Use of Robots and Artificial Intelligence

in Greenhouse Horticulture
Reaping the Benefits Step by Step

October 2022
Content
Summary
1 Robotics and AI in greenhouse farming: what are we talking about? 4
1 Great Potential Value of Robotization 4
2 Still many steps needed to reap fruits 11
2 Robotization and AI: An Evolution, Not a Revolution 12
3 Robotics and AI will radically change greenhouse farming and supply chain 19
3 Current Use and Supply of Robotics and AI in Greenhouse Horticulture 22

4 Reflections on the Long-term Implications of Robotics 27

5 Conclusions 31

Robotization, AI, and Rabobank 34

About the Study 36

Summary
Great potential value, but difficult process

• Robotics and AI are not panaceas for solving labor • The knowledge and skills of horticultural companies are
shortages or other product-related problems. But in the indispensable in developing robots and AI. For robotics
next decade, robots and crop support software will take players, connecting with the right growers, quickly
over certain tasks in the high-tech greenhouse. adapting products based on experience, service, and
• Growers will not produce more cheaply because of convenience are success factors.
robotization and digitalization, but rather differently. • There are several options for growers to adopt innovations
Managing larger, more international companies will in robotics and AI, from really coinvesting in the companies
become easier, management skills will change, and the to cautiously trying them out using a (trial) subscription.
cooperation with suppliers and customers will become • In the more distant future, a completely different
closer. Pests can also be dealt with earlier, more sustainably, automated cultivation system is a smarter solution for
or more precisely. The quality of the final product may also certain crops. But this requires substantial investments and
improve. All of this has great potential value that is difficult many adjustments, e.g. in varieties.
to estimate.
• Robotization and digitalization also have potential
• Lack of, and high costs of, labor are a major driver of drawbacks. Consider the growing dependence on large
robotization. The cost of labor and management in high- software companies, cyber risks, less flexibility, negative
tech greenhouses is about EUR 150,000 to EUR 250,000 per consumer perception, and less diversity of companies and
hectare, or about EUR 13bn worldwide. products.

3
 Shutterstock id: 1520468858

1. Great
Potential Value
of Robotization
Robotization and AI: What Are We Talking
About?
• For this report, the definition of robotization and AI in greenhouse agriculture includes software/digital platforms for data-
driven crop support or (semi)autonomous cultivation, drones, and various types of robots for the greenhouse and sorting
hall. Techniques commonly found in greenhouse robotics include crop support software, automatic guided vehicles (AGV),
grippers, and vision and sensing techniques.
• A robot is a programmable machine that can perform tasks independently and respond to environmental factors.
• Artificial intelligence (AI) is the ability of a machine, software, or device to exhibit human-like skills – such as reasoning,
learning, or planning. With the help of ‘machine learning,’ the machine, software, or device can learn from mistakes and thus
perform tasks more and more independently. Thus, a robot equipped with AI software should be able to gain experience just
like a human and perform its job ever better.
• Mechanization, automation, and robotization are already common on greenhouse farms. Heavy lifting, for example, has
already almost disappeared through the use of roller conveyors, conveyor belts, palletizers, and AGVs. This is also the case in
the packaging hall and, for example, when sowing, cutting, or potting plants; we see that much of the work has been taken
over by machines. Vision technology, which allows robots, cameras, or drones to monitor crops, has also developed
tremendously.

5
Many Factors Drive Robotization, but Labor
Is the Most Pressing Factor
Drivers of robotization and AI
• A recent Rabobank survey found that companies committed
to digitalization were more likely to achieve revenue growth
Increase
over the past three years than those that did not. flexibility
Market-
• Growers around the world face similar challenges. Finding Increase
productivity
oriented
production
and housing suitable employees, keeping customers happy,
optimizing harvests, and reducing costs are hefty daily tasks
Improve
for virtually every grower. reliability,
Greater span
of control for
quality,
•
management
We believe robots and AI can provide part of the solution for uniformity

some of these challenges, although robots and AI are not

panaceas.
• The most pressing reason to automate is the labor issue in Hygiene and
food safety Value of data
horticulture (see page 7). This involves not only production
work but also cultivation and energy management.
• The sector also faces enormous challenges in the area of Labor
Preserve
shortages
sustainability. AI can help optimize greenhouse climate and Improve
use of production resources. working
conditions

Source: Rabobank 2022

6
Labor Issues Increasing
Migrant workers are not a structural solution
Share of agricultural businesses in the Netherlands that
• On average, a greenhouse horticulture company in the
Netherlands in the year 2021 spent about 27% of its report experiencing obstacles due to labor shortages
25
total costs on labor (including third-party work),
according to data from Wageningen Economic
Research. 20

