Crop Root Zone™

The single most consequential price event in the global nitrogen market this quarter has a deadline, and the deadline is today. India's state-owned Rashtriya Chemicals and Fertilizers Limited issued an international tender on July 29 for 1.7 million tonnes of bulk urea, opened technical and commercial bids on August 11, and required every supplier to hold its offer valid through August 24 — with all shipments completed by September 24 (Profercy, Aug 7, 2026; Fertilizer Daily, Aug 7, 2026). On the west-coast tranche, a 1 million-tonne requirement, RCF received roughly 3.1 million tonnes of offers priced between $393.65 and $435 per tonne (Bloomberg, Aug 13, 2026).

Almost every account of this tender leads with the low offer. That is the least informative number in it. What the tender actually establishes is the shape of the world's exportable urea supply in late August — how much of it exists, how badly it wants a home, and how far apart the sellers are on what it is worth. All three of those live in the numbers that are not the headline.

1. What the tender is, and what today's date means

RCF's requirement splits 1 million tonnes to India's west coast and 700,000 tonnes to the east (Fertilizer Daily, Aug 7, 2026). Bids opened August 11. Offer validity runs to August 24 — today. Shipment must be complete by September 24.

An offer-validity window is not administrative housekeeping. It is a free option the seller writes to the buyer: for thirteen days, RCF may accept at the bid price no matter what the market does, and the supplier may not withdraw. Sellers price that option into the number. The longer the validity, the more a seller must charge to cover the risk that the market runs away from it in the interim — which means a thirteen-day validity in a market that has fallen 12% in ten weeks is a genuinely different product from the same tonne sold spot.

Source: Profercy, Aug 7, 2026; Fertilizer Daily, Aug 7, 2026.

The compression at the end is the part worth sitting with. If the full 1.7 million tonnes is awarded at or near the validity expiry, the entire volume has to be loaded and sailed inside roughly thirty days. That is about 56,700 tonnes per day of loadings, sustained, across every origin serving India — a Crop Root Zone calculation on the tender's own stated volume and deadline. At a 35,000-tonne handysize parcel, the number of cargoes implied is on the order of 49 vessels in thirty days, or about one and a half sailings a day.

2. The offers: a 10.5% spread for identical product

The west-coast tranche is the one with published numbers, and they are unusual.

Source: offers and price range, Bloomberg, Aug 13, 2026; tender volume, Fertilizer Daily, Aug 7, 2026. Ratios and spreads are Crop Root Zone calculations on those figures.

Two facts sit inside that table and they pull in opposite directions.

The first is the oversubscription. Three times the requested tonnage was put in front of one buyer for one delivery window. A supply that is genuinely scarce does not queue up at 3.1×.

The second is the dispersion. Urea is urea: 46-0-0, prilled or granular, fungible to a degree that few agricultural inputs are. When a single buyer receives offers on a single day for a single delivery window and the top offer is 10.5% above the bottom, that gap cannot be about product quality. It is about freight and origin — the distance and voyage economics from each producing region to the Indian west coast — and about how urgently each individual seller needs to move a cargo.

$41.35/t

The gap between the highest and lowest offer in one tender, for the same commodity, to the same buyer, on the same day — 10.5% of the low offer. (Bloomberg, Aug 13, 2026; spread is a Crop Root Zone calculation.)

Ground Truth: The low offer tells you what the most desperate seller with the shortest voyage will take. The spread tells you how many sellers there are. A tightening market compresses tender dispersion, because when cargoes are scarce every seller can hold out for the top of the range and the offers bunch near the ceiling. A loosening one widens it, because sellers who cannot compete on freight have to buy their way in on price. A 10.5% spread against 3.1× oversubscription is one message stated twice: there is more urea looking for a buyer than there are buyers. For a grower deciding whether to prepay fall nitrogen, that is a more durable signal than the headline $393.65 — a single low print can be one distressed cargo, but a wide spread across a 3.1-million-tonne offer book cannot be.

