The United Nations estimates that the world has a total population of about 7.7 billion; it also projects that the population will increase by 2.2 billion in the next 30 years and that it will hit 9.7 billion by 2050. Needless to say, that’s a lot of people to feed.

Topsoil is the upper, outermost layer of soil, usually 10–25 cm deep, where most of Earth’s biological soil activity occurs. That thin layer of soil is responsible for 95% of food produced for human consumption; it’s a vital ingredient in our production of food.

But just 3 cm of topsoil takes some 1000 years for the planet to create. For such a crucial ingredient in the production of our food, it’s a good thing we have an abundance of it right?

Our vast factory-like farms, which cover 40% of the world’s lands, are destroying the topsoil at a faster rate than the earth can create it.

Each year, over 75 billion tons of topsoil are lost to land degradation. Similarly, 12 million hectares of land is lost to desertification — an area that could produce 20 million tons of grains (for those wondering, that’s the equivalent of the world’s 9th largest producer of grain in 2019— Argentina). With an ever-increasing number of mouths to feed and with productive agricultural land becoming rarer, the world is facing a looming food security disaster.

Figure 1: World cropland, which covers about 40% earth’s landmass. Source: USGS

Although our modern farming techniques such as soil tiling (an agricultural technique that turns over and breaks the soil) result in greater yields of crop, they also destroy the soil to the point where it becomes unproductive. Unproductive soil results in lesser yields forcing farmers to till their land further resulting in even more unproductive lands. Granted, while humans are responsible for a lot of the lost topsoil, nature also destroys a fair bit as well. Between ourselves and nature, experts estimate that we have 60 years before our soil becomes unproductive. Food security will become a problem.

Figure 2: Crop yields vary due to climate, management practices, and the mix of crops grown. Map shows average yields for 16 major crops. Green indicates higher yield. Source: Institute of the Environment, University of Minnesota

Farming without soil

Hydroponics

Atits core, hydroponics is a method of growing plants without soil, by exposing the plant's roots to nutrients dissolved in a water solvent. The technology is not new — it supposedly traces its history back to Francis Bacon. Even today, the hydroponic market is worth an estimated $2 billion and is expected to continue growing at a CAGR of 20+ percent in the next 5 years.

Hydroponics is surprisingly efficient at producing food and could one day share in a greater percentage of our production.

In one study, lettuce grown hydroponically “offered 11 ± 1.7 times higher yields but required 82 ± 11 times more energy compared to conventionally produced lettuce.” At the same time, water usage is about 70% less than traditional farms.

Aquaponics

Aquaponics takes the idea of hydroponics a step further by combining the growing of land-based plants with the production of aquatic organisms (think fish). It aims to mimic nature using a near-zero water discharge closed-loop system that offers economic benefits for the production of both plants and fish.

Aeroponics

Developed by NASA for use in space, Aeroponics takes the idea of growing plants to the next level. Instead of water or soil, it relies on mists to deliver nutrients to the exposed roots of a plant.

Compared to traditional farming, it can reduce water usage by 98%, fertilizer usage by 60%, and pesticide use by 100% while also potentially making plants healthier and more nutritious.

The future of food

Currently, the majority of food production is done in rural areas away from larger cities. For certain crops, where a lot of space is required (think corn and grain), traditional farms will continue to play an important role in the foreseeable future or until our soil productivity collapses. For other crops, consumer staples that have continuous harvest seasons such as cucumbers, peppers and lettuce, hydroponics, and other indoor farming techniques in cities can play an important role.

For this reason, technology-enabled farming is not a competitor to our traditional methods of growing food. They should instead be viewed as partners — there’s a lot that can be learned from farmers in the same way that there is a lot to learn from data.

The concept of indoor farms is not new. The fields are barren with failed ventures given the high investment and startup costs. Even so, we’ve seen companies continue to raise huge amounts of venture capital funding. Plenty has raised over $400M in venture funding while 16-year-old Aerofarm, which is operationally profitable, raised $100M during their latest round, valuing the company at half a billion dollars.

Figure 3: An Aerofarm growing operation

Serving tomorrow’s needs with today’s tech

How emerging technology can play a role

With Moore’s Law coming to its end, Christopher Mims, writing in the WSJ, recently introduced the idea of Huang’s Law, which he named after Nvidia’s founder Jensen Huang: “it describes how the silicon chips that power artificial intelligence more than double in performance every two years.”