• In horticulture, the proportion of flexible labor is very

high at 60%. This brings risks and uncertainties. 15

percent
• Filling the labor needs of immigrant workers is 10
becoming more difficult. Housing is scarce, and
migrant workers are coming from further and further
away. This is not a sustainable solution. In addition to 5
rising costs, including required permits, housing, and
employee training, social acceptance and abuse of 0
labor migrants by employment agencies or human

Q1 2012

Q3 2017
Q3 2012
Q1 2013
Q3 2013
Q1 2014
Q3 2014
Q1 2015
Q3 2015
Q1 2016
Q3 2016
Q1 2017

Q1 2018
Q3 2018
Q1 2019
Q3 2019
Q1 2020
Q3 2020
Q1 2021
Q3 2021
Q1 2022
Q3 2022
traffickers are a (chain) risk.

Source: CBS Business Survey, August 2022

7
EUR 13bn in Labor Costs in Global High-tech
Greenhouse Horticulture
Indication of labor costs across operations and
• We estimate the total labor costs (incl. management) of subsectors in global high-tech greenhouse horticulture
global high-tech greenhouse horticulture (approx. 60,000 ha)
at about EUR 13bn per year. For Dutch greenhouse Other (including
Management strawberry and lettuce)
horticulture alone, this is over EUR 2bn. Per hectare, the labor
costs amount to between EUR 150,000 and EUR 250,000. Ornamental
Other
• Labor costs in greenhouse growing are spread across many cultivation
Post-harvest
different types of crops and operations, for example: operations Bell pepper

• Practical work such as pollinating, cutting leaves, applying crop Cucumber

protection, disinfecting, unwinding wire, harvesting, sorting, Other
cultivation Cost:
monitoring (to determine quality and ripeness, for example) work
EUR 13bn
• Management tasks to set strategy, make decisions, and direct
processes related to cultivation, purchasing, sales, energy, etc.
Tomatoes
• We did not include (rising) costs for housing, permits, and Harvesting
60,000ha
recruitment in our estimate of labor costs. We also did not activities

look at mid-tech greenhouses, where benefits can also be

gained through further automation.
Source: Rabobank 2022. This labor cost estimate uses various assumptions about
number of acres, labor hours, and labor costs per hour.

8
Robotics and AI Can Make Greenhouse Work
More Attractive and Partially Replace It
• Robotization should not be seen as merely replacing • Some of the labor that currently takes place on
labor. Robotization offers opportunities to make horticultural farms will shift to software and hardware
greenhouse work more interesting and lighter. suppliers, as they take care of programming,
• Only a limited portion of the costs currently spent on controlling, and maintaining the robots and software.
labor will become costs or investments in robotics and AI
within the next decade. Due to the complexity and
diversity of the greenhouse horticulture sector and high Estimated potential market value of robotics and AI in
labor performance, only some of the work can be global greenhouse horticulture (EUR/year)
automated in the short term. Also, current greenhouse
horticulture managers and employees do not yet always
have the will and capabilities to work with robots and AI.
2022 2032
• The total potential market for robotization and AI could
reach several billion euros per year, based on current
acreage. Successful affordable solutions could eventually
also be applied in mid- and low-tech greenhouses. In
that case, we are talking about a substantially larger < 1bn 2bn-8bn
potential market.
Source: Rabobank 2022
Area has been kept the same as the 2022 level.

9
Example
Robotics contributing to sustainability
Fighting powdery mildew with robot Thorvald
Mildew is a major problem in strawberry and wine grape crops,
among others. Growers fight this pest mostly by using
fungicides. This costs a lot of time and money and is not
sustainable.
The self-navigating platform Thorvald can control powdery
mildew with UV treatment completely autonomously. The
robot is now being used successfully on some crop farms in the
UK, Norway, and the US. Thorvald is a robot that can be used in
both low-tech and high-tech crops. Saga Robotics' ambition is
to expand Thorvald with other useful tools in the future, such
as a picking arm. "But the company has learned not to do
everything everywhere at once," says Albena Todorova, who is Thorvald is a robot from Saga
a member of Saga's Board Of Directors on behalf of the Rabo Robotics that can autonomously
Food & Agri Innovation Fund. According to Albena, who has control mildew in strawberries
been involved in several startups, “It is important for startups and wine grapes, among others.
to keep focus on fast applicable solutions that really benefit
the grower. In addition, it is very important to establish good
partnerships both inside and outside of the sector in which the
company operates.”
10
Example
Saving labor and delivering a better
product
GearStation can do more than just sorting
Gearbox Innovations says it provides "the digital employees of
the future to growers, breeders, and traders.” For example,
Gearbox has developed GearStation. A GearStation provides
in-line quality inspection and grading of vegetables, fruits, and
flowers, whether loose or packaged. It does this by using smart
cameras and through artificial intelligence. GearStation rejects
products if they do not meet the quality requirements set by
the user. It ensures that the wrong product does not end up
with the wrong customer. It also allows the grower to quickly
detect whether certain plants are diseased or if an employee is
not harvesting properly. GearStation from Gearbox is the
"Just as important as products that work well is the service and smart workstation for sorting and
relationship with the customer," says Simone Keijzer, one of quality inspection of fruits,
Gearbox's founders. "Our company works closely with the vegetables, and flowers.
customer, provides installers and trainers to educate growers
and staff, and continuously measures the performance of the
products."

