The Gates Foundation and Google are committing $100 million to expand artificial-intelligence tools for smallholder farmers. The goal is to quadruple their reach from 50 million to 200 million growers across Sub-Saharan Africa and South Asia. The two organisations announced the funding on 18 September, alongside technical support from Google’s own research teams.
Google.org vice-president Maggie Johnson said the partnership was about connecting frontier AI systems with “the farmers who feed a third of the world.” Gates Foundation chief executive Mark Suzman framed the goal in more basic terms. He argued that farmers deserve access to weather, crop and soil information regardless of where they farm, since that access can decide whether a harvest succeeds or fails.
What the tools actually do
Three tools sit at the centre of the expansion. A climate-forecasting system built on the TomorrowNow platform delivers real-time weather data tailored to specific farm locations. An Agricultural Understanding Platform maps field boundaries and identifies crops at sub-metre resolution, letting extension workers and lenders assess farmland more precisely than manual surveys allow. Open-source speech and text datasets spanning more than 40 African languages are meant to make the underlying advice usable by farmers who may not read English or type easily on a basic phone.
Google’s agricultural AI models have already mapped 140 million hectares of farmland across India and six African countries, giving the partnership a base to build on rather than starting from nothing. The named target countries include India, Kenya, Uganda, Ghana, Rwanda, Zambia and Nigeria, alongside Malaysia, Vietnam and Indonesia.
Field-boundary mapping at sub-metre resolution sounds technical, but it addresses a real problem. Without a reliable record of exactly which land a farmer works, banks and insurers have struggled to price risk or extend credit at scale. A more precise map of who farms what, and where, could in principle open access to financial products that have previously been available mainly to farmers with formal land title.
A patchy map of need
The reach of any single tool still depends heavily on where a farmer happens to live. In Uganda, one of the countries named in the rollout, the Karamoja region is facing one of the sharpest food-security deteriorations anywhere in East Africa this year. That is a reminder that better information for farmers does not by itself solve a drought-driven hunger crisis once it has already taken hold. AI-based forecasting can help a farmer plant and irrigate more precisely, but it cannot restore a harvest that has already failed.
The language-access component may end up mattering as much as the underlying forecasting technology. Voice-based tools built on speech datasets spanning more than 40 African languages are designed to reach farmers who might otherwise be excluded from digital advisory services. Those services were often built around English or a handful of dominant regional languages. Extension officers working across Sub-Saharan Africa have long identified language as one of the more stubborn barriers to getting agronomic advice to the farmers who need it most.
The Gates Foundation and Google have not set a fixed end date for the multi-year rollout, describing it instead as a roadmap that will scale as adoption grows. Independent verification of farmer uptake, rather than reach figures supplied by the funders themselves, will be the more useful test of whether the tools change outcomes on the ground. For now, the partnership represents one of the largest single commitments this year to closing the information gap that has long separated smallholder agriculture from the data-rich farming common in wealthier regions.
