Automating the farm

Everyone involved in farming understands just how complex […] a job running a farm really is. We believe that technology can make it easier to run a farm, and we aim to provide our growers with tools that help them achieve success.

Jesse Vollmar, CEO of FarmLogs, February 2019 message, “Tying it all together in 2019”, derived on 18.01.2021 from https://blog.farmlogs.com/where-farmlogs-is-headed-2019

The innovations in agriculture have been numerous and varied. The enhancement of plough design over thousands of years; the invention of barbed wire in American Midwest in the 1870s;¹ the introduction of petrol-driven tractors and later combine harvesters; the discovery of the benefits of antibiotics on livestock health, nitrogen on crop yields, and genetic modification on both crops and livestock – the advance of technology has clearly played a critical role in this field.

Today, however, despite these advances, farmers face significant challenges. Globalization has driven a focus on scale that makes it difficult for smaller operators to compete. It has also led to a homogenization of crop varieties and a lack of biodiversity that is exerting a negative impact on the environment and on human health. Increasingly extreme weather conditions caused by global warming create greater uncertainty in yields while boosting the prevalence of insects and disease. To solve these problems, fertilizers, pesticides, and antibiotics are overused in many countries. In addition, farmers must balance the challenges of running a commercial enterprise while dealing with a growing mass of regulatory requirements and government initiatives to improve sustainability and provide safe and healthy food.

We believe that technological innovations, and in particular intelligent automation systems, may provide at least part of the answer. Thanks to advances in technology, we now have more powerful computers, cheaper and more accurate sensors, a growing range of enabling technologies such as fast wireless networks, cloud storage and computing, and more software engineers in the world than ever before. We believe these elements in concert will lead to a golden era for innovation in robotics and automation and help solve many of the issues facing modern agriculture.

The world population is expected to grow to 9.7 billion people by 2050, a 25% increase from today.² That in itself already implies 2 billion additional mouths to feed, but expectations are that affluence will also rise materially in a number of countries with very large populations; as a result, the UN predicts that demand for food will grow by more than 50%.³

To supply this growing need, it is estimated that there are approximately 570 million farms operating worldwide today, which provide a livelihood for around 27% of the global work force (roughly 884 million people).⁴ Agriculture is often the critical economic activity in areas of extreme poverty and in some developing countries it can represent 25% of GDP.7 In more affluent countries, while commercial agriculture is critical since most of society rely on others to grow and process their food, it is a very small part of the economy, and globally it accounts for only 4% of global GDP.⁵ 

The environmental footprint of agriculture is far greater than its weight in global GDP. Half of the world’s habitable land is used for agriculture, far more than any other human activity. The irrigation of crops accounts for 70% of global water use, and agriculture directly contributes to approximately 11% of global greenhouse gas emissions, mostly from cattle.3 Agriculture is indirectly responsible for 20% of total greenhouse gases.⁶

Figure 1: Half of the world’s habitable land is used for agriculture

In 1962, farmers produced on average enough over their lifetime to feed 25.8 people. Today thanks to innovations in technology, machinery, crop science and genetics, a farmer feeds 155 people. However, Ernst & Young⁷ estimates that by 2050, a single farmer will have to feed more than 265 humans. A growing number of technology-based solutions and innovations are emerging to meet the associated challenges, and the amount of investment into “agtech” start-ups has jumped noticeably in recent years. In addition to these investments, established companies such as John Deere, Trimble, Hexagon, Topcon, and many others also continue to invest in research and development (R&D), devising innovative new technologies.

Before we look as some of the examples of innovation in agriculture, it is helpful to broadly classify the solutions into different categories. Many of the solutions focus on improving and managing yield: the day-to-day job of getting the best results from the raw inputs into the farm, be it seed or livestock. How to generate a large crop of beautiful produce, despite the variability of weather and the risk of pests and disease, while also limiting environmental impacts.