11
 Shutterstock id: 1520468858

2. Robotization and
AI: An Evolution,
Not a Revolution
 "When we start working with a grower, we tell them that it
can sometimes be a painful process, which is why we give
them a good bottle of wine beforehand."
Renji John, Honest Agtech

13
Robotics and AI Make Their Way Into High-
tech Greenhouses Step by Step
Order in which work in high-tech greenhouses will be robotized
• For work at the beginning and end of
the cultivation process, robotization i Observing,
has already made its appearance. counting,
Robotization for harvesting and predicting 4
other cultivation work took longer. ii Harvesting
• Various solutions are already
available for supporting 1 Other cultivation
management in a greenhouse, such activities
iii Sorting, packing,
as data dashboards and preparing orders
(semi)autonomous cultivation Filling pots, potting,
systems. But autonomous cultivation seeding, cutting, etc. 3 2
is also a multistage process (see page
16). The "autonomous greenhouse
challenge" has attracted worldwide Management (of
cultivation, energy, labor,
attention and demonstrated the marketing, etc.)
possibilities of autonomous growing.
Source: Rabobank 2022

14
AI Is Going To Help ‘Green Fingers’ Little-by-little
Development levels of autonomous cultivation
• Autonomous cultivation is gaining momentum and will go through
several phases in the process. Much cultivation knowledge is still in 0 No automation

the heads of growers and advisors. This cultivation knowledge is being

made more accessible by the possibilities of AI.
1 Cultivation support
• AI is usually a supplement to the crop manager's knowledge but in the
future will increasingly become an indispensable tool in a manager’s
decision-making. AI can take into account many different variables, 2 Partly autonomous: grower supervises (hands off)

such as expected market demand, weather forecasts, and crop

condition. It is difficult for crop managers to factor all these variables Conditional autonomous cultivation under conditions: grower
3 intervenes only in case of emergency (eyes off)
into decisions.
• The three editions of the Autonomous Greenhouse Challenge have
Autonomous growing where the grower still determines the strategy
shown that it is possible to control a greenhouse remotely based on 4 (brains off)
sensor data and algorithms. In the 2022 edition, in which lettuce was
grown remotely, the winner even managed to achieve 30 percent
more profit than the reference group of experienced growers. 5 Fully autonomous

Source: InnovationOrigins.com

15
Example
Artificial intelligence can help growers in
several areas
Growing autonomously with Crop Controller
Autonomous cultivation is a buzzword, but not a fad. A
system like Crop Controller collects all the data from climate
computers, sensors, and external data sources, such as
weather forecasts, among others. Based on that data,
cultivation models, the cultivation strategy indicated by the
grower (what the grower wants to achieve), and machine
learning, Crop Controller can continuously calculate and
realize the optimal climate by adjusting the set points in the
climate computer.
According to Ronald Hoek, CEO and co-founder of Blue Radix,
it allows growers to control larger acreages, have more time Crop Controller is a service
for other things, be more demand-driven, or grow more provided by Blue Radix that
sustainably. Before all growers are convinced of autonomous supports the grower by using all
growing, "a change in mindset is still needed," according to the data available and adds
Hoek. "And the same goes for other software and hardware artificial intelligence.
suppliers in the greenhouse. It has to become much easier to
integrate data and technology from different companies."

