Feeding the World Through 2036: Farmers Grow It. The System Delivers It.
I serve on the Grain Elevator and Processing Society’s Education and Training Committee, and we’ve been looking at our training library and what we might need to add or change for the membership down the road. That got me thinking past the training program itself, about where the whole grain industry needs to head globally.
I kicked some of those ideas around with a few good friends, my old companions Jack, Jim, and Jose, and later asked new friends Claude and Grok to help dig through some of the data. This post is what came out of that.
The purpose of this post is to stimulate thought, discussion, planning, and action. I don’t have a forecast to sell. If it gets a few people at the coffee pot arguing about where the next bottleneck will show up, it did its job.
Who’s going to feed the world over the next ten years?
Start with the farmer. No crop, nothing to dry, store, process, or ship. But growing the crop is only the start of the job.
Anybody who’s run an elevator knows the two questions that follow.
If a producing area harvests another million tons next year, where does that grain go? Who receives it, dries it, stores it, and moves it? Which part of the system runs out of room first?
Then flip it. If that area comes up short, the river drops, or the usual shipping route gets cut off, where does the customer find replacement grain, and what does it cost?
That’s the practical version of the next ten years. We need farms that can produce, and we need a system that turns those harvests into a dependable supply.
That means the right seed, getting the fieldwork done on time, decent storage, transportation, processing, and financing. It means people who can run and fix the equipment. And it means customers who can pay for what shows up.
I’m mostly talking corn, soybeans, wheat, and rice here. Those four cover a big share of the commercial grain system of food, feed, processing, and trade, even though they don’t represent the whole food supply.
The real question is whether production and the system behind it can move forward together.
Where we stand
Production and demand don’t line up neatly on the map.
Some big producing countries eat most of what they grow. Others produce way more than they need at home. Trade fills the gap, but domestic production still matters a lot.
Here’s a simple continental snapshot for roughly 2023 (2022 for freshwater). The production and export numbers cover only the combined weight of corn, soybeans, wheat, and rice, not all food. Rice is converted to an approximate milled equivalent so the crops can be added together. Exports are gross country reports, including trade inside continents. Freshwater means renewable freshwater resources, not water use.
Continent | World population | World arable land | World renewable freshwater | World fertilizer used | Four-crop production | Four-crop exports |
Asia | 59.0% | 34.4% | 25.8% | 56.4% | 43.2% | 9.8% |
Africa | 18.3% | 19.0% | 9.2% | 3.7% | 5.5% | 1.3% |
Europe | 9.2% | 19.6% | 15.4% | 10.7% | 13.8% | 27.9% |
North America | 7.5% | 16.0% | 15.9% | 14.4% | 22.2% | 23.6% |
South America | 5.4% | 8.7% | 29.7% | 12.8% | 14.4% | 32.7% |
Oceania | 0.6% | 2.3% | 4.0% | 2.1% | 1.0% | 4.7% |
Figures are approximate and compiled from the sources listed below. Reference years, methods, and regional groupings vary by dataset, so read the table for broad regional relationships.
Asia has about 59 percent of the people and a big share of production, but a much smaller share of the exports of the four crops. South America, Europe, and North America account for a lot of the exportable surplus even though they have far fewer people. Africa’s share of the people is close to its share of arable land, but its fertilizer use and production are way lower. Treat these as starting points. A continent’s water and land aren’t automatically sitting where the crops are grown.
The OECD–FAO Agricultural Outlook 2025–2034 has world cereal production growing about 1.1 percent a year, reaching roughly 3.2 billion metric tons by 2034. That’s a useful direction for looking ten years out. It’s a baseline, and real harvests will bounce around it.
Most of the extra production is expected to come from higher yields, not more acres. Yield growth is projected around 0.9 percent a year; harvested area growth is only about 0.14 percent.
That puts a lot of weight on how the existing acres are farmed. It also puts weight on the businesses that have to receive the extra crop.
More annual production doesn’t show up as a steady daily trickle. Harvest dumps grain into a short window. The operating questions are about peak arrivals, moisture, storage, and how long you have to move inventory before the next crop needs the same space.
A forecast in annual tons eventually has to become a plan in tons per hour, storage days, and available transportation.
Where the additional demand will come from
Population matters, but headcount alone doesn’t explain grain demand.
Food, livestock feeding, processing, and industrial uses all pull on production. OECD–FAO figures that by 2034 about 40 percent of cereal use will be food, 33 percent feed, and 27 percent biofuel and other uses. That last category is more than just fuel.
