Mineral’s work to live on with other agribusinesses
Mineral’s plant rover at work in a soybean field. (Photo: Alphabet)
Google parent company Alphabet will shutter Mineral, its agricultural technology operation, but said the team’s work will live on with other agribusinesses.
Mineral was launched as a separate business in January, 2023, after five years of development by X, Alphabet’s innovation lab. With a mission to help scale sustainable agriculture., the Mineral team’s work included robots designed to inspect every plant in a field and the application of artificial intelligence (AI) and generative AI to help companies tackle the challenges of food waste, forecasting and the collection of high-quality data.
In a blog post announcing the closure, Mineral CEO Elliott Grant wrote, “Our mission has always been to make a meaningful, positive difference to the global food system – which we knew was an audacious, high-risk undertaking – so we would regularly ask ourselves: ‘is this the best way to maximize impact?’ and ‘are we reaching the diversity of farmers we want to, worldwide?’ We’ve noted before that we’ve been looking for new and innovative partnerships that can transcend traditional approaches. As a result, earlier this year, we decided to rethink Mineral’s structure and our approach to achieving a positive impact for the greatest number of farmers.”
Mineral’s work focused on three areas: developing sensing technology that can generate rich data sets about plants, organizing agriculture data from disparate sources for machine learning and building software algorithms and conducting research to advance the understanding of plants.
In its research, Mineral said it found that most companies were not collecting the quantity, diversity or quality of data needed to take full advantage of machine learning, which led it to build tools to capture, curate, clean and augment multimodal data. The company developed plant rovers to capture large quantities of high quality images, and then expanded to the construction of generalized perception technology that can work across platforms such as robots, third party farm equipment, drones, sentinel devices and mobile phones.
Mineral said an important discovery from its work was that implementing machine learning as a service was quite unlike software as a service (SaaS). It said models must be monitored and continuously retrained with new data – especially in an agricultural system that changes annually.
Early on, Mineral partnered with fresh berry producer Driscoll’s to develop AI tools to improve crop phenotyping, better forecast yields, optimize quality inspections and reduce food waste. Alphabet said it has now transferred some of that technology to Driscoll’s.
California-based Driscoll’s is the first agribusiness to receive Mineral technology. Perhaps best known as a strawberry producer, the company first engaged with Mineral team to to improve data collection in its breeding operations. Crop breeding involves the collection of multiple datapoints about each plant variety throughout the growing season. The ability to collect data at scale throughout the season allowed plant breeders to focus on other tasks.
Insights gained from that data inspired Driscoll’s to work with Mineral on another challenge: predicting the yield of each plant. In a case study, Mineral reported that its engineers and data scientists helped Driscoll’s improve its forecast accuracy and increase its frequency, reduce forecast volatility and extend its forecasting horizon.