Indeed, Bill Dally, chief scientist at Nvidia noted that from November of 2012 to May of 2020, the “performance of Nvidia’s chips increased 317 times for an important class of AI calculations”

Figure 4: The performance of Nvidia’s chips increased 317 times since late 2012. Source: WSJ

Whether Huang’s Law holds in the same way as Moore’s Law does remains to be seen. However, continuing technological breakthroughs in chips will move emerging technologies such as AI and ML from the fringes to the forefront of technological change.

As David Roberts writes, Bill Gross once remarked that all commodities are finite with the exception of computing power, which gets cheaper and more powerful. So to ensure low cost in the long-term, you switch out as many inputs as possible for computing power. In other words, you gather data through sensors and cameras, you synthesize and analyze data through AI, and you automate through robotics. You then minimize water, pesticide and nutrient use with data and information technology.

In the world of indoor farms, where the biggest opex cost is labor — according to a Agrilyst survey, labor cost constitutes about 49% of opex cost in a hydroponic farm and 79% for an aquaponic farm — minimizing the number of employees while maximizing the productivity of the ones you have through existing and yet-to-be-thought-of digital technologies, ensures the best chance of succeeding.

The temperature and size of indoor spaces, water used by plants, nutrients dissolved in water, types of crops, among many other things, all differ considerably. Even the same crop grown a couple of hundred meters away from each other will have differing needs. The indoor farming environment is perfectly suited for AI and ML systems to recognize and control.

Figure 5: Survey of profitability by indoor farm system type. Green represents a profitable farm. Source: Agrilyst

Unlike traditional farms that are prone to drought, infestation, and other acts of God, an indoor farm’s environment can be reliably controlled given its sterile nature. More than that, an indoor farm equipped with AI can teach us something about how to best maximize yields and nutrients in the same way it taught itself to play Go and beat the best human.

If we re-imagine farming and the growth of food, the industry could become very comparable to manufacturing. Isn’t that what growing food is at the end of the day? Farming is about optimizing inputs to properly and accurately manage crop; in that same way manufacturing is about optimizing material and cost in turning raw materials into a finished good (though I will be the first to admit that there are more variables that go into growing foods than manufacturing goods).

Where does venture capital come in?

The market for edible plants grown indoors is still small; it doesn’t help that most operations are not profitable. But even then, there’s a huge opportunity in the space. Driven by technology, we are on the cups of seeing farming turn from blue-collar work into something that is more white collar. This said, the actual act of growing food is just a small part of the overall opportunity.

On one hand there’s biotech and the creation of new strains of crops that can better adapt to the realities of our changing environment; Joyn Bio, a joint-venture between Ginkgo and Bayer is already leading the charge. On the other hand, there’s indoor farming and the already existing opportunities.

Unlike most industries where a few giants dominate, agriculture and food is a big enough and always growing market where there’s room for hundreds if not thousands of colossal players. Then, in the middle, there’s the hardware and the software geared to agriculture and food production (indoors or otherwise) that will help propel the other two hands forward: AI, ML, software, sensors, cameras, robots, and other associated technologies.

I think the most interesting opportunity lies in the technologies — hardware and software — around the actual vertical farm. In time, it wouldn’t surprise me to see a dominant monitoring platform/operating system emerge where other applications and technologies can be built on-top.

With other traditional industries already well on their way to tech-driven disruption, agriculture is not far behind. If the future of our food is plant-based meats and lab-grown cultures, then plants grown in indoor farms by machines should also be part of the conversation.

Final thoughts

For the entirety of human history, we’ve farmed land to generate the food needed to sustain our population growth. In an ideal situation, we would be able to continue to farm the way we always have. We have the knowledge and have already invested a lot to create the supply chains to get the food we need. Not having to overhaul it in the next 50 years would be ideal. But obviously, that might not be possible. The economics of indoor farming hasn’t been great thus far, but we’ve made big leaps in key information technologies that would form the backbone of successful indoor farms.

For thousands of years, farming was done at the household level with farmers physically working a small plot of land. The industrial revolution brought new innovations and introduced machinery, making farmers more productive. Then, in the 1960s and 1970s, farms began to consolidate leaving us the giant mega-farms we often see today. At the same time, new agricultural techniques and innovations increased yields.

Despite all those changes, one thing remained constant — we grew most of our food in soil. Soon, that might finally change.

[Cross post from Authors Blog ]