3. The ladder: where $393.65 sits in 2026

The August offers are not a level, they are a point on a very steep line. India has been the world's price-setting urea buyer through a year in which the price it paid moved by a factor of two and a half.

Source: April and June levels as reported by Bloomberg, Apr 15 and Aug 13, 2026; August offers, Bloomberg, Aug 13, 2026. Percentage changes are Crop Root Zone calculations. The April figure is described in the source as "near $1,000" and is treated here as approximate.

The April spike was a war-risk premium — offers near $1,000 a tonne as Hormuz-related supply constraints bit (Bloomberg, Apr 15, 2026). The August print is what the same market looks like once that constraint eases. The percentage arithmetic is worth stating plainly because it is easy to under-feel: the low offer in August is 60.6% below the April level. Very few industrial inputs of any kind move that far in four months and then keep functioning normally.

Bars scaled so the April level is ten blocks. Source as above.

Note what the bar column makes visible that the percentages hide: June and August are nearly the same bar. The overwhelming majority of the 2026 decline happened between April and June. August's move is a 12.3% continuation, not a second collapse. A market that fell 55% and then fell 12% more is a market that has already done most of its correcting — and that matters for anyone reading the August number as the start of a trend rather than the tail of one.

4. What the tender does not settle

Three things are genuinely unresolved, and it is worth being precise about which.

The award is not the offer. Everything above describes what was bid, not what RCF accepted. A buyer facing 3.1× coverage does not have to take the whole book, and Indian tender practice routinely awards less than the full requirement when the price ladder rises steeply above the low. The tonnage that actually clears — and therefore the tonnage that actually leaves the export market — is not established by the offer range.

The 3.1-million-tonne figure is west-coast basis. It is reported against the 1 million-tonne west-coast tranche, not against the full 1.7 million tonnes (Bloomberg, Aug 13, 2026). The oversubscription ratio computed above is therefore a west-coast ratio. The east-coast tranche's coverage was not separately reported in the public sources reviewed for this piece, and no combined ratio should be inferred from these numbers.

Freight is inside the price and cannot be separated out. These are CFR offers — cost and freight to an Indian port. A $393.65 offer from a short-voyage origin and a $435 offer from a long-voyage one can represent the same netback to the producer. Nothing in the published range permits a reader to say which sellers are cutting price and which are simply farther away. That is a real limit on how much the dispersion can be made to carry, and it cuts against the reading in §2 rather than for it. What survives the objection is the volume: 3.1× coverage is a quantity fact, and freight geometry does not change it.

Ground Truth: The most actionable line in this tender is the shipment deadline, not the price. Every tonne awarded must sail by September 24, which places the physical loading of up to 1.7 million tonnes squarely inside the four weeks before the Northern Hemisphere harvest freight peak begins. Export urea and export grain do not compete for the same vessels in most trades, but they do compete for port labour, berth time and inland positioning at several shared origins. A grower or retailer waiting for a further leg down in landed nitrogen cost should watch the September loading schedule rather than the September price screen: if the tonnage moves cleanly, the market absorbed 1.7 million tonnes without a freight event and the softness is structural. If it does not, the next quote will carry a logistics premium that has nothing to do with the nitrogen.

5. What a US buyer should take from an Indian tender

Nothing in this tender prices a tonne of urea in Illinois. The mechanism is indirect and it runs one way: India is the residual buyer of the world's exportable surplus, so the level at which India clears sets the floor under what an exporter will accept from anyone else. A producer who can sell to India at $393.65 CFR has no reason to sell to a Gulf trader at less than the equivalent netback.

That is a floor argument, and floors travel slowly. The transmission from an Indian CFR tender to a US retail price runs through export netback, ocean freight, Gulf barge, terminal, and the retailer's own inventory cost — five margins, each with its own lag, and at least two with a physical inventory that was bought at an older price. The August tender is a real event in the world nitrogen market. It is not yet an event at a US farmgate, and the second piece in today's set takes apart exactly how large that gap has become.

