76. Banking carbon: Measurements & Moral Hazards (?)

Technology Trends for Food/AgTech leaders

“Software is Feeding the World” is a weekly newsletter about technology trends for Food/AgTech leaders. 

This week’s edition includes the following,

  • Analysis: A look at a Rabo’s Carbon Bank program to tease out larger trends. Is there a “moral hazard” situation brewing due to “additionality?”

  • Technology trends: Different flavors of autonomy, a huge gap between carbon credit adoption and emissions in agriculture, and a differentiated go-to-market strategy from Semios.

  • Research Review: Systemic risks of artificial intelligence on sustainability.

  • Tidbits from the world of Ag/Foodtech: Funding news, agtech, sustainability, news from developing countries.

  • Read, Listen, & Watch: Book: “Weapons of Math Destruction”, Podcast: The Slow Melt on the continuum of chocolate, Documentary:  


Banking carbon: Measurements and Moral Hazards (?)

Seana Day (Cultivian Ventures) wrote the following about carbon markets and soil carbon measurements.  

Whether the market will ultimately arrive at a ‘measure-plus-model,’ ‘model-only,’ or ‘measure-only’ approach, it is too early to tell.

I had confidently said that the market will settle on a measure + model approach, with examples from health care and banking. This week Rabo AgroFinance shattered my confidence with their approach.

Rabo AgroFinance announced their Rabo Carbon Bank program. The pilot program for 2020 included five farmers, getting paid $ 15 to $ 20 per ton of sequestered carbon. The company has partnered with Continuum Ag to identify management practices that lead to better soil health and carbon sequestration. Next year, the pilot will increase to 25,000 acres.

Rabo is not relying on models or remote sensing to measure the amount of carbon sequestered. They are doing actual in-field measurements to validate and Continuum recommends practices to increase the amount of C sequestered. Rabo is leaning into building trust through actual measurements, to verify carbon sequestration amounts. Rabo is not focusing on just corn and soy, but also on crops like broccoli, peanuts, potatoes, soybeans, sweet potatoes, tobacco, watermelons, and wheat.

Continuum Ag is a soil consulting company, with many years of experience related to soil health. The pilot includes fruit growers, cattle grazers, and specialty crop growers to evaluate which management practices would result in carbon sequestration. Continuum Ag uses historical yields, farming practice data, and soil sampling data to understand the impact of practices on soil health, including zone creation for sampling.

Rabo brings in demand for carbon credits, Continuum provides the software tools, and the agronomy expertise and measurement, and the farmers “supply” the carbon in their soil.

As is the case with agriculture carbon credits, Rabo’s program will pay for new practices due to the principle of additionality. Ironically, Mitch Hora, the owner of Continuum Ag will most probably not qualify for credits under the Rabo program!

On his own farm, Hora and his family have been no-tilling since 1978 and planting cover crops since 2013.

This is a tricky situation as farmers will be rewarded for change in their practices, but not for continuing good practices. It can create some interesting decision points for the farmer. How should a farmer decide whether they should enroll in a carbon program or not?

There are several factors to consider.

  • What are the projected costs and associated agronomic benefits of enrolling in the program?

  • Are there any incentives available to change agronomic practices (for example, incentives to buy cover crop seed, or make any equipment investments.)

  • What is the time to value to realize agronomic benefits?

  • How much time is required to get to a positive change in the quantity of carbon sequestered?

Given the complexity of the decision making on the agronomic side, Rabo’s approach of partnership with a group like Continuum Ag is commendable. Continuum has the expertise, the farmer trust, and the relationships to go with it to help through a complicated decision making process. Rabo will have to find entities like Continuum Ag, wherever they want to expand the program. Scaling a program based purely on measurement will be challenging due to the costs associated with measurement and verification. 

The price of carbon and your belief on how it will change in the future will play a role in the decision to participate in any carbon program.

Let us imagine two different scenarios of how the price of carbon will change in the future.

All else being equal, if you believe Scenario A is going to play out, you will delay your participation in the program. If you believe Scenario B will play out, you are more likely to opt into the program. Financial analysts, and people better than me in terms of modeling skills can run different simulations based on a set of modifiable assumptions. Any modeling will be dependent on the government subsidies available, which are different in different parts of the world.

Whichever way you model out different scenarios, it is probable that there are “moral hazard” scenarios associated with the adoption of these practices (if one assumes they lead to better outcomes).

No one ever said carbon markets in agriculture are easy.

Technology trends

Autonomy targets specific use cases with new and existing equipment

Drones target labor, input usage, and terrain requirements

Ground based machinery, aerial spraying using aircrafts, or manual spraying are the main solutions to protect food crops from pests and diseases. The farming terrain, local agronomic conditions for pest and disease, the crop type, the crop growth stage, are some important factors, when it comes to the choice of method to spray crop protection products.