16
Robotization and Digitalization in
Greenhouse Horticulture: An Evolution, Not a
Revolution • Beginning of large-scale use of robots for specific operations to replace labor
• Range of tasks that robots can do expanded
• Increasing cross-fertilization between innovation in greenhouse farming and
innovations outside horticulture (e.g. precision agriculture in outdoor crops)
• All larger greenhouse farms will use data dashboards, sometimes with AI
software

2022 2025-2032 >2032

• Testing and use of robots to alleviate and support specific tasks, • Use of robots that can perform multiple tasks
such as scouting, leaf cutting, sorting • Autonomous cultivation common among large companies
• Use of robots in propagation and packaging increasingly • Completely new production systems slowly make their appearance
commonplace • Development of varieties and technology better suited to robotics
• It becomes clear who the winners are among the robotics and AI
players
17
 Robotization Uptake Differs per Crop
Diversity within the crop
Rollout to low-/mid- Complexity of the group in terms of labor
Potential global tech market and crop (shape, type, tasks, method of Other factors that may determine development and adoption of
Crop group high-tech market outdoor cultivation etc.) harvesting, etc. robots and AI Final verdict

• Relatively large companies

Tomatoes • Attractive to investors
• Several robotics players already active
• Bell pepper plants very difficult for robots
Peppers
• Relatively little high-tech cultivation

• Relatively large amount of local cultivation

Cucumbers
• Low product price

Lettuce • Automated cultivation systems already available

• Several companies already active in robotics

Strawberries
• Large companies active in cultivation

• Very high diversity

Cut flowers
• High-tech cultivation mostly in the Netherlands

• Automated cultivation systems already available

Potted plants
• A pot around the product helps automation

18 Source: Rabobank 2022

Still Many Obstacles To Overcome
Barriers to robotization in greenhouse farming
• Greenhouse farming is a very diverse industry with many crops
and growing operations. And for robots, greenhouses are an Lack of
unpredictable and confusing environment with moisture, dirt, experience
with
and live material. All this makes development of robotics and Cost of automation
Resistance to
technology
AI complicated. Tomatoes harvested by the bunch make (e.g. sensors)
change

different demands on a robot than cherry tomatoes picked by

the piece. The cultivation of strawberries is completely Uncertainty
different from that of roses, so AI is also very different for the regarding
security,
Distrust due
to unfulfilled
two crops. In addition, much existing technology and software ownership of
data, etc.
promises

for greenhouse farms is already available, and it can be difficult

to integrate new technology with it. It takes time to find
technical solutions to these obstacles. Temporary
nature of Fit into
existing
• What should not be underestimated are adjustments in the certain work
activities processes

knowledge, skills, and way of thinking needed to robotize and

digitalize. It is also necessary to mitigate risks around
cybersecurity, and it is sometimes necessary to adapt Cost of use,
training
Cost of
acquisition
processes and products to new technology. Great
diversity in
products and
Source: Rabobank processes

19
Financing the Purchase of Robots or AI Need
Not Be an Issue
Forms of financing of robots and AI
• We mentioned the cost of acquisition as a possible
barrier to using robotics and AI. Often this is primarily a Form Pros and cons
mental barrier. In practice, we expect few problems with
Full ownership. Encourages commitment and use on the part of the
the cost and financing of a robot or software if it can Investment up
buyer. Uncertainty about service, warranties, and updates after
save a grower money or add value. front
purchase.
• The cost of a robot will be between EUR 150,000 and
Lease contract, with or without option to purchase. Compared to
EUR 500,000, depending on the tasks the robot will Financial or purchase, this offers more financial room and tax advantages.
perform. Labor costs for harvesting can easily exceed operational Disadvantages are possible decrease in value, contract obligation.
EUR 80,000 per hectare. So, for example, a harvesting lease Operational lease often includes service contracts. Leasing is not yet
possible for all robots.
robot will pay for itself within a few years. We already see
this in the sorting and packing process as well. A ‘smart’ Fixed fee per month, e.g. per hectare or per robot. Often more
sorting machine that saves three full-time employees, expensive than purchase or lease, but more flexible (easier to
for example, sometimes pays for itself within a year and Subscription terminate) with more services and updates included, also offers more
financial leeway to user. Because it is an operational cost rather than
can also provide valuable data. an investment, no debts or assets appear on the balance sheet.
• In addition to purchase, robots and AI are often available
through subscriptions or fee-per-use. With software, Pay for use: Fee based on actual use or services provided, e.g. fee per hectare of
SaaS, pay-per- leaf cutting or per tomato picked. Provides security, flexibility, and
‘software as a service‘ (SaaS) is already common. use, pay-for- service for user. Provides additional administration and risk for
service supplier. Often will require minimum monthly fee.

20
Example
Building an intelligent layer around the
greenhouse
Source collaborates with growers to create a smarter greenhouse
Source is building what it calls a layer of intelligence around
the greenhouse. According to Source, AI can take greenhouse
farming to a new level, making knowledge more accessible,
producing more efficiently, and reducing risk in the
greenhouse.

