49. What would Drake say?
Soil health, driverless operations, incumbent investments, and new tech
Hi, if you are new here, I am Rhishi Pethe and you have elected to become a member of the “Software is feeding the world” community. You will receive this free weekly newsletter at the intersection of technology and agriculture/food systems. I work as a product manager on Project Mineral at Alphabet X, focused on sustainable agriculture. The views expressed in this newsletter are my personal opinions.
This week’s newsletter is the “Canopy” version, which covers a few stories of interest from the world of agtech, research, sustainability, food & agriculture value chains etc. This week’s edition will cover the following five topics.
What would Drake say?
Path towards driverless operations
Monarch tractor partnerships
FMC Ventures invests in ScanIt
New tech for crop farming
Programming note: Please watch out for an interesting conversation with a human rights activist, journalist, and prolific writer next week!
What would Drake say?
Climate change, soil health, insurance, irrigation, and carbon markets!
😅 tl;dr (too long, didn’t read): Different studies continue to show the link between soil health, and resilience to climate change, better and less volatile yields, reduced insurance needs, and higher profitability. Despite all of this, carbon markets continue to dominate the discourse.
Improved soil health has an impact on many agronomic and financial metrics, that farmers and others case about. Soil organic matter is an important indicator of soil health.
A recent study done by Dr. Bradford, Daniel Kane et al, shows that soil organic matter not only protects US corn yields, but also lowers insurance payouts under drought.
Under severe drought, an increase of 1% soil organic matter (SOM) was associated with a yield increase of 2.2 ± 0.33 Mg ha-1 (32.7 bu ac-1) 22 and a 36 ± 4.76% reduction in the mean proportion of liabilities paid.
For example, the total crop insurance indemnity payments in the US in 2019 were at $ 8.1 billion.
Soil organic matter had a standardized marginal effect of 0.83 with a standard error of 0.04, meaning an increase of 1% soil organic matter was associated with an increase in yields of 0.83 ± 0.04 Mg ha-1.” (roughly 13-14 bu / acre for corn: 1 Mg / hectare = 15.9 bu/acre of 56 lb bushel corn)
The study shows that areas with higher mean SOM are correlated with lower corn yield loss due to drought. This relationship is the strongest for high intensity drought conditions and leads to lower crop insurance payments. This has direct policy implications to incentivize a transition away from pure year to year protection, based on APH (production history).
The results support the goals of such initiatives, highlighting how management of soil health might be a key adaptation strategy for sustaining agricultural production, rural communities, and food-supply security in a world experiencing increasingly extreme weather.
Another research study from MSU in “Nature Communications” (March 2021), shows that midwest farmers can use soil health management as a tool to combat climate change. By 2050, the midwest will need 35% more water to maintain current corn and soybean yields. The researchers found that the midwest is warming. The daily minimum air temperatures have increased, but the maximum daytime temperature has decreased based on a 120 year weather record.
As the increase in average temperature is coming from higher minimum temperature, the air is becoming more humid, making droughts more likely.
But instead of installing extensive and expensive irrigation systems that might only pay off under extreme droughts, Bassos advises farmers to invest in technology and regenerative soil practices that make plants more resilient and adaptable to climate change.
Another recent study from Soil Health Institute and EDF, makes a similar point that cover crops, and tillage practices help improve soil health and lead to profitability.
Figures from “Conservation’s Impact on the Farm Bottom Line” by the Soil Health Partnership, EDF, and K-Coe Isom
The numbers show the challenges in increasing adoption, as recent adopters see an increase in costs (when going from conventional tillage) and reduction in revenues and net returns. There is a learning curve in the adoption process, before you start to realize the benefits of the change in management practices. It is a policy, and tech question on how to get people over this hump on to the other side.
Daniel Aloysius Kane et al 2021 Environ. Res. Lett. in press https://doi.org/10.1088/1748-9326/abe492
Basso, B., Martinez-Feria, R.A., Rill, L. et al. Contrasting long-term temperature trends reveal minor changes in projected potential evapotranspiration in the US Midwest. Nat Commun 12, 1476 (2021). https://doi.org/10.1038/s41467-021-21763-7
💡Key takeaway: There are many benefits of improving soil health: profitability, resilience against climate change, avoidance of expensive irrigation equipment, policy implications etc. Recently, there has been a big focus on carbon markets, as a way to incentivize different value chain participants to sequester carbon. The excitement around carbon markets is understandable, and there is still a long way to go. Changes in management practices have the potential to have a significant impact in the future.