Other solutions focus on the farm as a business: how to efficiently manage the daily operations of cost management, staff scheduling, equipment maintenance, and seed procurement, and how best to sell in the marketplace. There are also tools that attempt to build a bridge from day-to-day operations to data from the end markets, so that farmers can adjust inputs dynamically to reflect changes in demand from consumers and supply from competitors. Of course, there are also solutions specific to different types of farms, from crop to livestock, dairy to meat, oils, seeds, nuts, fruits, as well as solutions suited to large-scale operations which already use a variety of advanced technologies, as well as those designed for much smaller farms whose only technology may be a mobile phone.

Figure 2: Investments in agricultural technology (“agtech”) start-up companies (in USD billion)

Due to the large number, breadth, and evolving range of new solutions, we have not attempted to cover everything here. Instead, we have selected a few key areas that we believe are most relevant to the theme of robotics and automation in agriculture.

One of the fundamental challenges long facing farming is that so many of the inputs are unpredictable. How much water and sunlight crops will receive today, tomorrow, or over the course of the season is unknown, and too much or too little can be disastrous for crop yields. Rain and frost can affect soil acidity and wash away valuable nutrients, and with global warming on the rise, extreme weather conditions are becoming more frequent.

Satellite imagery has been used in agriculture since the 1980s, allowing farmers to detect and monitor anomalies and seasonal changes. This information is applied to optimize irrigation and identify nutrient deficiencies and the spread of pests and disease long before they are noticeable to the human eye.

Today a growing number of companies, such as CropX, Deepfield Connect, Prospera, and SWIMM Systems, are enhancing these solutions with sensors or “probes” placed directly into the soil to provide accurate real-time information and alerts on any changes in soil conditions. Some offer even more complete solutions, using detailed weather forecast data, onsite weather stations, pressure and soil tension sensors, and dendrometers (to measure tree growth).

As sensors become cheaper, more accurate, and more reliable, and as analytics platforms reach a level of sophistication allowing them to transform the various data inputs into actionable insights, farmers should be better positioned to increase yields with less waste and a smaller environmental impact.

Farmobile, ProAgrica, CGIAR, AgDNA and many others offer data analysis platforms to collate the data from the sensors, perform analysis, and turn the signals into a useful and practical strategy. The EU’s Flourish Project seeks to bring drones, small robotic ground vehicles, satellite data, and other input data from an array of sensors together in a holistic solution.⁸

One of the more extreme solutions is to grow crops indoors or inside a greenhouse. This type of artificial environment allows the inputs to be controlled precisely and limits the damage, or “externalities”, to the environment. These systems can greatly reduce the volume of water, pesticides, and fertilizers used, help ensure optimal growing conditions and extended growing seasons, and even enable crops to be grown vertically on shelves.

Farminova claims that their factory farm system reduces water consumption by 85-95%, cuts fertilizer use by 60%, and entirely eliminates the need for pesticides. Furthermore, in a world where fertile farmland is increasingly scarce, as more land is assigned to residential and industrial operations, this type of factory farm can produce more than 10x the volume of crops for the same land area.⁹

Another example is the UK-based company HydroCotton, which uses hydropnonics to grow cotton indoors with 80% less water and no pesticides. When we imagine agriculture, most people think of food, but a large part of agriculture produces natural fibres for clothing as well as wood for construction and crops for feeding livestock.⁷

Urbanization has greatly depleted the size of rural communities and therefore the pool of skilled labor available for occasional and seasonal farm work. The evolution of the workforce continues today. In 1900, the rural population was 6.7x times larger than the urban population. Today this ratio has reversed, and now more of the world’s population live in urban settlements than in the countryside.¹⁰

As we see in other industries facing a growing shortage of skilled labor, robotics and automation solutions are filling the labor shortage gap in agriculture. Automated milking systems (AMS) on dairy farms have seen rapid adoption since their first commercial installation on a farm in the Netherlands in 1992, with 20%–30% of dairy farms in the Netherlands and Nordics using some form of AMS today.¹¹ Basic AMS systems require a certain amount of human involvement (directing the cows into the stalls and attaching the suction cups to the udders, etc.). However, recent innovations from companies such as GEA and DeLaval allow for fully automated systems that guide the cattle into the stalls, simultaneously feed and milk the cows, with portions of feed individualized depending on the milk yield and biometric data of each cow; the systems also automatically clean the udders and gently attach the suction cups. GEA’s DairyMilk M6850 also has integrated somatic cell count into the system, which uses EPT (Electrical Permittivity Threshold) technology to ensure the health of the herd.

