Technology trends for Food/AgTech leaders
“Software is Feeding the World” is a weekly newsletter for Food/AgTech leaders about technology trends.
Programming note: The newsletter will take a content break next week. You will still get an email, and it will include a reader survey for feedback. I would love to get your feedback about the newsletter, and any changes you would like to see in the future. I will call out some fantastic content creators.
My podcast on writing with Janette Barnard is out. The link to the podcast is towards the end of today’s newsletter.
This week’s edition includes the following topics:
Analysis: See & spray technology makes progress but significant challenges remain. A look at Greeneye, Deere, and Tartan Sense.
Technology trends: On-farm quality evaluation tools and their impact, insurance for small and smallholder farmers, and are grain marketplaces dead? (What about carbon marketplaces?)
Research review: A democratic alternative to true cost pricing of negative externalities?
Tidbits from the world of Ag/Foodtech: Are SPACs running out of steam?
Read, Listen, & Watch: Deere podcast with Matt Coutts, my Foster writing podcast with Janette, Barnard, video of 15 interesting small robots.
See & spray
💡Key takeaway: See and spray technology is making progress, but there are many unanswered questions on the application of this technology in different contexts (crops, regions, types of threats, timing etc.). The scalability of the technology and business models around it are very much in flux.
Israel based Greeneye Technology announced the first stage of a global rollout of its AI based precision spraying technology, with a launch in America. The initial phase will include corn and soybean farmers in the Midwest.
According to Greeneye, they have developed a precision spraying system with the following results obtained during field trials in 2020.
Detection: Recall rate (true positive) of 95.7% in multiple field conditions from all spraying applications (Green on Green and Green on Brown)
Savings: Average herbicides reduction of 78% from all spraying applications (Green on Green and Green on Brown)
The system is supposed to integrate seamlessly with any brand or size of commercial sprayer, indicating an aftermarket unit, which can be attached to a sprayer. The system is supposed to work at the same speed as broadcast spraying (20 km / hour), and so might not require any disruption or slowdown in normal farming operations. Cameras are mounted on a sprayer, capture images at 40 frames per second, which enables rapid detection and classification of weeds, including weed species. The system is able to calculate the amount of herbicide required, and sprays it on the weeds, and not around it.
Based on the diagram above, the system works across multiple crops (though the news report lists only corn and soy), works on the edge (real time operation with no connectivity required), is ISOBUS compatible (the unit can plug into the normal telematics of an equipment which supports ISOBUS), can do both pre and post-emergence, and can integrate with any sprayer.
Greeneye includes a dual spraying function to apply residual herbicides (broadcast), and apply non-residual herbicides precisely on weeds. The current practice of mixing residual and non-residual herbicide, and spraying it through the same nozzle wastes a significant amount of non-residual herbicide. The challenge is overcome by the dual spraying system.
A comparison between Green Eye’s technology, and John Deere’s See & Spray Select (announced in March 2021) based on ExactApply™ foundation, shows differences between the published material for the two systems.
Comparison between JD See & Spray Select & Greeneye system
There are many challenges to be solved around application of the see & spray technology at production scale, with a viable business model around it. These challenges are around technology - can you identify the threats, in real time and on the edge, with a variety of environmental conditions like soil patterns, rain, different crop types, lighting conditions, precision and recall rates, and spray the right amount and type of chemical at the right time etc. As Deere’s Blue River acquisition has shown, these are hard problems to solve, and will require novel approaches.
There are challenges on the business model in terms of pricing, deployment of technology, support and service needs etc. John Deere is taking a cautious approach, and is looking to integrate it within an existing piece of equipment, and to utilize its own distribution network.
In edition 74 (The Monkey and the Pedestal) I had written about companies following the “Rumba” strategy of building micro-robots for different use cases like reduction of chemicals, labor cost savings, and reduction in soil compaction. Companies like TartanSense in India are building electric powered robots for the smallholder market, which promise 50% reduction in chemical use, while lowering the skill set requirements for human labor.
