Let’s first look into the past
To attempt to understand CEA’s future, we need to understand its past to promote the narrative that innovation (especially in agriculture and for staple crops) requires time and patience and is much needed for the sector.
Organised plant agriculture began with open-field cultivation of precursor crops in the fertile crescent at least 12,000 years ago with semi-controlled structures, which we now refer to as greenhouses, arising in around 30 CE[i]. It took much longer to get to controlled-environment-agriculture (“CEA”) models (vertical farming falling under this definition), which have been around for around 72 years[ii] following the development of phytotrons.
After the development of phytotrons, research was initially focused on space exploration and looking forward to the next millennium. The advent of the most recent phase of CEA, with a greater emphasis on vertical farming, has only been around for the past 15-20 years or so and has been predominantly U.S.-led, riding on the coattails of the growing climate change narrative, substantial initial investments in the hundreds of millions of dollars into vertical farming operators, the availability of capital, technological improvements, and the search for a new asset class that could deliver on returns comparable with other climate-focused technologies, such as solar.
The concept of CEA is, therefore, nothing new, but the key question for today’s vertical farms that remains broadly unanswered is whether the combined technologies, models, and teams that make up the businesses involved in this sector’s “new wave” can simultaneously deliver on both consumer demands (ideally at price parity) and investor return expectations.
Hardware “bargain basements”
Investments that have gone wrong in this space, have been discussed in the press and even by Founders (widely referenced), highlighting many shared characteristics. Looking at many of these businesses from a review of their system (the hardware and software to operate it), either via available evidence or in some cases my own direct visits to their facilities after the fact, they share one or more of the following characteristics; high-maintenance moulded plastics that are susceptible to high bacterial loads over time, single spectrum or limited spectrum lighting; limited or no automation across the entire system; no software or third-party software that only performs very basic functions; and in general, assets with a short useful life that do not integrate and in most cases, would be incredibly difficult to scale. The list could be much longer, but that’s for another article.
Any vertical farm, whether low-tech or high-tech, involves many different technologies that need to seamlessly work together, with continuous data flow and the formation of a “system” that has the capability to become more productive over time, with a significantly lower dependency on humans from operations, data collection and interpretation.
This “integrated systems approach” is not being followed by a (very) high percentage of the sector; however, new vertical farming start-ups are announced almost daily online, and based on the equipment on view, the writing is already on the wall from day one in many cases. Sticking to the metrics that focus on the farm, these choices could be for many reasons. My main assertion from experience is that there is still a lack of understanding of the need for full integration and the benefits this brings, and combinations of ignorance, entrepreneurial optimism and constrained initial capital, which together overlook the need for continuous science-led enrichment and automation.
Vertical farming in some geographies is becoming a “bargain basement”, with equipment from failed farms sold online and then set up again under a different brand and broadly the same business model. Albert Einstein’s observation “doing the same thing over and over again and expecting different results is the definition of insanity” springs to mind. Moreover, a short-term view is rife, which assumes that lower cost, more “basic” systems from multiple suppliers is the best place to start and then scale.
Five or ten years ago, this would have been completely understandable, but there is now enough data in the public domain that clearly disproves this. Images are often enough in many cases to judge yield performance and other metrics to determine whether the farm is likely to succeed. This is, of course, discounting the existence of plant scientists, agronomists, and seasoned growers who can improve matters, but only to a point. Case in point, images of operators walking around a farm with an iPad or another form of tablet, taking photos, and manually checking crops, first, is never going to be scalable and second, suggests a non-integrated system, where the data infrastructure and technologies involved in the process should be doing the work.
Some suggestions from experience
Don’t go for the cheapest capital expenditure (“CAPEX”) option from the outset. Understand that higher-quality materials pay dividends in the long run. Consider the “all in” lifetime cost over a suitable time horizon accounting for maintenance of these assets. A vertical farm may not be with you for as long as a tattoo, but the same mentality should be adopted. Look for quality, experience, capabilities, and specialization. This almost always means spending more money at the outset, but 20 years down the line, when you look at your farm (or skin with your tattoo), you’ll know you’ve made the right investment—and by then, the initial cost difference won’t even come into your mind.
