raises strawberry yields by 40%. It is that a narrow, measurable deployment can be judged against farm outcomes that matter: fruit per area, grade-out rate, waste, water use, season length and revenue. That is a stronger procurement framework than buying a generic farm dashboard and assuming that more data will improve production.
For growers evaluating a similar system, the first comparison should be a small instrumented block against current practice. Irrigation advice should be logged alongside actual water applied, disease or stress observations, harvested weight by grade and rejected fruit. The system's value is the additional margin after sensors, communications, power, maintenance, training and crop-management changes—not the model's prediction accuracy in isolation.
Local processing also addresses a common constraint in digital agriculture: weak or expensive connectivity. A farm can keep core monitoring functions running during network outages and transmit summaries when a connection becomes available. The trade-off is that edge hardware must survive heat, humidity, dust and inconsistent power, and someone must maintain both the sensors and the decision rules.
What happens next
The researchers say they want to extend SmartBerry toward predictive climate modelling, greenhouse control and yield forecasting, and to test the approach on other crops. Those are plausible extensions, but each adds operational and validation risk. Automating a greenhouse response, for example, raises the cost of a false signal compared with simply alerting a manager.
Agrilinkage has separately examined a different strawberry-production intervention: an NC State textile cover that raised yield in tunnel tests. The two cases should not be compared as competing yield claims. One is a digital monitoring and management deployment in Nigeria; the other is a physical crop-protection treatment tested under different conditions.
The useful next evidence from SmartBerry would be a transparent methods and results paper: planted area, cultivar and plant density; baseline and comparison seasons; irrigation volumes; input and labour costs; grade definitions; weather; and statistical uncertainty. Until then, the project is best understood as a promising commercial-scale signal that farm-side sensing can improve the timing and quality of decisions—not proof of a universal yield response.
Methodology and sources
This article was prepared by the Agrilinkage Technology desk from the 19 October 2026 Nottingham Trent University project report, the UKRI grant record and the Innovate UK programme rules. Calculations of the relative change in market-grade share and turnover use the reported before-and-after values. Agrilinkage did not interview the participants or independently test the system.
Comments