There is a need to have working see & spray solutions, and it is going to take some iterations to get to it, given all the technology, and business model challenges.
Quality measurements on the go
💡Key takeaway: Product quality measurement technology reduces friction, and can potentially open up new sources of supply, especially in the developing world.
One of the challenges here in the US during CoViD has been the lack of inexpensive rapid at-home CoViD tests, even though they have been available in other places. The Panbio test from Abbott can be administered at home, returns results in 15 minutes, and is fairly accurate.
The grain and the produce industry faces a similar challenge of assessing the quality of harvested product quickly, and inexpensively. On the grain side, a quick assessment on the quality of their product, allows farmers to make decisions about storage, and transportation. On the produce side, it allows the producer to make decisions on pricing, product disposal, and evaluating alternative markets, if available.
Australian company Graincorp has invested in Hone, which brings the process of assessing the quality of a crop sample down to minutes from days. Hone includes a handheld device (see picture below), with a potential to use it for a variety of applications. The device weighs 700 grams (about 1.5 lbs), is IP65 compliant (outdoor use), with a 5 hour battery life, and it pairs with an iOS/Android device.
Image from Hone website
The spectral engine is powered by two sensors covering a wide spectral range spanning visible and near-infrared (NIR) regions and includes a large measurement area for consistency with heterogenous samples.
The company has the desire to extend the potential of this technology to soil carbon testing and measurement of carbon sequestration. A rapid, inexpensive, accurate, and on-farm soil testing kit could really change the economics on carbon markets, and open it up for more innovation.
Rapid and accurate quality testing in the developing world can open up newer supply locations for the global markets, as producers can provide products with consistent quality. A rapid feedback mechanism can help them improve the quality of their final products as well.
Everyone deserves insurance
💡Key takeaway: Policies and technologies to can cover more and more farmers with insurance, can improve livelihoods, increase productivity, and have better economic, social, and environmental outcomes.
The USDA is rolling out new, updated insurance policies through the Whole Farm Revenue Protection to US farmers who report less than $ 100,000 revenue per year.
Whole-Farm Revenue Protection (WFRP) provides a risk management safety net for all commodities on the farm under one insurance policy and is available in all counties nationwide. This insurance plan is tailored for any farm with up to $8.5 million in insured revenue, including farms with specialty or organic commodities (both crops and livestock), or those marketing to local, regional, farm-identity preserved, specialty, or direct markets.
The RMA’s research shows 85% of small producers report less than $75,000 in gross sales, so they set the revenue limit with the goal of including as wide a portion of small farmers as possible. The goal of the program is to make it more widely available, while reducing the paperwork requirements for insurance.
Pula and smallholders
Smallholder farmers have limited access to agriculture insurance, even though they account for 70% of production in Africa. The farmers are much more at the mercy of weather, input and market prices, labor availability compared to the “smaller” farmers in the developed world. It creates a vicious cycle of subsistence, poverty, and low productivity.
For truly smallholder farmers in Africa, Nairobi based Insurance tech company Pula is trying to fill the gap between the farmers and insurance companies.
Pula bundles insurance with input and credit services from aggregators, government subsidies to make insurance affordable for farmers, and less risky and profitable for insurance providers.
Pula has so far managed to sign up more than 4.7 million farmers in 17 countries on two continents. At least 154,000 farmers have made claims on the schemes to protect their income from crop failures. Pula says $6.9 million has been paid out to smallholder farmers this year alone.
Pula offers Area Yield Index Insurance, which makes it much cheaper to insure and administer policies, and does not require a huge amount of farmer level data.
The country is divided into agro-ecological zones (AEZs) based on historical rainfall, temperatures, prior yields etc, and average historical yield data is determined for each zone based on past data. At the end of the season, trained enumerators measure yield levels for each agro-ecological zone. With this information, farmers will receive compensation if yields in a specific agro-ecological zone are below a determined trigger level.”
The average annual insurance premium a farmer can afford is about $5 to $10. This includes the cost of product development, pricing, underwriting, claim adjustment and of course the claim costs. We use artificial intelligence, mobile-based registration systems—both USSD and mobile app—remote sensing, and end-to-end automation tools to make the numbers work.