In the high mountains of the Andes (the birthplace of potatoes), large tractors are not suitable for spraying on potatoes. Potato farmers have to spray as many as 14 times during the growth cycle to prevent severe yield loss.

Drones can provide a good alternative for this context. XAG has recently tested drone spraying capability through a partnership with Megadrone. The drones (4 atomized nozzles, 16L tank) fly 2 meters above the canopy, and deliver droplets directly to the base of the potato plants without drift or soil compaction. During test flights, it resulted in a 30% reduction in chemical use, and 90% water reduction.

Its 30% reduction in chemical use and 90% water saving allow farmers to effectively control pests and fungal diseases with minimum environmental impact. Most importantly, these drones can fly semi-autonomously reducing the need for skilled human labor.

Robots to assist human labor

Here in the US, farming robotics startup Burro closed a $ 10.9 million Series A, led by Toyota Ventures and S2G Ventures.

Burro’s platform features a unique patent pending approach called Pop Up Autonomy™ – a name reflecting how its autonomous people-scale robot works immediately out of the box by enabling every worker within an environment to become an operator. Burro’s robots do not require a centralized control or installation of complicated infrastructure – rather, they use computer vision and AI to learn “on the fly” to navigate autonomously “from A to B” while carrying various payloads.

Burro’s robots focus on assisting humans in tasks related to table grape farming. Given the huge labor costs involved in specialty crop farming, Burro is potentially going after a huge market. The challenge will be whether they can adopt their tech to work in different contexts of crop types, terrains, crop growth stage etc.

Autonomous pivots to reduce water and chemical usage

We talked about smart irrigation, micro-irrigation, and other forms of smart irrigation systems in the last edition of the newsletter.

There is a significant opportunity for smart irrigation systems with pivot irrigation, especially in places like Nebraska and Colorado in the US. Autonomous pivots, which incorporate cameras, and other sensors can understand growth stages, water needs, moisture levels, expectation precipitation, and calibrate the irrigation needs for their field quite precisely. It might be simpler to solve this problem, than say see and spray due to the slow movement of pivots. The presence of a long life pivot with a slow refresh cycle, requires solution providers to furnish aftermarket solutions, which can integrate with an existing pivot. The smart pivot connected to the sprinkler system to help eliminate weeds, detect diseases, and apply fertilizer or pesticides. 

Autonomous decision making & remote control to address the farm labor crisis

Agtonomy (Ag + Autonomy?) came out of stealth mode with a $ 4 million seed round from Grit Ventures and Village Global.

Agtonomy’s marketing tagline says, 

Building a sustainable future for local farms.

A hybrid autonomy and tele-assist platform for agriculture vehicles.

It is not exactly clear what Agtonomy does. Agtonomy might be providing an aftermarket sensor package to add additional capabilities to your existing agriculture equipment (based on job openings on their careers page). They work with original equipment manufacturers as a go-to-market channel (not clear who these OEMs are).

The company wants to provide a tele-assist service program, to allow for remote mode of operation, with built in flexibility and safety features. One can imagine an off-farm operations center, with operators operating equipment on various farms.

It is not clear how good their tech is, or whether they have established any existing relationships, but the model is an interesting one. I like the aftermarket sensor package, given the refresh cycles on ag equipment are unlike the refresh cycle on your smartphone. Using OEMs as a channel to go-to-market leverages the existing distribution and relationship network of the OEMs. The teleoperation reduces the burden to have full autonomy, and potentially accelerates their path to market.

Carbon credit adoption lags industry impact

Tthe decision to take part in a carbon program is complicated. If you look at carbon offsets as a whole, agriculture related emissions account for just over 1% of the carbon credit projects.

According to the FAO, agriculture accounts for a third of the total emissions. It shows,

Pre-farm gate agricultural food production — including usage of fertilizers and other soil inputs — is the biggest contributor within food systems; while methane from livestock farming and rice cultivation accounts for around 35% of food system emissions.

The Berkeley Carbon Trading Project has published data for carbon credits across industries, and agriculture is a tiny sliver of the total pie.

Typical projects for agriculture are related to soil carbon sequestration, impact of input runoffs, and livestock related emission mitigation.

If agriculture accounts for a third of the emissions, but only 1% of the total carbon credits, there is a lot of potential in agricultural carbon credits. In spite of the recent exuberance in the carbon markets related to agriculture, there are significant headwinds as we saw in the previous section. There is a lack of widely agreed upon, cost effective measurements, many of the models work in specific geographies only, and a lot needs to be done to validate those models. The process of physical soil sampling is expensive, time consuming, and oftentimes not very reliable.