Asia accounts for about 54 percent of the projected increase in cereal demand. Regions with the fastest population growth aren’t always the ones with the biggest increase in tons.
Africa’s rice situation is another piece. OECD–FAO projects African rice imports up 53 percent by 2034 from its base period. Meeting that takes domestic farming plus international buying, vessel discharge, storage, milling, and inland distribution.
China shows why you have to stay crop-specific. The same outlook has its corn imports falling about 20 percent as domestic production rises. A big consuming country doesn’t move the same direction on every commodity.
Don’t treat any region as permanently self-sufficient or permanently dependent.
Extra domestic production where demand is growing still adds to world supply even if the grain never leaves the country. A ton delivered to a local mill helps that customer the same as a ton unloaded from a ship.
So “who feeds the world” includes the established exporters, the domestic producers serving their own markets, and the people who make both of those supplies reliable.
Who can supply the additional grain?
No single country carries the whole load, and no single crop tells the whole story.
On corn, the United States, Brazil, Argentina, and Ukraine are expected to account for 91 percent of the growth in world corn exports through 2034. That’s their share of the increase, not of all corn exports. It still shows how much of the additional trade rides on a relatively small group of suppliers. OECD–FAO also expects some of the largest absolute gains in corn output in the United States, Brazil, and China. China’s presence is a reminder that extra production doesn’t have to become exports to matter.
Whole soybean exports are even more concentrated. Brazil and the United States dominate the trade, with Brazil the larger supplier in most recent years. Argentina is smaller on whole beans but is a major factor in the soybean complex through meal and oil. China is the main destination for beans. The Brazil section below gets into the same constraints: storage location, inland movement, and port access decide how much of a larger crop can actually move.
Wheat is more spread out. Russia has been the largest exporter in recent years, with the European Union, Canada, Australia, Ukraine, and the United States also supplying meaningful volumes. Weather and policy swings in any of those places can change the export balance quickly.
Rice is the most concentrated and the most policy-sensitive of the four. Only about 10 percent of production is typically traded. India accounts for a large share of that trade (often around 40 percent in recent years), followed by Thailand, Vietnam, and Pakistan (especially important in basmati). One exporter’s policy decision can move the whole market.
For planning across these crops, I’d separate three contributions.
First is additional exportable production. Growing a bigger crop isn’t enough. Domestic use, quality requirements, inventories, and delivery costs decide what can actually be offered overseas.
Second is stronger domestic production in consuming markets. That can cut import needs, improve local availability, or give an alternative when international prices jump.
Third is better preservation of what’s already been harvested. Grain saved from avoidable deterioration is usable supply without another acre.
Look at those together. Export growth matters, but it shouldn’t be the only scorecard for agricultural progress.
On the farm: turning potential into usable output
If most of the production growth is supposed to come from yield, the farm job is big.
The useful question isn’t “which technology posts the highest number.” It’s “what combination can this farmer get, afford, and use consistently.”
A variety has to fit the growing season, the stresses that usually show up, the harvest system, and the market. Yield potential matters, but so do maturity, standability, disease resistance, grain quality, and how it holds up when conditions aren’t ideal. The farmer’s final test is simple: Is suitable seed available before planting? Is it dependable? Can the crop support the cost? A seed program that misses the planting window hasn’t delivered, no matter how good the genetics look on paper.
Mechanization is about getting the work done on time, inside the available window. Owning, custom hire, cooperatives, and leasing all have a place. Every machinery proposal should answer the practical questions: Who runs it? Who fixes it? Where do parts come from? Is fuel available? Can the business carry the equipment between seasons? A machine isn’t productive capacity just because it got delivered.
Technology needs a clear job. Soil testing, irrigation scheduling, weather information, and remote sensing are all worth looking at, and so are simpler tools that help somebody act at the right time. Automation and artificial intelligence should face the same test as any other investment: What job will they improve, what evidence supports the result, and who stays responsible? The comparison should include cost per usable ton, timeliness, quality, and reliability.
Then follow the improvement downstream. If a new practice raises output or packs harvest into fewer days, what changes at the receiving pit, dryer, storage site, and transport connection? The production plan isn’t finished until those questions have answers.
Brazil: the crop must have somewhere to go
Brazil is a useful example because it puts production and delivery in the same conversation.
IBGE’s survey for the first half of 2025 put Brazilian storage capacity at about 231 million metric tons. Conab put the 2024/25 soybean and corn crops alone at a little over 310 million metric tons, both records. That shows the scale of the system, but it doesn’t give you a simple national number for how many more bins are needed. Annual production and static storage measure different things. Crops arrive at different times, inventory moves during the year, and some space turns more than once. A national total also doesn’t tell you whether storage is available in the district where the crop is piling up.