The crop support system that Source is developing supports

growers in defining and executing an optimal strategy.
According to Rien Kamman, CEO and co-founder of Source,
partnering with growers is essential to develop technology of
direct value to the industry: "Our AI is not built to replace Dutch company Source
growers, but to give them the ability to manage more acres (Source.ag) is developing
more efficiently. From the beginning of our partnership with technology to grow fruits and
Harvest House, we looked over growers' shoulders and were in vegetables sustainably with AI-
the greenhouses of Agro Care and Rainbow, among others, on powered greenhouses.
a weekly basis to really understand how AI could support
them."

21
 3

Shutterstock id: 2132351563

3. Current Use
and Supply of
Robotics and AI
in Greenhouse
Horticulture
Both the Use of Robotics and Investment in
Robotics Are Relatively Limited
Venture capital investments in ‘agrifood tech,’
• The installation of industrial robots globally grew at an average according to AgFunder
annual rate of 9 percent from 2015 to 2020 (source: IFR). Due to
travel restrictions and lack of computer chips and other
materials, the growth rate dropped in 2020 and 2021.
• In Europe, the use of robots is lower than in the US and Asia.
• According to CBS, 6 percent of all Dutch businesses use robotics.
This percentage is lower for SMEs and the agricultural sector. The
limited use has to do with the lack of capacity, time, knowledge,
and financial resources, according to CBS.
• Although the labor issue is large and widespread, venture capital
is finding its way into automation in a limited way. Globally, USD
51.7bn in venture capital was pumped into ‘agrifood tech’ in
2021, according to AgFunder, of which USD 900m was in
robotization, automation, and mechanization of mainly outdoor
crops. From the conversations we had with several robotics
players, the relatively long development time is the main reason
for the relatively low investments.
Source: AgFunder 2022
Note: Financing is stated in US dollars.

23
The Breeding Ground for Greenhouse
Farming Robotics and AI Differs per Region
Drivers of successful greenhouse robotics and AI development
• Robotics and AI for greenhouses are being
developed in different types of organizations –
think technical and agricultural universities
and especially startups. Sometimes those • Presence and extent of high-tech greenhouse
startups are spin-offs from universities. In Demand horticulture (potential 'testing ground' and
market)
addition, several existing suppliers to the factors • Existence of bottlenecks (e.g. labor shortages) in
horticulture chain have now developed greenhouse farming
applications in the field of robotics or AI or are
collaborating with players in robotics.
• Robotics and AI companies specifically for
greenhouses can be found in the US, the
Netherlands, Israel, Belgium, Japan, and the
• Presence of greenhouse horticulture supply
UK, among others. companies
• Several factors determine where development Supply • Availability of capital, venture capital
• Incentive from the government
of robotics and AI for greenhouse horticulture factors • Presence of rotobics companies for other industries
arises and where they are eventually
• Scientific research
successfully applied.

24
Robotics and AI: A Diverse Playing Field
Examples of players in AI and robotics for greenhouses
Existing suppliers (greenhouse) horticulture Startups and scale-ups Other

Several existing suppliers (e.g. in automation, Most companies in robotization are startups and scale-ups, some with a Other companies and organizations
transportation, and sorting) are active in background in agribusiness, some in engineering developing robotics and AI include
robotization or AI through participation or in- • ADI universities or companies in the industry
• Lenzeel
house R&D • Ellips (incl. Aris BV)
• Arugga AI Farming • Organifarms
• Certhon (including with Denso) • MTA Group
• Automato Robotics • PATS Indoor Drone
• GREEFA • Panasonic
• Blue Radix Solutions
• Havatec • Siemens
• Corvus Drones • Polybee
• Hoogendoorn (LetsGrow) • VDL CropTeq Robotics
• Dogtooth • Root AI (AppHarvest)
• ISO Group • Vineland Research and Innovation
• Ecoation Innovation Solutions • Saga Robotics
• Kind Technologies (Crux Agribotics, Centre
• Egatech • SAIA Agrobotics
Martin Stolze) • Wageningen University & Research
• Four Growers • Source.ag
• Octiva (Priva Kompano, Octinion)
• Gearbox • Tortuga AgTech
• Ridder (MetoMotion)
• Honest AgTech • Xihelm
• Royal Brinkman (Berg Hortimotive,
HortiKey) • Inaho Inc. • 30MHz
• Veditec Greenhouse Robotics • Koidra

Sources: Various public websites. Note: List is not exhaustive.