Beyond dairy farms, when we look at large scale “row crop” farms, innovations such as “precision agriculture” have already achieved significant penetration. Precision agriculture uses satellite-based positioning systems (GNSS, or the more familiar term GPS), sometimes supported by laser-based equipment on the ground, to show the precise position of a tractor or other piece of farm machinery in the field in order to guide such machinery accurately around the field. In combination with a variety of topological and geomorphic data, the system can determine the optimal level of inputs (seed, fertilizer, water, pesticide, etc.) for each area of the field and adjust the system’s sprayers accordingly.

Here too, however, recent innovations are enabling some important upgrades to the solutions. The use of drones now makes precision agriculture far more affordable by eliminating the need for data from satellites or light aircraft. Also, a number of companies are developing systems to enable precision agriculture without a human driver behind the wheel.

Autonomous tractors have been slow to come to market for the simple reason that a commercial tractor can weigh well over 10,000 kg,¹² and therefore any failure in the software could result in catastrophic physical damage. However, recent developments from several companies (John Deere, CNH Industrial, Pattison Liquid Systems, Raven Autonomy, Iseki & Co. and others) suggest that fully autonomous tractors and perhaps other large farm equipment, such as combine harvesters, will be commercially available soon.

While fully autonomous large vehicles are slowly making their way into the field, a menagerie of smaller robotics are already making their presence known in a broad variety of niche applications.

Harvesting is a good example, because the variety of crops and the variation in shape, color, and size of each piece of produce require a dedicated system for every type of fruit and vegetable, usually employing highly sophisticated machine vision systems. Harvest CROO Robotics of Tampa, Florida, has developed an automatic harvester for strawberries; Abundant Robotics of California has one for apples, which uses suction to pluck apples from the tree; and Root AI has developed one for harvesting tomatoes. Weeds present a similar technological challenge of how to identify and remove them without damaging crops. Naio Technologies of France has a range of weeding robots for row crops, vegetables, and vineyards, and a Bosch subsidiary, Deepfield Robotics, has weeding robots specifically for certain crops.

More generic multi-purpose solutions include Fendt’s “Xaver” system. This is a small autonomous ground vehicle (AGV) that operates in a fleet, or “swarm”, to monitor and document the precision planting of corn. Satellite-based GNSS navigation and cloud storage of data and schedules allow for around-the-clock operation, and data on the exact position of every seed and the time it was planted provides the potential to automate other processes such as watering, fertilization, and pest control.

BlueRiver Technology has developed an intelligent “see and spray” system that claims to reduce the use of pesticides by as much as 90%. The system uses machine vision and machine learning to distinguish subtle differences between cotton plants and weeds of many species and sizes, and to spray precisely on the weed only. Bosch – in partnership with BASF Digital Farming – John Deere, and other companies also have intelligent “see and spray” systems.

There are many other areas where technology is making significant inroads, such as Farm Management Systems linking farming operations with financing, insurance, and business and market information. Some systems try to bridge the gap between yield management and business management, allowing farmers to dynamically manage their budget, price assumptions, and day-to-day operations in the field, to accommodate for any changes in supply and demand in the end market.

Some of the simplest solutions are among the most powerful. Skira is an online trading platform for grain, while Foodla connects local producers with online retailers. In developing countries, many of the most effective innovations require nothing more than a mobile phone. FarmX, for example, is an online marketplace for farmers in India to buy and sell their crops, as well as agricultural commodities, in bulk without an expensive middleman. As another example, the US Farmers Business Network is a digital marketplace to buy or rent the inputs of production; it is effectively an Uber-type rental platform for tractors and combine harvesters. Similarly, Farming Revolution offers a “Weeding as a Service” (WaaS) solution, using a range of AGV robots from Bosch subsidiary Deepfield Robotics.

