Alphabet brings agtech startup out of stealth with data from 10% of world’s farmland, 3 major customers

This month Alphabet Inc. , the parent company of 谷歌 , launched Mineral, an agtech business it’s been incubating in stealth within its X, the moonshot factory since 2017. Louisa Burwood-Taylor caught up with Mineral CEO Elliot Grant...

Mineral aims to provide foundational and actionable data and analytics for companies across food, agriculture, and technology to better understand the natural systems driving our food supply and find new and improved ways to manage the industry in the face of climate change.

And it solves for what Elliott Grant , CEO of Mineral, says has been a key issue holding back digital agriculture advancements for years: low levels of high-quality, diverse, and scalable agricultural datasets.

“It is still early days for AI/ML and sensing in agriculture. I love how much innovation is happening — and we need it desperately — but this industry is still at day zero for maturation, so we saw the opportunity to create foundational technologies to help other companies go faster,” Elliot Grant, CEO of Mineral, tells?AFN.?“Not enough data have been collected and that’s an industry-wide problem.”

Mineral is gathering data at scale. It is combining proprietary sensing technology — a?camera-laden rover?capable of collecting large imagery datasets — with “large, multimodal, unstructured sets of the world’s agricultural data, sourced from satellite images, farm equipment, and public databases.”

It is also conducting plant research with the goal of being able to provide clients with analysis and insights on the holy grail of ag insights: the relationship between crop genetics, environmental impacts, and management practices on the farm, according to Grant.

Vast Data Needs

Mineral has already surveyed and analyzed 10% of the world’s farmland and it says it will increase the number of data points pulled and analyzed from any one farm by more than 20 times by 2050, from an average of 190,000 data points per day in 2014.

With these data, Mineral has developed 80″high-performance” machine-learning models to help a range of businesses, farmers and researchers and breeders to “predict crop yields, increase production, target pests and weeds, reduce waste, minimize chemical and water use, and reduce the impact of agriculture on the planet,” according to a company factsheet.

Extensive and reliable datasets are essential for artificial intelligence and machine learning (AI/ML) to work effectively. In other less variable industries — and those with shorter cycles to contend with than agriculture’s typical year or biannual growing seasons — collecting repeatable data is much easier. Agriculture changes from one day to the next as crops grow, weather changes, weeds and pests appears, and more!

“It’s a harsh environment to collect data from; combined with the astronomical complexity of the problems people are trying to solve across diverse domains that are changing every day, that’s a very hard AI problem to solve [particularly with a low level of data to start from]. I don’t think people appreciate how much diversity there is in ag.” He referenced that some of his team members who had joined from other industries such as advertising were generally surprised about how much harder it was than expected.

To boost its datasets, Mineral even created simulated images of plants “that are so realistic we can use them for training a model, even if [our rover] has never seen that plant in the real world,” Grant told?The Smithsonian?last year.

Data is key but there have also been considerable developments in artificial intelligence and computing power enabling Mineral to take full advantage of the deep levels of expertise, manpower, and digital infrastructure at parent company Alphabet, adds Grant.

Productizing Mineral

Mineral has worked closely with three core partner customers for years to help develop its platform.

1. Major US berry producer Driscoll's , “who have been consistently early adopters of technology, bringing AI (artificial intelligence) to their business processes including work with some of our robotic tech,” according to Grant.
2. Global ag inputs provider Syngenta.
3. CGIAR Climate Impact Platform , the global food security group with deep knowledge of what Grant describes as our “forgotten crops that are incredibly important for food security such as?beans.”

It has also started working with new customers, including, according to a company factsheet, “many of the agriculture and food industry’s most influential companies and research institutes” who are using the platform “to more deeply understand plant physiology, discover more resilient crops, increase production, and improve the bottom line for agribusinesses and farmers alike.”

Grant says that the company is still exploring the business model but referenced being a service for large companies as well as startup businesses looking to accelerate their innovations with the data, insights and infrastructure Mineral could provide, without having to do it all themselves from scratch.

What do you think? Is Mineral a game-changer for the agtech industry? Email Louisa Burwood-Taylor?

#agrobotics #artificialintellegence #AI #bigdata #machinelearning #remotesensing #agribusiness #farming #agriculture #innovation