25
Horticulture Itself Will Need To Invest Time
and Capacity
Robotics and adaptations by greenhouse entrepreneurs
• Successful development and application of
robots and AI in greenhouse farming requires
grower co-creation. That means growers
investing time in testing innovations and
sharing knowledge to improve development. Selection: choice of robotics partner –
Opinions differ on the need for growers to 1 the one who can solve concrete
coinvest financially in the development of Selection problems and adapt quickly
robots and AI.
Pilot phase: invest capacity and
• When an innovation is ready for market, a time to provide accurate
grower will also need to innovate socially to
use the new technology successfully. That 2 information and feedback;
information on labor input, revenue
sometimes means adjustments to processes, Pilot phase model, etc.
crops, or types of employees. Management
skills must also grow with technological Structural adaptations: train
advances. 3 employees or recruit other
employees to work with robotics
Structural
adaptations and AI, to analyze data, etc.

26
 3

Shutterstock id: 2132351563

4. Reflections on
the Long-term
Implications of
Robotics
Long-term Consequences To Consider
Robotics and AI will eventually change greenhouse horticulture structurally

Competitive positions are changing Less crop diversity and more industrial More sustainable and efficient
Differences in innovation rates are changing production Application of robots and AI can help, for example,
competitive relationships and shifting production Because some crops are more amenable to reduce use of production resources (for example,
locations. Digitalization and robotization make low automation than others, there is a chance that this water or crop protection) or reduce waste of
labor costs less of a determinant of growing could reduce crop diversity. There is also a chance finished goods.
locations and will result in growing taking place that some consumers will reject products from a
closer to consumers. more industrial production system.

Economies of scale and organization New players in horticultural technology Risks related to cybersecurity, privacy,
The organization of the highly automated and the end of some traditional players dependency
greenhouse horticulture company of the future will New players in AI, for example, could mean the end Dependence on technology also carries risks. A lot of
look different than it does today. Automation makes of existing players in greenhouse automation. data is collected that others can potentially use.
it easier to manage large companies and companies Cybercrime can completely cripple a company. And
with more locations and for external players to enter companies can become very dependent on large
horticulture. Competencies such as analytical skills software players, for example.
and creativity will become more important for
managers at horticultural companies.

28
Towards Completely Different Growing Systems?
Promising for a number of crops, but potential drawbacks

• Replacing humans one-on-one with robots in the greenhouse is a complex task. Therefore, is an entirely different
cultivation system a better solution than robotizing diverse work within existing cultivation systems?
• In ornamental horticulture (e.g. pot orchids) and cultivation of leafy crops (lettuce, herbs), in some cases, cultivation is
already largely automated, and the plants leave the greenhouse when they are ready to be harvested/sold. In vertical
farming, we also see cultivation systems that no longer require people in the cultivation facility itself, but only outside it.
This is easier for lettuce to achieve than for cucumbers or tomatoes, which usually grow on long tethered plants. An
automated cultivation system without people in the greenhouse has the added advantage – besides labor savings – that
risks related to diseases and food safety are further reduced and greenhouse utilization can be improved. Of course, there
are also disadvantages to a completely different cultivation system, including the large investment that will be required,
different varieties needed, completely different working methods, etc.
• Despite the big move to completely new cultivation systems, we see several parties working on this. Examples are the
Gronos growing system for tomatoes from greenhouse builder Certhon (see page 30) and SAIA, a tech-spinoff of
Wageningen University & Research.
• There are also potential drawbacks to fully automated, more industrial production systems for food, flowers, and plants.
There will be a group of consumers who prefer not to have ‘industrial’ food and instead need small-scale and/or more
‘naturally’ produced tomatoes or flowers. There may also be resistance to the growing power of large global robot and
software suppliers and further economies of scale that may result from further automation and digitalization of
horticulture.

29
Example
A new growing system for tomatoes
Gronos is a new cultivation system for tomatoes
Sustainability requirements and labor needs are among the
biggest challenges tomato growers face. The current tomato
growing system is difficult to fully automate, in part because
tomato plants grow on a long wire and produce ripe tomatoes
at different times over nearly a year.
Certhon has therefore designed a new cultivation system with
small, compact tomato plants that have a short cultivation
cycle of 15 weeks. Cultivation takes place in an indoor farm or a
combination of indoor farm and greenhouse. In this system,
tomatoes can be grown fully automatically without the need
for labor. They can also be grown with minimal water use, no Gronos is a fully automated
pesticides, little waste, and with only electricity (i.e. no other cultivation system for tomatoes
sources of energy or heat). According to CEO John van der developed by Certhon, developer of
Sande, "Gronos allows cultivation in all corners of the world. high-tech horticultural projects
Gronos is complementary to existing cultivation systems and
responds to the trend of locally produced, fresh, tasty, and safe
food.''

