For as long as I can remember, people have been hyping vertical farming — growing crops indoors, using vertical space to intensify production.
Its virtues, relative to conventional agriculture, have long been clear. Indoors, the climate can be controlled year-round. Pests can be minimized, and with them pesticides. Water and nutrients can be applied in precise quantities. By going up rather than out, a vertical farm can produce more food per acre of land. And by siting close to an urban area, it can reduce long distribution chains, getting fresher food to customers’ tables, quicker.
Its drawbacks have become equally clear. They mainly come down to cost. Farming well requires deep know-how and expertise; it has proven extraordinarily difficult to expand vertical farms in a way that holds quality consistent while driving costs down. Optimizing production at a small scale is very different from doing so at a large scale. The landscape is littered with the corpses of vertical-farming startups that thought they could beat the odds (though several are still alive and kicking).
Now a young Silicon Valley startup called Plenty thinks it has cracked the code. It has enormous expansion plans and a bank account full of fresh investor funding, but most excitingly, it is building a 100,000 square foot vertical-farming warehouse in Kent, Washington, just outside of Seattle, your author’s home town. That farm is expected to be open and delivering produce locally by midyear, and is designed to produce 4.5 million pounds of greens annually. Your author, in keeping with coastal elitist stereotypes, is a fervent lover of greens.
In part because I now have a personal stake in the matter, I thought I’d take a look at the company, its prospects, the environmental benefits it promises, and — perhaps most importantly — some of the unnerving social and political implications of a vertical farming revolution.
Plenty wants to build a farm near your city
Plenty is at the center of a veritable hurricane of buzz at the moment.
It checks all the boxes: It recently got a huge round of funding ($200 million in July, the largest ag-tech investment in history), including some through Jeff Bezos’s investment firm, so it has the capital to scale; it is leaning heavily on machine learning and AI; it has endorsements from several Michelin-rated chefs (“I’ve never had anything of this quality,” a former sous-chef at French Laundry, Anthony Secviar, told Bloomberg); it is in talks with several large distributors in the US and abroad; heck, it even lured away the director of battery technology at Tesla, Kurt Kelty, to be executive of operations and development. (You’re nothing in Silicon Valley without an ex-Tesla exec.)
“I wanted to figure out where I would contribute to the next big wave,” Kelty told Bloomberg. “I see my next 10-year-run as growing Plenty.”
So, what’s the big deal?
If you want to really dig in, Bloomberg has the best feature story on Plenty (see also Fast Company), but I’ll quickly run through what the company is up to. It’s helpful to read what follows against this list of nine reasons vertical farms fail, by Chris Michael, CEO of vertical-farming company Bright Agrotech. In a sense, Plenty is a response to previous failures.
The company is run by CEO Matt Barnard, a former private equity investor, and CTO Nate Storey, an agronomist who did his doctoral work in tower farming. (Storey also founded Bright Agrotech, which he left to join Plenty. In June, Plenty acquired the company.)
Plenty grows plants on 20-foot vertical towers instead of the stacks of horizontal shelves used by most other vertical-farming companies. Plants jut horizontally from the towers, growing out of a substrate made primarily of recycled plastic bottles (there’s no soil involved). Water and nutrients are fed in from the top of the tower and dispersed by gravity (rather than pumps, which saves money). All water, including from condensation, is collected and recycled.
The plants receive no sunlight, just light from hanging LED lamps. There are thousands of infrared cameras and sensors covering everything, taking fine measurements of temperature, moisture, and plant growth; the data is used by agronomists and artificial intelligence nerds to fine-tune the system.
The towers are so close together that the effect is a giant wall of plants.
Currently, Plenty is focusing on leafy greens and herbs — varieties of lettuce, kale, mustard greens, basil, etc. — but it says it can use the system to grow anything except root vegetables and tree fruits. Strawberries and cucumbers are coming up next. (It’s worth noting that anything beyond leafy greens requires more light and thus more energy, so the source and cost of an indoor farm’s electricity is of keen interest.)
There are virtually no pests in a controlled indoor environment, so Plenty doesn’t have to use any pesticides or herbicides; it gets by with a few ladybugs. The produce from Plenty’s San Francisco warehouse is certified organic, but leaders in the industry also like to stress that vertical farming is local, with an entirely transparent supply chain. (Why yes, you can also get that at your local farmers market.)