Above all, I will share our biggest learning from our days as operators and why my company, Vertical Future (“VF”) has chosen a different path to support the industry. Buying a system that is not fully integrated from seed-to-harvest with the system-wide software capabilities to increase productivity across the life of the asset means that you are foregoing long-term, valuable, low-cost data learnings and settling for predominantly human-led learnings and know-how, which (while valuable), will never be able to compete in speed, interpretation, and outcomes. Of course, a plant can only grow so fast and produce so much biomass, and we will at some point reach a performance plateau for every cultivar, at which time all future improvements will have to be derived from operations, cost of production savings, and wider operational value accretion, but this can still be reached much faster through better data.
Can you hear the music?
For those who have seen the movie Oppenheimer you will recognize that I am borrowing this title from the words spoken by a prominent Danish Physicist, Niels Bohr when referring to Algebra. He was referring to the difference between being able to do the math and not really understanding it. Market data shows that many new founders, companies, and investors are not “hearing the music” and moreover, history is repeating itself as the same mistakes are occurring.
For the Vertical Farming Industry, I am hearing two “notes.” These are (a) system-wide integration through in-house technology production; and (b) the obvious need for data-and-science-driven growing algorithms to accelerate progress. Together these notes create the music of a completely integrated system that, using a base hardware that will, by its nature depreciate in value, combines with software at every point to make a whole system that appreciates in value by becoming more productive over time and eventually, self-learning. By linking these “notes” facilitates the ability to collect, collate, and efficiently analyse data from many sources. This also allows for the efficient development of better-growing algorithms and the expedient repopulation of these learnings into our customers’ systems.
Again, whilst some human intervention, oversight, and expertise will always be required, walking around a vertical farm (or greenhouse) with a touchscreen iPad (returning to this example) and visual inspection of crops will never be able to compete with the capabilities of a fully integrated software system, and this is true for both small farms and large farms, except the negative externalities are proliferated for the latter.
How are we applying these notes? To cite a few examples from our work at VF, our in-house Software-as-a-Service, DIANA, moves beyond simple functionalities exhibited in your “average” vertical farm, such as temperature, humidity, and light control. System-wide, DIANA tracks electrical current, water pressure, light modulation every 6m2 of growing space, crop height at germination stage triggering autonomous transfer to grow zones, and the entire lifecycle (i.e., stage of growth) of each 0.25m2 tray from seed through to harvest, facilitated through barcode tracking.
Through full integration of every component across a VF system, irrespective of farm size, we’re able to develop better-growing algorithms and better understand and improve preventative and predictive maintenance regimens. Tracking current, for example, notifies the system immediately if there is an issue with an LED, whether in a 5,000m2 system or a 50,000m2 system. This results in better, faster, and more efficient maintenance and knock-on cost savings. The larger the facility, the greater the efficiency savings and economies of scale.
What’s next: my predictions
1. Market consolidation will take place in specific geographies and will, in many cases, be a forced choice with more acquisitions than mergers on the cards, creating somewhat of regional oligopolies with sizeable production sites in peri-urban and rural areas. Probabilistically, this will progressively push “backyard” and “urban farms” (which do deliver a lot of value, especially regarding education, training, and health) into increasingly niche markets after higher pricing strategies and localized partnerships are attempted.
2. The scarcity of components used in vertical farms will create competitive tension and elongate delivery times, especially for larger projects. This is a natural outcome of the manufacturing capabilities and control over natural resources.
3. Patents will continue to be (broadly) a waste of time and resources for vertical farms (just like they are for traditional farmers). Investors love these, but except for core, breakthrough features, an entire vertical farm cannot be patented. In rare exceptions where these types of claims (or others) have been granted, these are likely not defensible (I am, of course, talking broadly, not to any specific patent application).
4. Genetic breakthroughs for more crop varieties will act as an external propellant for yield increases in quality improvements, which will support the entire industry, subject to a suitable mechanism to access improved seed varieties. This will be in parallel with the legalization of gene-edited seed variants in more countries following the U.K.’s legalisation earlier this year[iii].
[i] Paris, H. & Janick, J. 2008 What the roman emperor Tiberius grew in his greenhouses 1 Cucurbitaceae 2008, Proceedings of the IXth EUCARPIA meeting on genetics and breeding of Cucurbitaceae Pitrat, M. INRA Avignon, France 21–24 May 2008
[ii] Kramer, P., Helmers, H. & Downs, R. 1970 SEPEL: New phytotrons for environmental research Bioscience 20 1201 1204
[iii] https://www.gov.uk/government/news/genetic-technology-act-key-tool-for-uk-food-security
This article was first published on 3rd June 2024 on Jamie Burrows’ (Founder & CEO of Vertical Future) ‘The Vertical Farming Journey’ newsletter. View the original newsletter here.
