Inside the Warehouse Where Lettuce Costs More Than It Should

On the top floor of a former textile mill in Newark, New Jersey, beneath a ceiling hung with 47,000 LED lights arranged in rows so precise they could have been planted by a surveyor, David Chen stands beside a vertical tower of butter lettuce and calculates, again, whether his company will survive the next eighteen months. The lights hum at a frequency just below human hearing. The air smells faintly of nutrients — iron, magnesium, a trace of something chemical he can't quite name. Each tower holds 312 plants. There are 89 towers on this floor. He knows these numbers the way a ship captain knows draft and freeboard.

Chen, who is forty-three and has the slightly bewildered air of a man who has spent a decade explaining the same thing to people who never quite believe him, founded GreenTower Systems in 2017 with $8.5 million from three venture capital firms that no longer return his calls. The pitch was elegant: grow food inside cities, eliminating transportation costs, pesticides, and the vagaries of weather. Use 95 percent less water than field agriculture. Harvest every day of the year. Feed nine billion people on a warming planet.

The reality is a $4.20 head of lettuce that costs $6.80 to produce.

"People think we're growing lettuce," Chen says, leaning against a tower and crossing his arms. "We're not. We're running a real estate operation with an energy bill that would make an aluminium smelter nervous." He pauses. "And we're competing with a farmer in California who owns his land, gets his water subsidised, and has five hundred acres of sun he doesn't pay for."

The Promise That Raised $3 Billion

Vertical farming — the practice of growing crops indoors in stacked layers under artificial light — emerged as a serious commercial proposition around 2013, when LED technology became efficient enough to make photosynthesis economically plausible without sunlight. The sector attracted $2.9 billion in venture capital between 2014 and 2023, according to PitchBook data. AeroFarms raised $238 million. Plenty secured $541 million, including backing from Jeff Bezos and Eric Schmidt. AppHarvest went public via SPAC in 2021 at a $1 billion valuation.

By 2024, AeroFarms had filed for bankruptcy. AppHarvest traded at $0.32, down from a peak of $28. Plenty had closed two facilities. Kalera, a Florida-based operator, filed for Chapter 11 in April 2023 after burning through $250 million. Bowery Farming, once valued at $2.3 billion, laid off 20 percent of its staff in January 2024.

The premise was not absurd. Global food systems generate 34 percent of greenhouse gas emissions, according to the Food and Agriculture Organization. Agriculture consumes 70 percent of freshwater withdrawals. Transportation of produce across continents burns fuel and generates waste; the average head of lettuce in a U.S. supermarket has travelled 1,500 miles. Meanwhile, soil degradation affects 33 percent of the world's arable land, and climate volatility is making traditional farming riskier. A 2021 report by the World Resources Institute estimated that feeding ten billion people by 2050 will require increasing food production by 56 percent while reducing agricultural emissions and land use.

Vertical farming appeared to solve all of it. Grow food where people live. Use hydroponics or aeroponics to eliminate soil. Recirculate water. Control pests without chemicals. Harvest twelve to fifteen crop cycles per year instead of two or three. The first time Chen saw a working vertical farm, in a research lab at Cornell in 2015, he called his wife and said, "I think I just saw the future."

What he did not see was a functioning business model.

The Physics You Cannot Negotiate With

Dr. Cynthia Rosenzweig, a senior research scientist at NASA's Goddard Institute for Space Studies who has studied urban agriculture for thirty years, sits in her office on the Upper West Side of Manhattan and explains the problem with a directness that suggests she has had this conversation many times. "You're fighting the sun," she says. "And the sun is free."

Photosynthesis requires light. In a field, sunlight provides roughly 1,000 watts per square meter on a clear day, at no cost. In a vertical farm, LEDs must replicate that energy, and LEDs require electricity. Even the most efficient LEDs convert only 40 to 50 percent of electrical energy into photosynthetically active radiation; the rest becomes heat, which must be removed by climate control systems that consume more electricity.

Chen's facility in Newark uses approximately 1.2 megawatt-hours of electricity per day. At New Jersey's commercial rate of $0.14 per kWh, that is $168 per day, or $61,320 per year, just for power. Add rent ($42,000 per year for 18,000 square feet), labor (seven full-time employees at $45,000 average salary), nutrients, water, packaging, and transportation to retailers, and the cost per head of lettuce approaches $6.80. Whole Foods pays him $4.20. He loses $2.60 on every unit sold.

"I can cut labor," Chen says. "I can negotiate rent. I cannot negotiate with the laws of thermodynamics."

The Crops That Might Work (And Don't)

Leo Marcelis, a professor of horticultural production systems at Wageningen University in the Netherlands, has spent fifteen years trying to identify which crops might make vertical farming economically viable. He speaks with the weary precision of a man who has repeatedly done the math and arrived at the same disappointing answer.

"Lettuce is actually one of the better candidates," he says over a video call from his office in Wageningen, where a whiteboard behind him is covered with equations relating light intensity to biomass yield. "It has a short growing cycle — twenty-eight to thirty-five days — and high market value relative to weight. If you cannot make lettuce profitable, you will not make tomatoes or cucumbers profitable."

