Harvest

Harvest is where a season’s work converts into either income or loss. It is the most time-pressured phase in agriculture — windows measured in days, weather that doesn’t negotiate, equipment that cannot fail. Globally, approximately 14% of all food produced is lost between harvest and the retail level — before a consumer ever sees it (FAO, The State of Food and Agriculture 2019). In the American specialty crop sector, the losses at the farm and packing stages alone run far higher.

What It Is

Harvest is the point where biological production becomes economic reality. For commodity row crops — corn, soybeans, wheat — highly mechanized combines cover enormous acreage efficiently, and farm-level losses are typically low (1–5%). For specialty crops — leafy greens, berries, melons, tree fruit — the harvest picture is dramatically different: losses of 20–40% at the farm and pack-shed level are common (USDA ERS), and the causes are as much economic and political as they are biological.

How It Works

Commodity harvest is a triumph of mechanical engineering. A modern combine harvests, threshes, and separates grain continuously, moving at speeds up to 6 mph across thousands of acres. The limiting factor is not the machine’s capability but the weather: too much rain softens the ground and prevents the equipment from entering fields. A $750,000 machine that gets stuck in mud is not a machine that’s harvesting.

Specialty crop harvest depends almost entirely on human hands. Strawberries, lettuce, asparagus, and most tree fruit cannot be harvested mechanically without damage. The labor for this work is supplied almost entirely through the H-2A temporary agricultural worker visa program — a federally managed system bringing workers primarily from Mexico and Central America for the harvest season. Visa processing delays of even two weeks can mean crops over-ripen and rot before workers arrive.

Three non-weather factors drive the high loss rates in specialty crops:

Economic abandonment occurs when a region-wide bumper crop collapses the wholesale price below the cost of harvesting, packing, and shipping. If a flat of strawberries is worth less at the packing shed than it costs to get it there, the farmer loses less money by leaving the crop in the field. This is rational; it is also an enormous waste.

Cosmetic standards set by major grocery chains specify size, shape, and color ranges for every commodity. A perfectly nutritious but slightly crooked carrot or an apple with a minor blemish has no major-retail buyer. Secondary markets (canning, juicing) pay pennies on the dollar, often not enough to cover freight.

Equipment DRM is an emerging but serious factor. Combine software locks that require dealer authorization to clear fault codes mean that a software glitch during peak harvest can delay fieldwork for days while a dealer technician is dispatched — time that a deteriorating crop cannot afford to wait.

Why It Matters

Harvest loss is the most visible waste in the food system — food grown, inputs spent, water used, and labor deployed for a product that never feeds anyone. For specialty crops, the waste represents not just food but the H-2A labor that planted and tended it, the irrigation water drawn from stressed aquifers, and the logistics capacity that will never be needed.

Crop insurance, the primary financial safety net for farmers, was designed to cover yield losses from weather and disease. It does not cover the “Bountiful Harvest Paradox” — the scenario where growing the crop successfully is what destroys the margin. Economic abandonment of perfect crops is an uninsured loss, and it repeats whenever growing conditions are uniformly excellent.

What’s Being Done

The harvest loss problem is large and well-documented — but it is not standing still. Across right-to-repair law, labor reform, robotic harvesting, and surplus food recovery, real progress is underway. The challenge now is scaling what works and filling the gaps that no existing program yet addresses.

Efforts Showing Results

John Deere $99M Settlement and 10-Year Diagnostic Access (April 2026). In April 2026, John Deere agreed to a $99 million class-action settlement covering farmers who paid Deere or authorized dealers for equipment repairs from January 2018 onward. The settlement’s binding 10-year requirement to provide all digital diagnostic tools to farmers and independent repair providers is the most enforceable repair-access commitment ever extracted from a major agricultural equipment manufacturer. Unlike the voluntary 2023 MOU with the American Farm Bureau Federation, this is a court-enforceable obligation — and the separate FTC antitrust lawsuit, whose motion to dismiss was denied in June 2025, continues to create sustained regulatory pressure. The FARM Act (H.R.5857), introduced in October 2025, would codify these access rights in federal statute across all manufacturers. The priority now is extending Deere’s commitments to AGCO, CNH Industrial, and Kubota before another harvest season passes. (Arnold & Porter, April 2026; Congress.gov H.R.5857)