References

1. Profercy — "RCF announces fresh urea purchasing tender to close 11 August," Aug 7, 2026. https://www.profercy.com/insights/rcf-announces-fresh-urea-purchasing-tender-to-close-11-august
2. Fertilizer Daily — "India issues 1.7 million tonne urea tender through RCF with bids opening on August 11," Aug 7, 2026. https://www.fertilizerdaily.com/20260807-india-rcf-urea-tender-1-7-million-tonnes-august-2026/
3. Bloomberg — "India Urea Offers Decline 12% as War Supply Squeeze Eases," Aug 13, 2026. https://www.bloomberg.com/news/articles/2026-08-13/india-urea-offers-decline-12-as-war-supply-squeeze-eases
4. Bloomberg — "India Gets Fertilizer Offers Near $1,000 as War Stokes Costs," Apr 15, 2026. https://www.bloomberg.com/news/articles/2026-04-15/india-gets-fertilizer-offers-near-1-000-as-war-stokes-costs
5. SunSirs — "New Tender After Two Months! India Seeks 1.7 Million Tons of Urea—A Key Market Bellwether," Aug 2026. https://www.sunsirs.com/commodity-news/petail-34939.html
6. StoneX — "India Urea Tender Reshapes Global Fertilizer Pricing," 2026. https://www.stonex.com/en-gb/insights/india-urea-tender-reshapes-global-fertilizer-pricing/

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

There are two urea markets and they are not currently telling the same story. In one, international sellers offered India urea at $393.65 to $435 a tonne CFR in the August tender round, down from offers near $1,000 in April (Bloomberg, Aug 13 and Apr 15, 2026). In the other, the average US retailer quoted urea at $678 a short ton in the week of August 10–14, against $826 in the week of March 23–27 (DTN/Progressive Farmer, Aug 19 and Apr 1, 2026).

Both markets fell. One fell 60.6%. The other fell 17.9%. The interesting question is not which number is right — both are correctly reported, on their own bases — but what happened to the difference.

1. The matched-endpoint comparison

Most commentary on this gap founders on mismatched dates: an international price from this week set against a retail price from last quarter, or a four-month move set against a twelve-month one. The comparison below uses the same two endpoints on both sides.

Source: international offers, Bloomberg, Apr 15 and Aug 13, 2026; US retail, DTN/Progressive Farmer prints of Apr 1, 2026 (week of Mar 23–27) and Aug 19, 2026 (week of Aug 10–14). Percentage changes and the pass-through ratio are Crop Root Zone calculations.

The pass-through ratio — the retail decline divided by the international decline — is 29.5%. Slightly under three-tenths of the world price move has reached a US farmgate quote. The date mismatch that remains is stated plainly: the DTN endpoint is the week of March 23–27 and the Bloomberg April figure is dated April 15, roughly three weeks later, at what appears to be the top of that spike. That mismatch flatters the international decline somewhat, and the ratio should be read as approximate. It does not come close to closing a gap this size.

2. Product by product, over the same five months

The retail side is worth breaking out, because the nitrogen products did not behave alike and the phosphates did the opposite of everything.

All prices $/short ton, delivered retail. Source: DTN/Progressive Farmer, Apr 1 and Aug 19, 2026. Percentage changes are Crop Root Zone calculations on those published averages.

Bars scaled so the largest absolute move is ten blocks; sign shown in the numeric column. Source as above.

Anhydrous is the outlier that matters most, because it is the cheapest nitrogen on the board and the one most likely to go on in the fall. Its five-month decline is 6.9% against the world market's 60.6% — a pass-through ratio of 11.4%, less than half of urea's. That is the specific reason the fall-application decision this year is not being made in the same market as the one setting international prices.

3. Both markets on one basis: dollars per pound of nitrogen

Short tons of one product and metric tonnes of another are not comparable, and the comparison above only works because it is like against like within each row. To put the two markets on the same ruler, both have to be converted to the unit the crop actually consumes.