30
 Shutterstock id: 1520468858

5. Conclusions
Conclusions
• The market for robots and artificial intelligence represents enormous potential value. The potential value of further
sustainability, a higher-quality harvested product, better working conditions, and a higher product price through better
alignment with demand is difficult to express in euros. Labor and management costs in global high-tech greenhouse
horticulture are currently about EUR 13bn. In part, these costs will be replaced by the costs of investing in robotics and AI.
• It will take at least another decade before robots and AI prove their worth and are widely used in greenhouse farming. This
will go step by step. We already see robots at the beginning and end of the production process and expect them to be used
for specific tasks such as crop protection, monitoring, and leaf cutting soon. Slowly, additional tasks will be added, such as
harvesting. A robot that can replace all tasks is still far in the future – just like completely different cultivation systems, which
we are already seeing the first examples of.
• Co-creation of robots with growers is necessary for successful robot development. By testing and freeing up time and
capacity to provide feedback and share knowledge on labor costs, for example, innovation can be accelerated. It is
important to collaborate here with those parties who outline honest expectations and can make quick adjustments.
Awareness about the use and ownership of data that AI and robots collect and create is also important.
• In the long run, robotization and digitalization are changing competitive positions and the balance of power in the
greenhouse chain. This produces winners and losers. Other possible negative consequences to consider are cybersecurity,
dependence on (likely very large) software suppliers, and consumer resistance to ‘industrially’ produced food.

32
Critical Success Factors for Robotization and
AI in Greenhouse Horticulture
Business model for both Partnerships Service and Social innovation
supplier and user convenience
Innovating in a greenhouse
Solutions for robotization and We are convinced that the Greenhouse horticulture horticulture company goes
AI must make money for both right technology will be companies do not want to beyond acquiring and using
provider and user. This sounds developed, but whether that trade the problems new technology. Changing
like a no-brainer but is often technology will find its way to surrounding labor for management methods,
not the case in practice. Very users depends, among other problems surrounding the use collaborating with
high labor performance things, on the extent to which of robots. Defects must be technology, redesigning
makes it challenging to parties are able to cooperate fixed quickly, robots must be business processes, etc. is at
replace humans with robots. with each other. This involves easy to use, and solutions least as important for
cooperation between must fit into current processes successful innovation.
technology suppliers and with and systems as much as
growers and technology possible.
companies outside
horticulture, so that useful
innovations from other
sectors can also quickly find
their way into horticulture.

33
 Shutterstock id: 1520468858

Robotization, AI,
and Rabobank
Some of Rabobank’s Services
Support for startups and inspiration and financing for greenhouse horticulture

Rabo Innovation Loan Robohouse The startup and scale-up team

We are happy to help finance innovations, even if there Rabobank wants entrepreneurs to grow future-proof, and Successfully founding a startup and scaling it up to a
are no proven results, cashflow, or paying customers yet. Robohouse is using its field lab to put the possibilities of mature company is challenging. The Rabobank Startup
Are you a new business working on developments that robotization and automation into practice. For example, and Scale-up Team wants to make a positive contribution
contribute to digitalization/robotization or making society entrepreneurs can be thematically inspired at demo days. to your chances of success and the speed at which your
more sustainable? Please request the brochure for the They can work with other entrepreneurs during Knowledge company develops. We do this by helping you with the
Rabobank Innovation Loan and share your plan with us. Sprints and receive one-on-one guidance when they want to most suitable financial solutions and insights and by
https://www.rabobank.nl/bedrijven/zakelijk- start working on a concrete use case. making valuable connections.
financieren/financieringen/rabo-innovatielening https://robohouse.nl/activities/news/coming-up-kennis- https://www.rabobank.nl/bedrijven/startups-en-scale-
sprint/ ups/artikelen/011168651-het-start-up-scale-up-team

Rabo F&A Innovation Fund

Rabo F&A Innovation Fund - Rabo Investments

FoodBytes! by Rabobank is a global digital platform for startups, corporates, and investors. For this, young companies can apply that we think will make a difference in the F&A sector. These are companies that respond to
trends and challenges in the world of food & agribusiness, for example in the field of sustainability or consumer trends. We help these startups on their way by connecting them to the established order of large companies,
investors, and consultants from Rabobank's broad network. For more information see www.foodbytesworld.com & http://www.foodbytesdigital.com/

35
 3

Shutterstock id: 2132351563

About the Study

Contributions
Many thanks to all those who contributed to this publication through interviews,
information, or photographs:
Albena Todorova – Rabo Corporate Investments / Saga Robotics Sebastiaan Schreijen - RaboResearch
Bart van Tuijl - SAIA Agrobotics Simone Keijzer - Gearbox Innovations
Bram Tijmons - PATS Indoor Drone Solutions Tom Coen - Octinion
Colinda de Beer - InnovationQuarter Wilco Schoonderbeek - Horticoop
Henk Jan van Ooijen - ISO Group Wim van der Meijden - ISO Group
Jan Meiling - StartLife
John van der Sande - Certhon
Lambert van Horen - RaboResearch
Marc Zuidgeest - Frestia
Maren Schoormans - Honest AgTech
Peter Hendriks - LetsGrow
Renji John - Honest AgTech
Rien Kamman - Source.ag
Ronald Hoek - Blue Radix
Ruud Barth - SAIA Agrobotics