All of this says there is a huge opportunity for startups, and incumbents (and everyone in between) to solve some of these difficult problems and bring those solutions to market. If the price of carbon continues to go up, there is potential ROI on the other side.

Semios leads with a different go-to-market strategy

Semios has been on a tear, when it comes to M&A activities and fundraising. Within the world of AgTech, Semios has been around for almost 11 years now. They initially started to help tree, vine, and nut growers manage pests and collect data about their crops. They deploy sensors to measure leaf wetness, soil moisture etc, and recommend trapping insects and dispensing pheromones.

Over the last 12 months, the company has made multiple acquisitions, including Australian company AgWorld. These acquisitions have expanded Semios’s geographical, and most importantly crop footprint.

  • Altrac: Frost and irrigation control

  • Centricity: Farm activity record-keeping

  • Agword: Whole season farm management and documentation

Semios is following a strategy similar to Telus. Telus has acquired multiple companies to address needs for farmers, across multiple crops, and geographies.

The interesting part is Semios’s go-to-market strategy. As we have discussed innumerable times in “Software is Feeding the World”, it is extremely difficult to go “direct-to-farmer” with digital solutions, especially in the row crop space. Most of the agronomic and farm management digital solutions are loss leaders, to help input companies establish digital relationships, create additional moments of collaboration in the customer journey, with the ultimate goal to sell them an input or a piece of equipment or to increase LTV (lifetime value) and lower CAC (customer acquisition costs).

Semios has chosen the path to go direct, and has built out a sales and service team to enable stronger customer relationships. It does slow down growth, and scaling is expensive, but the recent $ 100 million raise will provide the necessary funds for more acquisitions, partnerships, and growth.

Research review

Artificial intelligence, systemic risks, and sustainability

“Software is Feeding the World” frequently talks about the power of artificial intelligence and machine learning to better understand agriculture and food systems. The power of AI/ML combined with Big Data can lead to better decisions and economic, environmental, and social outcomes. AI/ML has a big role to play to make agriculture more sustainable, both economically as well as environmentally.

Latest research published in the journal “Technology in Society,” investigates artificial intelligence, associated system risks, and its impact on sustainability outcomes. The research combines findings from literature, with new data, and asks some new and interesting questions.

  • Where in the world, and into which sectors directly relevant for biosphere-based sustainability, is AI and associated technologies diffusing?

  • Which sustainability systemic risks could emerge as the result of this diffusion?

  • To what extent do current notions and principles related to “responsible AI” acknowledge systemic risks related to sustainability?

  • Which possible governance mechanisms could be developed to help mitigate these risks?

The research highlights “increased applications of AI and associated technologies could lead to more effective uses of land- and seascapes, augmented environmental monitoring capacities, and improved transparency in supply chains; it could also create new systemic sustainability risks as AI technologies diffuse into new social, economic, and ecological contexts.”

Based on Crunchbase data, even though the number of startups in AI for Agriculture is the highest by a large margin in the US, China is investing a significant amount of resources. There is a new cold war for investment and talent brewing between the US and China, as can be sheen from the chart below.

Fig. 1. Global distribution of AI technologies and investments in farming, forestry and the marine/aquaculture sectors. Fig. 1A. Geographical and sectoral distribution of companies that develop applications of IoT, sensors, robotics and AI-supported analytics for aquaculture, forestry and agriculture. Total number of companies N = 1114. Fig. 1B. Geographical distribution of investments in companies listed in 1A. See Supplementary Information for details about methods and data.

AI/ML has the potential to drive optimization and simplification at both local and global levels. This might undermine resilience, if it causes loss of local knowledge, diversity, and a sense of place.

The research article highlights the following risks created by diffusion of AI:

1. Systemic risks due to networked nature of systems

The recent cyberattacks on JBS, New Cooperative, and the hacks into John Deere and CNH systems highlights some of the systemic risks.

2. Algorithmic bias and allocative harms

If the AI systems are designed poorly or with limited or biased data sets, then the train data could bias the emergent AI system. Modeling can transfer context bias, if AI systems are designed for one agro-socio-ecological context, and are transferred to another context, without checking on transfer learning or the feasibility of transfer learning a model.

3. Unequal access, benefits, and impacts

Resource constraints and access to information, introduced some additional risks as AI technologies start to defuse. Widespread use of AI in farming to a specific context may result in capital concentration and higher levels of inequality. If AI models create massive information asymmetries, then the AI model could help extract asymmetric rents, push local farmers to unfavorable contracts, or price discriminate.