The grain marketplace is dead, but what about the carbon marketplace?
💡Key takeaway: Grain marketplaces in commodity row crops in the developed world have had challenges to gain traction, in the true sense of a marketplace. There is significant room to improve the user experience. Carbon marketplaces have some suitable conditions for a marketplace, but they still have many challenges to gain traction.
Last week Bushel acquired Grainbridge. Grainbridge was the joint venture between ADM and Cargill to build out tools for giant marketing. Bushel is an ag-tech platform out of Fargo, North Dakota.
The acquisition complements Bushel’s strategic focus to enable, not disrupt, the agriculture industry and lead it into the digital age with strategies to solve real pain points and provide value for all stakeholders in the ag and food value chain. By providing permission-based data sharing and visualization, there will be improved data-based decision-making. These connections can support all points of the supply chain from grower to grain origination to consumer packaged goods (CPG companies).
Mark Johnson, CEO of Grainbridge wrote an article “The Day the Grain Market Died.”. Mark Johnson simplified his logic to two things
There are a limited number of buyers due to transportation costs of physical goods (this is why the supply chain infrastructure of the ABCD majors is so strategic and critical)
There are a limited number of farmers in a region, and the grain buyer knows all of them.
It is debatable if the same conditions apply to grain and produce markets in the developing world, due to limited infrastructure, both physical and digital. Due to this, we have seen a plethora of marketplaces open up in India and China, with Pinduoduo expanding the number of buyers for produce by reducing the role of the middleman.
Coming back to grain markets in North America, the inimitable Shane Thomas had made similar points about grain marketing, in the context of Indigo.
Online marketplaces have specific traits that allow them to thrive. One of those tendencies is fragmentation of the market itself. Disparate buyers and disparate sellers who struggle to find one another, when needed. A marketplace allows those two parties to connect seamlessly. The majority of farmers know their grain buyers, there typically are only a few near them.
There are other tendencies: decreased transaction costs from the marketplace, better experiences for both parties, information asymmetry in the market, high frequency of purchases, network effects and ideally an expandable market. When we look at whether a marketplace in grain has those tendencies listed, the conclusion tends to be no, or “kinda”, which shouldn’t in-still confidence that a marketplace is viable.
So what does the acquisition give Bushel?
Increased grower profitability at the field level through more data-driven decision-making. Bushel’s data insights capabilities will help educate grain buyers and sellers about forecasted trends and best practices related to grain marketing.
Increased visibility for food processors sourcing for specific traits such as higher protein or oil content within their customer base.
Verification to support sustainability credits and corporate targets for CPG companies through a “grain passport.”
If we apply the same logic to carbon marketplaces, are the conditions different for carbon marketplaces?
There is no physical transportation of carbon, as it has to stay in the ground.
The number of buyers is potentially large as different companies (and even individuals in the future) try to offset their carbon footprint (or inset in many cases).
The number of suppliers in a given region is still limited (at least here in the US).
The carbon credit could be securitized and then broken up or bundled up in different tranches to provide more liquidity to the market.
So do carbon marketplaces have a better chance of succeeding?
There are still some challenges. The transaction costs are still quite high (measurement and verification costs), and in the current scheme, the frequency of purchases is quite low (once or twice a year). Given the potentially large number of buyers, there will be network effects and the market can expand.
There are still significant challenges to drive down transaction costs, ramp up network effects, and create liquidity for carbon marketplaces.
A democratic alternative to true cost pricing
Reference: Patel, R. A democratic alternative to true cost pricing. Nat Food 2, 632–634 (2021). https://doi.org/10.1038/s43016-021-00347-0
💡Key takeaway: The dollarization of food systems’ externalities carries economic and political risks. Local democratic experiments navigate these risks by embracing both the complexity of valuation and the processes necessary for systemic political change. In spite of this, true cost pricing programs seem like a mirage, which will not deliver the stated results of this program.