Post-harvest, beyond the farm gate, in the areas of food processing, packing, and distribution, there are still further innovations and solutions in the market and others in development. Logistics automation and supply-chain intelligence is a key area. There is also a desire for greater transparency and security in the food supply chain, which is driven by both regulators and consumers. Food companies are reacting to this demand and are now starting to invest in solutions to provide consumers with the precise provenance of their food.

Perhaps further away on the horizon is the adoption of “blockchain for food”. In general, this is still in an early stage, since the cost-benefit tradeoff for farmers is in some cases not obvious, and often there is much work needed to set up the basic data flows required to run such a system. This task is made complex by the fact that data standards across agriculture are not uniform and highly siloed. However, IBM’s FoodTrust network is one of the more established blockchain solutions for agriculture, connecting a broad ecosystem of participants across the food supply chain with a permissioned, permanent, and shared record of food system data.

Our conviction in the robotics and automation theme as a long-term investment opportunity is based on the fact that we see a rising tide of innovation, an increasingly varied range of use-cases and applications, and a large number of new ventures entering the market.

We believe this is happening because the building blocks for technology innovation, including building robotics and automation solutions, are more accessible and affordable today than ever before. Equally important, there is a growing population of talented people who know how to put these pieces together.

The individual companies mentioned on this page are meant for illustration purposes only and are not intended as a solicitation or an offer to buy or sell any interest or any investment.

To the extent that these materials contain statements about the future, such statements are forward looking and are subject to a number of risks and uncertainties and are not a guarantee of future results/performance.

1.Tim Harford, “Fifty things that made the modern economy”, 2017. Thomas Malthu
2.The United Nations, “World population prospects 2019”, 17.07.2019. Derived on 12.01.2021 from https://population.un.org/wpp/.
3.Harvard Business Review, “Global demand for food is rising. Can we meet it?” 07.04.2016. Derived on 16.01.2021 from https://www.researchgate.net/publication/302466629_Global_Demand_for_Food_Is_Rising_Can_We_Meet_It.
4. Food and Agriculture Organisation of the UN, “Statistical Handbook, 2020”. Derived on 07.01.2021 from http://www.fao.org/3/cb1329en/online/cb1329en.html.
5. The World Bank, “Agriculture and food overview”, 2019. Derived on 04.01.2021 from https://www.worldbank.org/en/topic/agriculture/overview.
6. Climate Watch, the World Resources Institute (2020): 18.4% of total GHG are accounted for by agriculture, forestry, and land use, and a further 1.7% by energy in agriculture and fishing.
7. Ernst & Young, “Digital agriculture: enough to feed a rapidly growing world?”, by Rob Dongoski, 26.04.2018. Derived on 12.01.2021 from https://www.ey.com/en_gl/digital/digital-agriculture-data-solutions.
8. ICT-AGRI-Food (n.d.) Flourish - Aerial Data Collection and analysis, and automated ground Intervention for precision farming. https://ictagrifood.eu/node/36249 Derived on 15.01.2021.
9. Farminova Plant Factory (2019). Derived on 07.01.2021 from https://www.farminova.com/En.
10 United Nations, “The World’s Cities in 2018” [in 2018 an estimated 55.3% of the world population lived in urban settlements]. Derived on 14.01.2021 from https://www.un.org/en/events/citiesday/assets/pdf/the_worlds_cities_in_2018_data_booklet.pdf.
11. Frontiers in sustainable food systems (2020), “Innovation uncertainty impacts the adoptions of smarter farming approaches,” March 20, 2020. Derived on 05.01.2021, from https://www.frontiersin.org/articles/10.3389/fsufs.2020.00024/full.
12. A typical lightweight tractor is the John Deere 4044M “compact utility tractor”, which weighs just 1,705kg, while the large commercial tractor John Deere 9420RX weighs a colossal 24,494 kg. In both cases, this is the “base machine weight” before the addition of equipment.