Urea is 46-0-0. A metric tonne is 2,204.62 lb; a short ton is 2,000 lb.

Conversions are Crop Root Zone calculations at 2,204.62 lb per tonne, 2,000 lb per short ton and 46% N. DTN independently publishes urea at $0.74/lb.N for the same week, which the calculation reproduces. Source: Bloomberg, Aug 13, 2026; DTN/Progressive Farmer, Aug 19, 2026.

1.90×

US retail urea nitrogen, per pound, against the lowest offer India received in the August tender — $0.7370 versus $0.3882. (Crop Root Zone calculation on DTN, Aug 19, 2026 and Bloomberg, Aug 13, 2026.)

The multiple in March was not remotely this. DTN's late-March print put US retail urea at $0.90/lb.N (DTN/Progressive Farmer, Apr 1, 2026), while the April tender round was drawing offers near $1,000/t — about $0.99/lb.N on the same conversion. At those two points the US farmgate was below the price India was being asked to pay for a landed tonne. Five months later it is nearly double it.

That comparison carries the same three-week date mismatch flagged in §1 and should be treated as indicative rather than exact. The direction, however, is not in doubt: whatever the precise March multiple was, it was near or under 1.0, and the August multiple is 1.90.

Ground Truth: A wedge of this size is not evidence of anyone behaving badly, and reading it that way leads to the wrong decision. It is evidence of inventory age. A retailer's quote is set by what is in the shed, not by what is on the water, and the tonnes in the shed in August were bought into a market that had not yet fallen. The observable consequence is a timing rule rather than a fairness complaint: the pass-through arrives when the next fill is bought, not when the world price prints. That is why the two liquid products with the fastest inventory turns — urea and UAN32, both at −17.9% — have moved almost three times as far as anhydrous at −6.9%, which sits in dedicated storage and turns over on a seasonal rather than a monthly cycle. The pass-through ranking is an inventory-velocity ranking.

4. What the wedge is worth per acre

The abstraction becomes a decision at the rate sheet. A corn crop fertilized at 180 lb N per acre — a common Corn Belt rate, used here as an illustrative basis rather than an agronomic recommendation — costs the following at each of the prices above.

Per-pound-of-N figures for the four US retail products are DTN's own published values for the week of Aug 10–14 (DTN/Progressive Farmer, Aug 19, 2026). The India figure and all per-acre extensions are Crop Root Zone calculations. The 180 lb/ac rate is illustrative.

The spread between the cheapest and dearest US retail nitrogen at a fixed 180 lb rate is $37.80 an acre. The spread between US retail urea and the landed international price is $63.32 an acre — larger than the entire within-US product spread, and it is not a choice any grower can make. No one buys nitrogen at an Indian CFR tender price in Iowa.

Ground Truth: The practical read is that the product decision is worth less this year than the timing decision, and the ranking of the two has flipped. In a normal year the $37.80 spread across nitrogen forms is the biggest lever a grower controls, and the standard advice — buy the cheapest pound of N your equipment can apply — captures most of what is available. This year the timing wedge is 1.7 times the size of the form wedge. The question that is worth more money is not which nitrogen but when the retailer's next fill was bought, and that is a question a grower can actually ask at the counter. A retailer who has already refilled since July is quoting from a different cost base than one still selling spring tonnes, and there is no way to tell them apart from the posted price.

5. Limits on this reading

These are averages of a wide distribution. DTN's national averages conceal a real regional spread; the tender range conceals a freight spread. Neither number describes any particular transaction.

Retail includes services that CFR does not. A delivered retail tonne carries inland freight, storage, blending, application logistics, financing and a dealer margin. The 1.90× multiple is not 90% of profit; a large and unquantified share of it is real cost incurred between a vessel and a field. This piece does not attempt to decompose it, and no reader should treat the gap as a margin estimate.