The problem is that most valuable crops are energy-intensive. Strawberries require high light levels and have a long production cycle. Tomatoes need even more light. Staple crops — wheat, rice, maize — are physically impossible to grow profitably indoors; the energy cost exceeds the commodity price by orders of magnitude. A 2022 analysis by researchers at Cornell found that producing one kilogram of wheat in a vertical farm would require $23 worth of electricity, versus a global market price of $0.30 per kilogram.

This leaves vertical farming in a narrow economic niche: expensive greens sold to affluent urban consumers willing to pay a premium for freshness, local provenance, or the vague idea of sustainability. It is not a food security solution. It is a luxury product in a market with abundant cheap alternatives.

"We sold the vision of feeding cities," Marcelis says. "What we built was a business that sells salad to Whole Foods."

The Subsidy the Farmer Doesn't Know He Gets

In California's Salinas Valley, where 70 percent of the United States' lettuce is grown, a farmer named Maria Castillo runs a 240-acre operation that her grandfather started in 1967. She has never met David Chen, but she is his existential problem.

Castillo pays nothing for sunlight. Her water comes from the Central Valley Project, a federal irrigation system that delivers it at $70 per acre-foot — about 15 percent of the unsubsidised cost. Her land, inherited, carries no mortgage. Her operating costs are labor ($2.8 million annually for forty seasonal workers), diesel for tractors and harvesters ($47,000), and fertiliser ($83,000). She produces 4.2 million heads of lettuce per year. Her cost per head is approximately $0.74. She sells to distributors at $1.20, netting $0.46 per unit.

"I don't understand what problem they're solving," she says when told about vertical farms. She is standing beside an irrigator at 6:30 on a May morning; the sun is already warm. "We grow food. It's cheap. People can afford it. Why would you make it more expensive?"

This is the uncomfortable truth that vertical farming advocates rarely address: conventional agriculture works extraordinarily well at producing cheap calories. The global food system, for all its environmental and social costs, has driven food prices to historic lows. The average American household spends 8.6 percent of income on food, down from 23 percent in 1950. A head of iceberg lettuce at Walmart costs $1.68.

Vertical farms are trying to compete in a market shaped by a century of subsidies, mechanisation, and economic externalisation. Castillo does not pay for the carbon her diesel emits, the nitrogen runoff that pollutes the Salinas River, or the aquifer depletion that will make her farm unviable in thirty years. Those costs are borne by the climate, the ecosystem, and the future. The market price of lettuce reflects none of them.

The Technology That Might Change the Equation (Someday)

At the University of Arizona's Controlled Environment Agriculture Center in Tucson, Dr. Murat Kacira is testing LEDs that convert 65 percent of electrical energy into photosynthetically useful light, up from 42 percent in commercial systems. In his lab, experimental fixtures hang above trays of basil; spectrometers measure light wavelengths to the nanometer.

"If we reach 70 percent efficiency, and electricity costs drop to $0.08 per kWh with renewables, the unit economics start to work," Kacira says. He is cautious, which suggests he has seen promising technologies fail before. "That's a lot of ifs."

Other researchers are pursuing gene-edited crops optimised for indoor conditions. In 2024, a team at the John Innes Centre in the United Kingdom published results in Nature Plants describing a modified lettuce variety with 30 percent higher photosynthetic efficiency under LED light. The plant carries edits to three genes involved in chlorophyll synthesis and light absorption. In controlled trials, it reached harvest maturity in twenty-two days instead of thirty, cutting energy costs by 27 percent.

The seeds are not commercially available. The regulatory pathway in the United States requires field trials, environmental impact assessments, and USDA approval — a process that typically takes seven to twelve years. In the European Union, gene-edited crops face even stricter rules and public resistance. No vertical farm is using them.

Meanwhile, automation offers marginal savings. Robots can transplant seedlings and harvest mature plants, reducing labor costs by 40 to 60 percent, according to a 2023 report by the American Society of Agricultural and Biological Engineers. But labor is only 15 to 20 percent of operating costs in a typical vertical farm. Cutting it in half saves $0.80 per head of lettuce. The energy bill remains.

The Exit That Never Came

In Newark, David Chen has laid off three employees since January 2025. He is negotiating with his landlord for a rent reduction he knows will not come. His largest investor, a venture firm in Menlo Park, stopped responding to emails in November. He has enough cash to operate through October 2026.

"I believed the pitch," he says. It is late afternoon; the LEDs have dimmed to simulate dusk, part of a lighting protocol designed to optimise chlorophyll production. The towers cast long shadows across the concrete floor. "I still believe the pitch. We will need this. Just not yet. Maybe not for twenty years."

The history of agricultural technology is littered with ideas that were correct but premature. Mechanical reapers existed in concept for decades before Cyrus McCormick made them economically viable in the 1830s. Drip irrigation was invented in 1866 but not widely adopted until the 1970s, when plastics became cheap. Vertical farming may be such a technology — brilliant in theory, waiting for the economic and technological conditions that make it necessary and possible.

Or it may be something else: a Silicon Valley fantasy that confused technological sophistication with economic logic, that assumed innovation could overcome thermodynamics, that mistook a niche product for a revolution.

Chen walks to the window. Outside, the sun is setting over the Newark skyline, painting the sky in shades of orange and purple — light, free and inexhaustible, that no LED will ever replicate at a price anyone can afford. He watches it for a moment, then turns back to the towers, the lights, the silent hum of machines doing what nature does for nothing.

"The sun is free," he says quietly. "That's the whole problem."