37
Sources:
Agfunder, “2022 AgFunder AgriFoodTech Investment Report.”
Agrimatie.nl – informatie over de agrosector – Wageningen University & Research.
Ballestar, M., Díaz-chao, Á., Sainz, J., Torrent-Sellens, J., “The impact of robotics on manufacturing: A longitudinal machine learning perspective,”
10.1016/j.techfore.2020.120348, Technological Forecasting and Social Change, 2020.
Beer, C. de, Vliet M. van, Sar, P. van der, Maijers, W., “The role of digitalisation in Feeding and Greening the Megacities: Digitalisation vision for Greenhouse
Horticulture” June 2021.
Beer, C. de, “Autonomous cultivation is the holy grail for greenhouses; but what does that look like?” Innovation Origins, June 11, 2022.
Groentennieuws.nl, “Winnaar en tactieken autonome slateeltwedstrijd bekend,” July 4, 2022.
Hardeman, S, Snoep, N en Kampen van, N., “Succesvolle digitalisering vraagt om aandacht voor technologie én bedrijfsvoering,” Rabobank, April 1, 2022.
Horen, L. van, ‘Een goed klimaat voor tuinbouwtechniek,” Rabobank, December 2020.
International Federation of Robotics, “Presentation World Robotics press conference extended version,” October 28, 2021.
McKinsey & Company, “Industrial robotics: Insights into the sector’s future growth dynamics,” July 2019.
Pekkeriet, E and Splinter, G, “Arbeid in de toekomst; Inzicht in arbeid en goed werkgeverschap in de tuinbouw,” Wageningen, Wageningen University & Research,
report 2020-002.
Raaphorst, M and Benninga, I, “Kwantitatieve Informatie voor de Glastuinbouw (KWIN) 2019,” Wageningen University & Research, 2019.
Rijswick, C.W.J., van en Merrienboer, S. van, “Samenwerken aan toekomstbestendige arbeid land- en tuinbouw,” Rabobank, May 4, 2020.
Sparrow, R. and Howard, M. 2020. “Robots in Agriculture: Prospects, impacts, ethics, and policy,” Precision Agriculture Online First, October 24, 2020.

38
Contact details
RaboResearch Sector Management
Rabobank Rabobank

Cindy van Rijswick René Gomersbach

Global Strategist - Fresh Produce Sector Manager Horticulture - F&A
Sector Management
Mobile +31 6 12446402
Email [email protected] Mobile +31 6 53716982
Email [email protected]

Startup & Scale-up Team

Rabobank Rabobank

Gerben Dijksterhuis Niels Candy

Startup and Scale-up banker F&A Theme Manager Innovation

Mobile +31 (6) 20837724 Mobile +31 6 83173644

Email [email protected] Email [email protected]

Rabobank

Baibing Yan
Finance Specialist F&A

Mobile +31 (6) 29473915

Email [email protected]
Disclaimer
This document has been prepared exclusively for your benefit and does not carry any right of publication or disclosure other than to Coöperatieve Rabobank U.A. (“Rabobank”), registered in
Amsterdam. Neither this document nor any of its contents may be distributed, reproduced or used for any other purpose without the prior written consent of Rabobank. The information in this
document reflects prevailing market conditions and our judgement as of this date, all of which may be subject to change. This document is based on public information. The information and
opinions contained in this document have been compiled or derived from sources believed to be reliable, without independent verification. The information and opinions contained in this
document are indicative and for discussion purposes only. No rights may be derived from any potential offers, transactions, commercial ideas et cetera contained in this document. This document
does not constitute an offer or invitation. This document shall not form the basis of or cannot be relied upon in connection with any contract or commitment whatsoever. The information in this
document is not intended and may not be understood as an advice (including without limitation an advice within the meaning of article 1:1 and article 4:23 of the Dutch Financial Supervision Act).
This document is governed by Dutch law. The competent court in Amsterdam, The Netherlands has exclusive jurisdiction to settle any dispute which may arise out of or in connection with this
document and/or any discussions or negotiations based on it. This report has been published in line with Rabobank’s long-term commitment to international food and agribusiness. It is one of a
series of publications undertaken by the global department of RaboResearch Food & Agribusiness.