Food system generates externalities. It includes on farm externalities like nutrient and water run off, erosion, GHG emissions. A report published by KPGM in 2021, concluded that for every $1 of revenue generated by the food industry in 2010, the industry generated $2.24 of environmental harm.
Research by Raj Patel evaluates approaches to quantify the negative externalities in dollar terms, and present a ‘true cost price’ which might promote more sustainable production and consumption practices. There are two views about true cost pricing.
One follows the Peter Drucker slogan “you cannot manage what you cannot measure.”
The other view says “not everything that counts can be counted.” There are a different set of parameters to be considered to count the dollar cost of externalities. For example, the Los Angeles school district decided to use a weighted average score of different parameters, without doing a true cost piercing analysis. For example,
Humane treatment of animals; and
Health and nutrition
The Good Food Purchasing Program now works with 2.5 million students, across 50 different jurisdictions. As you can see below, it does not do a true cost account, but it assigns some weights to different factors to can measure externalities.
Fig. 1 A sample Good Food Purchasing score card. This score card contains the ratings of a sample food provider along dimensions of local economy, environmental sustainability, fairness to workers, humane treatment of animals and healthiness of food served. Credit: Center for Good Food purchasing
True cost accounting punts the difficult task of democratizing the food system to someone else, when the ultimate goal of true cost accounting is precisely to inaugurate a transformation towards redistributive justice within the food system.
Diopsis? New show on Netflix? No, it's a scanner. Syngenta’s Diopsis insect scanner to give growers insight into the balance between pests and beneficial insects in their crop.
SPAC in a spat? AeroFarms and Spring Valley Acquisition Corp. mutually Agree to Terminate Business Combination Agreement.
QR codes: Don’t just show me the menu, can you trace my food as well? A Stockholm startup uses QR codes to track item movement in the supply chain. The software can generate insights for brands to manage risk, compliance, and claims.
You want a popsicle? How about a POP-ST? Inflexion and Almaco have announced a new tech called POP-ST to enable a broader spectrum of biological and niche treatments to be directly applied over-the-top of already-treated seed. POP-ST enables farmers to maintain biological seed treatments in optimal state until they are deployed in real time. It is added as an upgrade to existing planters, and creates an entirely new category of in-field, real time prescriptive seed treatments.
The Fertilizer Institute and its retail members announced a campaign to increase the use of 4R nutrient stewardship to 70 million acres by 2030.
Tillable and Compeer Financial digitize the process of farm credit. Tillage acts as a front end to pass data to the Compeer system, and Tillable does not keep any farmer data.
Brazil’s farmer cooperative, Coopcerrado won the UN Equator Prize due to its work to develop farmer-to-farmer mutual support. Farmers and local communities have developed a model of knowledge & resource sharing, while restoring the biome and providing a livelihood to thousands of vulnerable families.
Read, listen, and watch
📚 Book: “Lentil Underground: Renegade Farmers and the Future of Food in America” by Liz Carlisle. Liz tells the story of “audacity rich but capital poor Montana farmers who thought lentils were the answer and stuck with them until proved right.”
🎧 “Matt Coutts - Deere: Centuries of Farming Innovation” Fantastic episode featuring Matt Coutts on the Business Breakdown podcast under the Colossus’ Invest Like the Best brand from Patrick O'Shaughnessy. This newsletter features in the Blocks section of the podcast!
🎧 “Writing about niche topics with Janette Barnard” My podcast with Janette Barnard about her writing process. Definitely worth a listen, as Janette breaks down her writing process. Janette writes the always thoughtful newsletter “Prime Future” about animal ag and it is definitely subscription worthy.
📺 15 impressive farming robots (part 2)
So, what do you think?
💗 If you like “Software is feeding the world”, please hit the “Like” button at the bottom or share with a friend.
🙏 If you don’t mind answering 3 questions anonymously (2 are optional), I would love to get your feedback.
My name is Rhishi Pethe. I lead the product management team at Project Mineral (focused on sustainable agriculture). The views expressed in this newsletter are my personal opinions.