Harvest CROO Robotics — Commercially Viable Strawberry Harvesting (April 2025). Harvest CROO, backed by approximately two-thirds of the US strawberry industry as investors, announced in April 2025 that its automated strawberry harvesting system achieved performance on par with human pickers in commercial field trials. The machine uses 16 independently operating robots powered by NVIDIA vision chips and 13 patents. Critically, the business model is a service-level agreement — growers pay per volume harvested rather than purchasing the machine outright — removing the capital barrier that has historically limited adoption by mid-size farms. “Commercially viable in trials” and “deployed at scale” are still separated by years of manufacturing ramp-up, but the proof of concept is real, and the financing model is designed for broad adoption. Multi-crop extension — asparagus, stone fruits, wine grapes — remains years away and needs USDA specialty crop research investment to accelerate. (harvestcroorobotics.com)

National Gleaning and Farm to Food Bank Project. USDA’s Farm to Food Bank Project connects Feeding America food banks with farmers to rescue food that doesn’t meet retail standards. Food Forward recovered 94 million pounds of produce in 2024 — their largest-ever annual haul — and similar organizations operate in dozens of states. These programs work: they move real food to real people who need it. The structural limitation is that gleaning recovers less than 2% of the estimated 34 billion pounds of edible produce left in American fields annually, because gleaning at scale requires volunteer coordination, transportation, and liability management that the current grant structure cannot fund adequately. The model is proven; the funding gap is what constrains it. (fns.usda.gov; foodforward.org)

Imperfect Produce Retail Expansion. The merged Misfits Market and Imperfect Foods operation has rescued nearly 500 million pounds of surplus and cosmetically imperfect produce and operates across 44 states. The direct-to-consumer subscription model has faced financial pressure and layoffs, but grocery inflation is creating a more durable channel: Kroger’s “Peculiar Picks” program and similar retailer in-store imperfect produce sections reach far more consumers than subscriptions ever could, and 86% of retailers were donating surplus food as of 2025. Inflation has accomplished what years of advocacy could not — making imperfect produce economically attractive to mainstream retail buyers. (grocerydive.com)

Where More Work Is Needed

Economic Abandonment Has No Safety Net. When market prices fall below the cost to harvest, pack, and ship a crop, farmers make a rational decision to plow it under — and no existing federal program intervenes at that moment. Tariff-driven disruptions in 2025–2026 collapsed Florida tomato prices from $16 to $3–4 per box against an $11 break-even, and Maine wild blueberry losses exceeded $28 million with acreage left unharvested. Crop insurance covers yield loss from weather and disease, not price collapse. Gleaning organizations cannot mobilize fast enough when an entire region decides simultaneously to abandon its fields. A targeted USDA “cost-of-harvest” safety net — paying specialty crop farmers the marginal harvest cost when market prices fall below it, conditional on the food going to food banks or USDA programs — would directly address this gap. Section 32 rapid-procurement authority could be adapted for this purpose without new legislation.

Multi-Crop Robotic Harvesting Is Years Away for Most Specialty Crops. Strawberry harvesting is approaching commercial viability, but asparagus, stone fruits (peaches, cherries, plums), wine grapes, and most other high-value specialty crops present physical challenges — irregular canopy structure, variable ripeness, fruit fragility — that current robotic platforms cannot yet address. Private investment concentrates on the highest-volume crops, leaving mid-tier specialty crops without a robotic solution on any near-term horizon. USDA NIFA and NSF co-funding for modular, multi-crop harvesting platforms and university-industry consortia focused specifically on asparagus, stone fruit, and wine grape harvesting would compress development timelines and spread R&D costs. The technology window is real but requires sustained public investment to close before labor shortages force further specialty crop abandonment.

Cosmetic Grading Standards Remain Unreformed. More than 20 billion pounds of cosmetically imperfect but food-safe produce is wasted annually in the US because USDA voluntary grading standards contain cosmetic specifications — size, shape, color uniformity — that functionally exclude edible food from retail. These standards were designed for uniform packing and long-distance transport, not nutrition or safety. No federal revision has occurred, and coordinated retailer reform has stalled repeatedly since early pilots in 2016–2019. A USDA revision separating cosmetic specifications from food safety requirements — combined with federal procurement leverage requiring school lunch programs and SNAP-authorized retailers to carry a minimum percentage of grade-B produce — could create institutional demand that sustains the supply chain without depending on consumer re-education alone.

Fourteen percent of all food produced globally is lost between farm and retail — and in the American specialty crop system, that number climbs to 20–40% for the fruits and vegetables people most need to eat. The causes are not mysteries: a tomato field gets plowed under not because the tomatoes are inedible, but because the price to pick them fell below what it costs to get them to market. A harvester sits broken during a three-day window not because the repair is complicated, but because the software is locked to a dealer who cannot arrive in time. Strawberries rot on the vine not because no one wants them, but because the workers who could pick them are waiting on paperwork. These are engineering, legal, and policy problems — and in 2024–2026, serious progress is being made on all three fronts simultaneously. The food is there. The need is there. The distance between them is smaller than it has ever been.