Die VAÖ-Homepage berichtete bereits darüber, dass Google mit der Tochter Mineral in die Landwirtschaft eingestiegen ist. Nun hat Mike Wilson, der frühere IFAJ-Präsident und Chefredakteur des US-amerikanischen Fachmediums „Farm Futures“ , einen Bericht über dieses Vorhaben des IT-Giganten geschrieben, den wir hier im Originalwortlaut abdrucken. Er gibt detaillierten Einblick in eine bald mögliche Agrarzukunft.
Chefredakteur Mike Wilson
Chefredakteur Mike Wilson ist Executive Editor und Content Manager bei FarmFutures.com. Er wuchs auf einer Getreide- und Viehfarm in Ogle County, Illinois, auf und erwarb einen Bachelor-Abschluss in Agrarjournalismus an der University of Illinois. Er wurde zweimal von der American Agricultural Editors’ Association zum Writer of the Year ernannt und ist ehemaliger Präsident der Organisation. Er ist außerdem ehemaliger Präsident der IFAJ, der International Federation of Agricultural Journalists.
Hier der Text im Original:
Mineral’s goal: Building tech to fix big problems in agriculture
In January Google’s parent company launched its moonshot ag business to make global agriculture more sustainable. Here’s how it’s going so far.
In January, X, the ‘moonshot factory,’ graduated its ag innovation project, Mineral, as an Alphabet subsidiary aimed at boosting global agricultural sustainability.
What, exactly, is Mineral? The company is building innovative analysis tools around advanced machine learning and artificial intelligence. These tools are being licensed to industry partners and used on everything from 10,000-acre row crop operations to half-acre smallholder farms.
Early tools include Mineral’s solar-powered rovers that can analyze a range of crops like corn and soybeans, melons, berries, lettuce, oilseeds, oats, and barley—from sprout to harvest. The machines, laden with sensors, cameras, and lights, autonomously drive through these crops, take high quality pictures, and gather a mountain of data that is turned into robust crop models.
In another example, Mineral’s quality assurance software can be used by food processing companies to determine miniscule defects in individual pieces of fruit. Using a mobile phone, users can take photos or videos of berries and sensors automatically and instantly detect defects. Mineral is working on other ways to use mobile and perception software to identify weeds, diseases, and pests in multiple crops.
So how has it been working so far? We asked Chief Commercial Officer Erica Bliss to give us a little more insight on how Mineral is innovating in agriculture.
Alphabet ‘graduated’ Mineral to commercial status in January. How has the ag business world responded?
Bliss: We’re getting excitement from the industry, and that’s been great. We have been developing tools with a number of partners since 2017 and now we’re really scaling to make these tools available to the market. It’s a sign of how hungry the industry is for people to demystify ML and find ways for it to impact growers.
For example, we’ve taken the imagery and data collected on our rovers and deployed that capability to drone-based weed scouting. We now see the model accuracy continue to improve across different geographies and crop types, not just in North America but South America and Europe. It’s been a full court press to continue refining products like these with partners that get this technology into the hands of farmers.
Mineral’s go-to-market strategy is through partners, not to farmers directly. How do these agreements work?
Bliss: We’re still at a point where we customize models for each customer. So today, it’s not just off the shelf, here’s a licensing agreement, go out and deploy it. Companies have unique needs that require additional tweaking, and every use case or setting in agriculture is unique and complex in its own ways.
We have solutions that enhance the ability to ‘sense’ on farm as much as possible, such as identifying weed species or plant growth stages at high speeds, or counting soybeans in a pod to estimate yields.
Additionally, when a partner comes to us and they have a vision to build something, but they don’t quite know how to do it, we evaluate if working together to customize and build the thing together can help them achieve their goals. For example, we partnered with Driscoll’s to develop the mobile phone app that can detect disease and bruises on fruit because it was a major pain point for them. These are commercial deals and we’re still developing and refining these engagement models.
What sets Mineral apart from other ag tech companies?
Bliss: Building models like the ones I just described are expensive and hard; there are other folks out there doing it as well. But we think of ourselves as a neutral third party, making those models available to everyone in the market, not just one company. So that’s a different lens than a company building a closed system. There are so many different niches on farms; we want to be able to enhance the ecosystem with incredible capabilities for the whole industry, so many people can build and co-create.
We want to make sure precision capabilities are accessible on all kinds of equipment. I believe this can reduce costs and increase efficiencies, and make all applications work better.
Can you give an example of a current partnership?
Bliss: We’ve been working with Syngenta for many years, helping to develop weed models. We’ve done work in the specialty space with Driscoll’s, starting with yield prediction and moving into quality assurance. That’s an extension of the computer vision capability we developed. We’re having conversations across the value chain as to how we can develop tools like these to transform things like quality assurance across many different crops.
What are some tools Mineral is working on that may soon work inside existing technology or products?
Bliss: We’ve continued product development, such as weed detection and classification. There’s been a lot of R&D to train the models and make them robust enough for different geographies and crops. We are having conversations with equipment companies using cameras inside equipment to optimize efficiency and monitor grain quality; Or in real time, to understand nuances of a harvester or any piece of equipment. Most of the work we’ve done is cracking the technology problems first; we’re more interested in making sure our partners have incredible tools and technical expertise from our team to work with.
What do you do for Mineral, and what got you interested in ag tech?
Bliss: I drive go-to-market for Mineral and identify opportunities for partnership. We ask, how do we solve our partners’ big problems, or help them achieve their visions? How do we pair our superpowers to develop solutions that scale? I’ve always been driven by a curiosity in how to have a positive impact in a global industry like ag. I started out working in dairy cooperatives but wondered, what if I could be the one who could drive positive change from the other side? I spent time working with companies like Coca-Cola, building supply chains, and Walmart in their direct farmer program, going downstream to understand how to improve the value chain. From those lenses I could see a lot of challenges in handling complicated food chains that multinationals work with.
On the other side of the chain, I worked with Syngenta for several years, bringing new products to market, including partners in lending and insurance. That brought me to Mineral. I feel lucky to work with a neutral party with empathy for different parts of the value chain and how to raise all boats in technology that can be applied to agriculture. At Mineral, we want to dig into deep challenges – to think big and solve big problems – and be audacious but practical. Try things, and if it doesn’t work, it’s time well spent if it helps point the partner and us in a more productive direction down the road.
Fotos: Mike Wilson, Mineral