Bottom line: Relative to conventional agriculture, Plenty says that it can get as much as 350 times the produce out of a given acre of land, using 1 percent as much water. “It is the most efficient [form of agriculture] in terms of the amount of productive capacity per dollar spent,” Barnard has said. “Period.”
It’s worth reading those claims again, as they are pretty eye-popping. The next grandest claim in the industry is AeroFarms, a Newark, New Jersey company with nine indoor farms, which says it can get to 130 times the amount of produce per acre.
What’s more, Plenty says its products taste better than most of what customers now have access to. Around 35 percent of fruits and vegetables eaten in the US today are imported. Leafy greens travel an average of 2,000 miles to reach your plate. Some produce has been on ships and trucks for two weeks before it reaches the table — having lost, by some estimates, 45 percent of its nutritional value along the way. Produce is bred to survive that long journey with its aesthetics, but not necessarily its flavor, intact.
Plenty plans to build warehouses not inside major cities, but just outside them, next to distribution centers, to minimize the time its food spends in transit — it wants produce to go from harvest to table in hours, rather than days. If it can do that, the company will be able to grow and sell a wide variety of rare and heirloom breeds, which are more tender and flavorful than what’s available at the supermarket, but less resilient to long journeys.
In fact, Barnard says he will save more money on trucks and fuel than he spends on facilities and power.
The company’s goal is to build an indoor farm outside of every city in the world of more than 1 million residents — around 500 in all. It claims it can build a farm in 30 days and pay investors back in three to five years (versus 20 to 40 for traditional farms). With scale, it says, it can get costs down to competitive with traditional produce (for a presumably more desirable product that could command a price premium).
If it can back up those claims in practice, Plenty might not revolutionize global agriculture, but it will sure as hell establish vertical farming as a real thing.
Now, to be very clear: It would be a terrible mistake for anyone to take investment advice from me. I’m not an industry analyst. I have no idea if Plenty will ultimately succeed. It could face difficulty finding affordable urban land; it could have trouble replicating the carefully controlled conditions of its initial warehouse; quality could slip as it output rises; consumers could reject the products for any number of reasons. Many previous alt-farming startups got similar buzz, only to fail. Scaling up is full of peril.
But I do think, if Plenty is not the early Google of this space, some other company will be, soon enough. The traditional barriers of cost and energy that have blocked the industry from growing are crumbling. And the way it’s happening carries some fateful lessons.
Plenty is replacing stuff with intelligence
Entrepreneur Bill Gross said something in a talk once that I never forgot. Every commodity, he said, is finite, and is eventually going to get more expensive — except computing power. Computing power just gets cheaper and cheaper and cheaper. So to the extent you want to ensure low costs in the long term, he said, you substitute computing power for other commodities — intelligence for stuff, as I like to put it.
By intelligence I mean, roughly, the ability to gather data (through sensors), synthesize it (through computing power), and use it to optimize operations (through machine learning). Optimization wrings waste, i.e., extraneous stuff, out of the process.
I have argued before that the current energy transition may well move faster than previous transitions, precisely because it is driven by information technology. To put it crudely, if past energy transitions have replaced giant stuff with other giant stuff, this one is going to replace giant stuff, at least in part, with intelligence.
That same process is taking place in agriculture; that’s part of what vertical farming represents. Zack Bogue, a Plenty investor through Data Collective, a San Francisco venture capital fund, put it this way: “We’re pretty excited about that space because some of the hardest problems in agriculture are now lending themselves to an algorithmic or computational or applied machine-learning solution.”
As Barnard himself put it, agriculture is a “giant optimization problem.” The challenge is to use just the amount of energy, water, and nutrients necessary to produce food, and no more. Big Ag has struggled with optimization for decades, of course, but it remains extraordinarily wasteful — nitrogen runoff producing dead zones in the Gulf, methane and carbon emissions heating the atmosphere, profligate water use leaving water tables depleted, etc.
Plenty uses cameras and sensors to optimize light, temperature, and humidity levels. It is automating the growing process “as much as possible,” Barnard told Business Insider. It even has little robots (“Schleppers”!) that transplant seedlings, because the towers and plants are getting so dense that it’s difficult for a human to operate among them.