4. Shocks, cascading failures and attacks

The complex interactions between humans, and machines, and machines and machines with the ecosystem, can lead to cyberattacks on farming, and food systems. Digital twin farms based on data from sensors, can be analyzed with AI to provide meaningful insights.

Industrial agricultural landscapes around the world now generate high yields of a few crop species, but have led to declines in many other ecosystem services also valued by societies, including biodiversity, scenic beauty, and climate or flood regulation.

The systemic risks associated with improper implementation of AI/ML are real, and they are not something to watch out for in agriculture only. (See book recommendation)

Reference: Victor Galaz, Miguel A. Centeno, Peter W. Callahan, Amar Causevic, Thayer Patterson, Irina Brass, Seth Baum, Darryl Farber, Joern Fischer, David Garcia, Timon McPhearson, Daniel Jimenez, Brian King, Paul Larcey, Karen Levy, Artificial intelligence, systemic risks, and sustainability, Technology in Society, Volume 67, 2021, 101741, ISSN 0160-791X, https://doi.org/10.1016/j.techsoc.2021.101741.


Funding, M&A

US farm data company Ag-Analytics acquired land research and evaluation platform AcreValue from Granular, a subsidiary of Corteva Agriscience. Did Acrevalue not fit within the Granular portfolio? Does Granular want to focus on agronomy to help its parent company?

Semios, a leading precision farming platform has raised $ 100 million in funding from a PE and VC firm. The funding will be used to accelerate its product roadmap, commercial growth, & potential acquisitions.

The trend of autonomous robotics companies raising venture funds continued with Burro raising $ 10.9 million. Burro works in table grapes, and their robots work as assistants to humans.


Weather provider for insurance measurements in India (Skymet) builds tools to help financial services mitigate risks arising due to climate change. The tool will provide information such as where the land is located, which crop was sown in the last four seasons, the soil condition of the farm, irrigation status of the farm, yield received every season etc.

The Wells Fargo Innovation Incubator (IN2) helps startups test technologies at scale, up to hundreds of acres on farms within the FBN network. FBN will leverage its network and detailed agronomic and environmental datasets to curate ideal farms for trials to test each startup’s solution in the right agronomic conditions.


There is a battle for precious landraces in Turkey. The Turkish government is battling British archaeologists and seized the specimens of the seeds of traditional crops from the British Institute in Ankara, grown in isolation for centuries. The long defunct crops potentially contain invaluable genetics to create more nutritious, disease and drought resistant grains.

Trees for Jane (from Jane Goodal) is partnering with the Trillion Tree Campaign, an initiative led by the German NGO Plant-for-the-Planet, and 1t.org, a World Economic Forum project supported by Salesforce CEO Marc Benioff, to reach the trillion trees goal by 2030. While tree-planting on such a massive scale could sequester vast amounts of carbon, there are concerns about the potential social and environmental impacts, especially if trees are planted at the expense of native ecosystems or if land targeted for planting is seized from local communities.

The share of agricultural land owned by non-operating landowners and leased to farms and ranchers varies by state and region. In the US grain belt, the percentage of rented land varies from 40% to over 50%. Land ownership is a big factor in the decision to change practices.

The agriculture sector was responsible for 15% of Australia’s emissions in 2019.

Developing countries

Africa’s Great Green Wall initiative, which aims to restore arid and semi-arid lands – has 393 million hectares of land with restoration potential. An FAO analysis found that 17 million hectares of land have been turned into new cropland since 2000, representing a 5% increase.

Investments in data and innovation ecosystems to help transform Indonesian agriculture. It contributes 14% of Indonesia’s GDP, employs a third of the country’s labor force and is dominated by smallholder farmers (93%).

Read, listen, and watch

📚 “Weapons of Math Destruction” by Cathy O’Neil. This book is relevant to the research review earlier in the newsletter, about the impacts of AI/ML.

Weapons of math destruction, which O’Neil refers to throughout the book as WMDs, are mathematical models or algorithms that claim to quantify important traits: teacher quality, recidivism risk, creditworthiness but have harmful outcomes and often reinforce inequality, keeping the poor poor and the rich rich. They have three things in common: opacity, scale, and damage.

🎧 The Slow Melt: The first podcast to cover the continuum of chocolate by Simran Sethi. If you want to geek out on chocolate, this is a good series to get into.

From flavor and physiology to chemistry and conservation, from global markets and gender to climate change, social justice and beyond, The Slow Melt highlights the people, places and processes behind this $100 billion industry.

📺 “An Innovative Farming Model for the Next Generation” TedX talk by Clara Coleman.

Clara is the daughter of renowned farming pioneer Eliot Coleman. She is a second-generation organic farmer, consultant, writer, and speaker on sustainable four-season farming.

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.

About me

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.