Path towards a driverless tomorrow
😅 tl;dr (too long, didn’t read): The improvements in vision and sensor tech, drop in prices for storage and compute, will accelerate the development and adoption of autonomous and electrified vehicles.
Today autonomous vehicles hold a less than 1% market share of total construction equipment sales worldwide, but global equipment sales are projected to grow 2.5% each year for the next couple of years.
Today’s ag vehicles are equipped with GPS, image sensors, telematics etc., but they cannot replicate a human. This will take time as tasks other than pure movement need to become automated. Autonomy for farm equipment is a continuum. Currently there are four types of autonomous agricultural equipment.
Driver assistance tools reduce complexity of field operations to improve efficiency.
With auto-steer, the operator controls tasks such as object detection and avoidance, safety, end-row turning, etc.
Vehicles with supervised autonomy with a vehicle to vehicle technology. Wireless connectivity between two vehicles can exchange data. The lead vehicle driven by a human, guides and controls the movement of other vehicles on the field.
Unmanned autonomy with a set of sensors to include path planning, obstacle management, and reaction to unforeseen issues.
The tech is past the second stage. Today’s ag robots are equipped with sensor packages, which let them “see” and “make sense” of their environment, and act accordingly. The robots will often use edge computing, with deep learning algorithms to perform certain specific tasks. Some of the current specific tasks today include monitoring growth, precise weeding, pollination, and harvesting.
The next set of advances in vision will transform tractor-attached equipment to actuate certain specific actions (e.g. see & spray) while operating in the field. As the cost of GPUs/TPUs/storage, networking etc. continues to drop, innovators will continue to push the envelope for different and novel use cases.
💡Key takeaway: The timing is uncertain, but what is more certain is that a future with increasing autonomy is coming.
CNH invests in Monarch
😅 tl;dr (too long, didn’t read): Incumbents continue to manage their innovation portfolio, through a combination of internal research, joint partnership, acquisitions, and venture investments.
For new tech, every large corporation (incumbent) can take multiple paths. It can invest to build and commercialize the tech in-house, partner with other companies, acquire the tech, or invest in startups. Several considerations come into play: time to market, investment needs, incumbent’s strengths and weaknesses, competitive pressure etc.
(I am a bit of a geek when it comes to autonomy and electrification. I had talked about Monarch tractors launching its electric tractors in edition 41 of the newsletter.)
One of the largest OEMs (CNH) has decided to make an investment and take a minority stake in Zimeno, the company behind Monarch Tractors. Zimeno has autonomous and data gathering capabilities. CNH wants a key piece of technology from Zimeno and Monarch. The tractor includes vision and proximity sensors for autonomous operations, and sensors for environmental variables like wind, and navigation.
Zimeno has also attracted a $ 1.5 million funding (2% stake, valuing the company at $ 75 million) funding from Indian tractor manufacturer VST Tillers Tractors Ltd.
Monarch tractors provide a high torque electric powertrain and can distribute its power electronically to axles, and hydraulics as appropriate. High torque is important to handle terrain of different kinds, and increase the reliability of your operation.
Monarch tractors provide a conventional operator’s platform and controls so that it can be operated like a normal tractor, though it has full autonomous capabilities, including a semi-auto mode when movements can be controlled with hand gestures (Jedi style)!
💡Key takeaway: OEMs will continue make acquisitions, joint partnerships, and investments in the autonomy and electrification space, as this tech is difficult to develop and scale.
😅 tl;dr (too long, didn’t read): Chem incumbents want to transition to selling outcomes, vs. selling products. This can be done only with a combination of digital tools, and physical products. Incumbents who do not invest in harnessing the power of data, digital and ML/AI will be left behind.
FMC Ventures (FMC’s venture capital arm), another incumbent, has invested in Scanit Technologies. ScanIt provides disease detection-as-as-service, with a combination of a hardware device for spore capture. The device is hooked to machine learning models for spore classification and identification.
Spore classification and identification model
Some of the key characteristics of the ScanIt solution are:
It can work indoors and outdoors, and key areas like post-harvest and cold storage.
It provides continuous monitoring. Continuous monitoring is key as timeliness of action is critical to prevent large losses.
It is a subscription based model. ScanIt stays the owner of the hardware device, but sells the data and model output as a service to its customers. The density of adoption will create additional opportunities at a regional level.