Part of what has convinced investors that Plenty has a shot is the radically declining costs of LED lighting. (The efficiency of LEDs puts Plenty on par with conventional agriculture, carbon-wise, at least for some crops, at least when distribution impacts are taken into account; Storey has said he thinks indoor agriculture will be more sustainable in the long haul, especially as the grid gets greener.)
But just as big, possibly bigger, are recent advancements in AI and machine learning. “Utility computing, [internet of things], machine intelligence, wasn’t effective enough five years ago,” Barnard told Fast Company, “much less affordable.” Now, those technologies have reached a level of cost and performance that enables Plenty to fine tune its process. In five more years, computing will be double again as powerful and half again as cheap, enabling yet more automation and optimization.
And that’s great! Mostly.
Plenty is also replacing people with intelligence
I’ve read about 30 articles on Plenty and not one of them has squarely addressed the elephant in the room. To wit: Plenty will succeed insofar as it eliminates food production jobs.
The reason is simple, and found among the aforementioned nine reasons vertical farms fail: “Labor is always your biggest cost.” In the same piece, Matt Liotta of Podponics, a company that tried growing produce in shipping containers and went bankrupt in 2016, is quoted putting it even more bluntly: “People are the problem.”
Industrial agriculture has made ruthless use of scale and automation to minimize labor, but labor remains a huge cost, especially in tasks like harvesting delicate crops like strawberries that can’t easily be mechanized. The same is true for indoor vertical farming: The biggest cost is people.
To compete with industrial ag, vertical farming will have to be even better at reducing the need for human planters and harvesters. In other words, to compete, it’s going to have to create as few jobs as possible.
That’s basically the history of technology — getting more value out of less labor.
The great promise of Plenty is that through automation and optimization, it can make clean, low-input food cost competitive with (morally and chemically) unclean, resource-intensive food. That could potentially save an enormous amount of water and (insofar as it is electrified and powered by renewable energy) radically reduce the carbon emissions of the agricultural sector. Plus it could give millions of people access to fresher, more flavorful, more nutritious fruits and vegetables, making Michelle Obama (and the public health community) happy.
But to do any of that, it has to minimize labor.
Barnard is well aware of that, as is everyone in the industry. “Small-scale growing in 2017 is not a profitable enterprise, and there are a lot of systemic reasons for that that aren’t going to change,” he told Fast Company. “Growing at a small scale, you can’t get to the labor efficiencies that you need. It requires, in essence, too many people.”
“Too many people” is not a great message to communities who might host a farm, though, so Barnard is quick to say that a full-size warehouse like the one planned for Washington will employ as many as several hundred people at skilled, full-time jobs. “While robots can handle some of the harvesting, planting, and logistics,” writes Selina Wang at Bloomberg, “experts will oversee the crop development and grocer relationships on-site.”
Barnard also emphasizes that he’s not competing with traditional agriculture or small-scale urban farming, just adding to the portfolio, seeking to keep up with demand.
But if vertical farming scales as fast as Barnard expects, competing purely on price and efficiency, it will represent a familiar pattern in the US economy — a relatively smaller number of high-skill jobs replacing a relatively larger number of low-skill jobs. In the bigger picture, it is a good thing, to get more and better food for fewer labor and material inputs, but displaced workers tend not to care much about the big picture. And right now agriculture and related industries provide about 11 percent of US employment, according to USDA.
Kevin Drum of Mother Jones recently published a big piece about robots taking all our jobs, thanks to the relentless advance of artificial intelligence. Lots of economists and pundits push back on that kind of thing, citing the lack of movement in productivity statistics.
I have no idea how that will sort out. But I see automation coming for all drivers pretty soon — taxi, bus, garbage truck, delivery van, backhoe, you name it. And now I can see it coming for the agricultural sector. Whether unemployment will spike, as Drum says, or there will just be more and faster churn, there are going to be lots of angry people out of work.
And what are we going to do with all those truck drivers and agricultural workers? I don’t think either party has a good answer. Republicans stomp their feet and insist the jobs will return. Democrats wave their hands at “retraining.”
We better figure it out. We will eventually teach robots to grow our food in giant climate-controlled buildings, optimizing production using AI and machine learning that we can’t yet fathom.
Plenty wasn’t possible five years ago. What will be possible in five more?