Picture a load of soybeans leaving a farm near Sorriso in Mato Grosso, deep in the interior and a long way from salt water.
If it heads north, it rides a truck up BR-163 to the barge terminals at Miritituba on the Tapajós River, goes by barge down to the export terminals at Santarém or Barcarena, and gets loaded on a vessel. If it heads south, it goes by truck to the rail terminal at Rondonópolis and rides the train to Santos. Either way, it has to reach a receiving facility, get conditioned where needed, go into storage or keep moving, and connect with the next mode until it reaches a ship.
At each point, ask what limits the movement.
Truck availability? Road access? Unloading? Drying? Storage? Rail or barge dispatch? Terminal inventory? Berth availability?
A bigger export terminal can remove one constraint and expose another. Faster unloading inland can fill storage faster without increasing outbound movement. More trucks can just lengthen a queue.
Brazil is better looked at as a set of developing corridors than as one national storage subtraction. Northern and northeastern ports, the Arco Norte, have taken a growing share of soybean exports, so the routes keep shifting. For each corridor, test normal movement, peak harvest arrivals, and an interruption. Figure out what can be held safely, what can be rerouted, and what has no practical alternative.
The infrastructure assignment
Follow a load from the field to the processor and the infrastructure needs get clearer.
Receiving facilities need capacity for how trucks actually show up at harvest, and the daily average won’t tell you that. Dryers need to handle a wet harvest and have dependable fuel and power. Storage needs to protect quality for the holding period.
The road into the facility belongs in the same plan. A facility can’t do its job if the approach road, bridge, or truck circulation pattern chokes deliveries.
Inland transportation has to connect those facilities with processing or export markets. That takes equipment, dispatching, transfer capacity, maintenance, and a plan for seasonal restrictions. A line on a map doesn’t move any grain.
At the export terminal, inland arrivals have to match up with storage, segregation, reclaim, loading, and vessel schedules.
Two distinctions matter a lot.
Static capacity is not throughput. Storage capacity tells you how much a facility can hold. Throughput tells you how much it can handle over a period. A site that fills once and holds grain for months has a different job from a terminal that receives and ships continuously. Both can be necessary, but they aren’t doing the same work.
Rated capacity is not dependable capacity. The planning question is what the system can deliver under real operating conditions. What happens when a dryer is down? When power is interrupted? When trucks arrive unevenly? When a transfer point is unavailable for several days? Plan around the capacity you can count on during an interruption. The nameplate is the best case.
Not every answer is new construction. Sometimes the best improvement is maintenance, better dispatching, a different schedule, or more reliability in the equipment you already have.
Identify the constraint first. Then choose the investment.
Protecting the harvest is part of production planning
Food loss and food waste are two different problems.
FAO estimates about 8.4 percent of cereals and pulses were lost globally between production and retail in 2023. For the grain business, the immediate concern is what happens between harvest and delivery: condition, insects, rodents, handling damage, spillage, and delays that go past the safe holding period.
Loss numbers vary depending on the crop, the region, and how they’re measured. That’s a reason to go look at the problem on the ground, and no excuse to ignore it.
Start local. Where is the grain being lost? What’s causing the loss? What does prevention cost? How will you measure the result?
The answer might be better drying, sound structures, suitable hermetic storage, pest management, improved handling, or a shorter holding period. The right choice depends on the crop, the scale, the environment, and the route to market.
Protecting grain already grown can be an economical source of additional usable supply. Treat it with the same discipline as a project to raise yield. Measure the grain that reaches the customer in usable condition, not just the grain that leaves the field.
The receiving country is part of the system
The food supply story doesn’t end when the export vessel sails.
The receiving country still has to unload the cargo, store it safely, move it to processing, and distribute the resulting food or feed.
Trade reporting on Egyptian and other African wheat markets keeps coming back to execution, finance, and inland logistics along with access to grain. Terminal performance varies, so a discharge rate from one facility shouldn’t be treated as a national average.
The operating questions are familiar.
How quickly can the ship discharge? Is storage available? Can the mill take the grain? Are trucks or rail equipment ready? Are customs and payment arrangements done?
A fast unloader can’t fix a full silo or a shortage of inland transport.
I saw this firsthand on a project for the U.S. Soybean Export Council in Guatemala. Vessels were being delayed at Puerto Quetzal for up to 90 days with a $25,000/day demurrage bill. The grain was there and the buyers were there. What held things up was infrastructure, both portside and inland. Every one of those days cost somebody money, and that cost ended up in the price of feed and food in Guatemala.