It takes 4 weeks to train a model for a particular context. ScanIt creates actual conditions, by placing sensors in growth chambers or bench-top bio-tents with inoculated host material. The challenge is to train the models to work in different real life conditions. The ability to create context aware models efficiently and at scale is a key component and challenge.
This investment pushes the envelope to expand its precision agriculture solutions (Arc Intelligence by FMC.) FMC’s precision agriculture solution provides monitoring, and forecasting for different types of pests. The service provided by ScanIt, beefs up their data collection, and modeling capabilities.
Real time predictive modeling, will help ensure that the right product is applied at the right time, in the right quantity. The understanding of outcome drivers, the ability to influence them with data and predictive models, and to tie a business model are the keys to transitioning to outcome-as-a-service model for input companies.
Image source: Arc™ farm intelligence by FMC
💡Key takeaway: Scaling of ML based models is tricky due to the need to have data to train your models, that work in different conditions. Incumbents will invest to accelerate and assist in their digital transformation.
New tech for crop farming
😅 tl;dr (too long, didn’t read): New tech around genetics, nutrition, and crop protection, monitoring, process control and autonomy / robotics will have a profound impact on agriculture in the coming decades.
Technology has played a role in production agriculture, with significant productivity increases over the last 100 years. Mechanization, scaling of the Haber-Bosch process, hybrid seeds, GMOs, data and analytics, better knowledge diffusion has led to a tripling of U.S. aggregate farm output in the last 70 years, without any corresponding increase in aggregate input.
Based on a recent report from the Center of Commercial Agriculture, Purdue University, the adoption of production agriculture is expected to accelerate in the coming years.
Crop farming changes
The use of precision agriculture and the management of production is moving towards managing smaller parts of a field. You see this trend in medicine (personalized medicine), health (quantified self), education (online education and self-paced learning) etc.
As we understand the environmental, biological, and economic factors that define the context, we will be able to use field, subfield and plant specific information to manage them at the field, sub-field, and plant level. This will have an impact on yields, profitability, and resource usage (soil, water, inputs etc.)
“The transition of production agriculture from an industry that grows crops to one that biologically manufactures raw materials with specific attributes and characteristics for food and industrial use products is well underway. “
1. Biotechnology, Nutritional and Crop Protection Technology
Progress will continue with growth manipulation (e.g. faster growth with less input), attribute development (e.g. grow sweeter berries) and deterioration process (e.g. longer shelf life.) The combination of nutritional (fertility as well as nutrition for end users)/biotech (e.g. CRISPR), crop protection (biologicals, different modes of action etc.) will have a big impact.
2. Monitoring, Measuring, and Information Technology
With the rise of sensors, perception at scale at various resolutions, the measurement of plant growth, and environmental variables, our understanding of the context and the growth of plants within that context will continue to improve. This includes technologies like GPS, remote sensing, weather systems, and other sensors.
3. Process Control Technology
The sophistication, fidelity and speed of monitoring and measuring systems opens the door for automated process control to stage the right interventions to close the gap between potential and actual performance. For example, the best corn yield is north of 600 bu/acre, but the average is 200 bu/acre.
Indoor farming utilizes tech to manage sunlight, humidity, temperature, and other characteristics of the plant growth environment. Irrigation systems are an example in field crops, with automated turn on and off tied to monitoring systems that measure soil moisture, weather information etc.
4. Autonomy and Robotics
Depending on the type and region of agriculture, labor availability, safety and wellbeing are huge challenges. Autonomy and robotics (with electrification) will address some of the existing problems.
Boehlje, M. and M. Langemeier. "Importance of New Technologies for Crop Farming." farmdoc daily (11):32, Department of Agricultural and Consumer Economics, University of Illinois at Urbana-Champaign, March 5, 2021.
💡Key takeaway: Multiple tech areas will have a profound impact on agriculture in the next few decades, with data, ML/AI, and precision agriculture being the underpinnings of the new tech.
Join me and Shane Thomas, Tim Hammerich, Janette Barnard, and Travis Martin every Thursday at 1 PM Pacific for the Clubhouse Ag Power Hour. Join the AgTech & Agriculture club as a member. The club hosts multiple events every week, with different focus areas (VC, founders, agtech, agribusiness etc.)
I have a few Clubhouse invites to share (iOS only). If you want to get an invite, reply to this email with your phone number. First come, first served!
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.