Delay has a cost. Say delay adds $500,000 (20 days) to a cargo of 40,000 tons. That’s $12.50 a ton, and somebody pays it. Run that clock out to 90 days and you can see why execution belongs in any discussion of affordability.
Port capacity, processing demand, inventory policy, and inland distribution need to be planned together.
The exporting end and the importing end are parts of the same delivery system. Improving one while neglecting the other leaves the customer exposed.
Plan for the year that does not follow the average
A ten-year outlook helps set direction. It doesn’t replace a plan for the bad year.
For every major corridor, test at least three conditions: a normal harvest, the peak of the harvest rush, and a significant interruption.
Then ask what can be stored, what can be rerouted, how much replacement supply is available, and who can afford the change.
I lived through one of those years. In the 1988 low water, hundreds of loaded ADM corn barges sat trapped on the Illinois River for weeks. Crews opened covers, ran portable aerators, and applied propionic acid to save the corn. That’s what a backup plan looks like when it has to be invented on the spot. It’s a lot cheaper to work it out ahead of time.
Weather belongs in this discussion as an operating condition.
A regional look should cover heat during the growing season, rainfall timing, harvest moisture, drainage, and river conditions where they matter. Before acting on any claim about changing conditions, pin down where, over what period, compared with what, and what the evidence shows. Then turn it into decisions. Does drying capacity need another look? Is drainage adequate? Does a transport route need a contingency? Are planting and harvest assumptions still suitable?
Inputs deserve the same attention as grain.
Reporting on Middle Eastern fertilizer disruption in 2026 shows why supply planning has to include fertilizer and fuel routes. Sharply reduced regional fertilizer shipments amid disruption around Hormuz are a reminder to identify dependencies on a single supplier, route, energy source, or critical input before that dependency becomes an emergency.
A backup option also has to be usable. An alternative route that can’t get equipment, can’t take the volume, or can’t operate at an affordable cost doesn’t give much protection.
Who builds it, and who pays?
Farmers, cooperatives, merchants, processors, transport companies, governments, and lenders all have roles. Whoever builds it, ask what constraint the project removes, who owns it, who maintains it, and where it really stands. Announced, financed, under construction, and operating are four different things, and only the last one can handle the next shipment.
For every capital proposal, ask for an operating plan alongside the construction plan: utilization, maintenance, staffing, working capital to carry inventory, and a recovery plan for an interruption. Explain how the facility stays useful after the ribbon-cutting.
The ten-year assignment
The practical job is to connect production targets with delivery plans.
Then keep asking the two questions we started with.
If another million tons is harvested, where does it go?
If the expected crop or usual route is unavailable, what is the replacement plan?
The next ten years can’t be planned only in acres and yield. They also have to be planned in the ability to protect the harvest, move it, and deliver it at a price the customer can pay.
The plan has to work when the crop is large, when it is short, and when the usual route can’t be used.
That brings me back to where this started, our GEAPS training library. If these are the questions the next ten years will ask, our people need to be ready to answer them. A few topics I’d want covered: throughput versus storage capacity and how to measure both; dryer performance and reliability in a wet harvest; contingency planning for a corridor when the river, rail, or port isn’t available; measuring loss from receiving to load out; and building an operating plan, with staffing, maintenance, and working capital, to go alongside a construction plan.
Thank you for reading. Like I said up top, the point of this one is to get people thinking, talking, planning, and acting. Feedback from elevators, processors, terminals, and anyone who participates in or watches how grain gets to market helps shape the value of GG50. Your experiences and opinions will definitely add more value than Jack, Jim, and Jose!
Regards,
Grain Guy Fifty
Sources
1. OECD/FAO, OECD-FAO Agricultural Outlook 2025-2034, Cereals chapter
2. IBGE, Pesquisa de Estoques (storage survey), first half of 2025
3. Conab, Acompanhamento da Safra Brasileira de Grãos, 2024/25 crop, 12th survey (September 2025)
4. FAO, food loss estimates for 2023 (SDG 12.3.1)
5. World Bank, review of post-harvest maize loss measurement
6. Trade reporting on Egyptian and other African wheat import logistics
7. Reporting on Middle Eastern fertilizer shipments, September 2026 (Kpler data via The National)
8. UN/OWID population data; FAO/World Bank arable land, freshwater, and fertilizer data; FAOSTAT production and trade series (continental table calculations)
9. Jim Voigt, Feeding the World Through 2036 research outline and data pack (author’s working materials)
Comments