One direction is untested. The pass-through ratio computed here describes a falling market. Nothing in these figures establishes that retail passes through rising international prices at the same 29.5% rate — and the anhydrous line hints strongly that it does not, since that product is up 27% year over year (DTN, Aug 19, 2026) while its five-month decline is the smallest on the board. Asymmetric pass-through is a well-documented pattern in input markets generally, and this data set is consistent with it, but a single falling episode cannot demonstrate it. It would take a matched rising episode to test.

References

1. DTN/Progressive Farmer — "5 Fertilizer Prices Slightly Lower Than Last Month; 3 Higher," Aug 19, 2026 (data week Aug 10–14, 2026). https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/19/5-fertilizer-prices-slightly-lower-3
2. DTN/Progressive Farmer — "4 Fertilizer Prices Rise Double-Digits; Anhydrous Jumps Above $1,000 Per Ton," Apr 1, 2026 (data week Mar 23–27, 2026). https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/04/01/4-fertilizer-prices-rise-double-1
3. DTN/Progressive Farmer — "Fertilizer Prices Fall as UAN32 Leads Significant Drops Heading Into August," Aug 5, 2026 (data week Jul 27–31, 2026). https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/05/fertilizer-prices-fall-uan32-leads
4. Bloomberg — "India Urea Offers Decline 12% as War Supply Squeeze Eases," Aug 13, 2026. https://www.bloomberg.com/news/articles/2026-08-13/india-urea-offers-decline-12-as-war-supply-squeeze-eases
5. Bloomberg — "India Gets Fertilizer Offers Near $1,000 as War Stokes Costs," Apr 15, 2026. https://www.bloomberg.com/news/articles/2026-04-15/india-gets-fertilizer-offers-near-1-000-as-war-stokes-costs

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Retail phosphate is doing something unusual this month: it is the only nutrient family on DTN's eight-product sheet that is rising. In the week of August 10-14, DAP averaged $917/ton and MAP $960/ton, both slightly higher than a month earlier and up 11% and 7% respectively against a year ago, while five nitrogen products fell (DTN/Progressive Farmer, Aug 19, 2026). That has sent a lot of growers to the same arithmetic — cost per pound of actual phosphate, rather than cost per ton of product — because the two products carry different analyses and the headline prices are not comparable. That calculation is correct as far as it goes. It also produces the wrong answer, for a reason that has nothing to do with phosphate.

1. The straight phosphate calculation, and what it says

What's new: DAP is 18-46-0 and MAP is 11-52-0. Per short ton, DAP delivers 920 lb of P2O5 and MAP delivers 1,040 lb. Divide the retail price by the phosphate delivered and the higher-priced product is the cheaper one.

Evidence: At $917/ton, DAP costs $0.9967 per lb of P2O5. At $960/ton, MAP costs $0.9231. MAP is 7.4% cheaper phosphate despite carrying a $43/ton premium on the invoice.

Source: DTN/Progressive Farmer retail fertilizer survey, week of Aug 10-14, 2026 (published Aug 19, 2026). Analyses are label grades. $/lb P2O5 is a Crop Root Zone calculation.

This is the calculation most ag-retail spreadsheets run, and it is the one that gets quoted back across the counter. It is also incomplete in a way that is easy to miss, because the thing it leaves out does not appear anywhere in the phosphate column.

2. Both products are also nitrogen, and nitrogen is not free

What's new: A ton of DAP carries 360 lb of nitrogen. A ton of MAP carries 220 lb. The 140 lb difference is not a rounding item — at this month's retail nitrogen prices it is worth $83 to $112 a ton, which is more than twice the $43 spread between the two products.

Evidence: The same DTN survey prices nitrogen four ways: urea at $0.74/lb N, anhydrous at $0.59/lb N, UAN28 at $0.80/lb N and UAN32 at $0.72/lb N. Credit each product's contained nitrogen at any of those, subtract it from the ton price, and divide what remains by the phosphate delivered. The ranking reverses on every one of them.

Source: product prices and $/lb N from DTN/Progressive Farmer, week of Aug 10-14, 2026. Net-of-nitrogen phosphate costs are Crop Root Zone calculations.

The break-even is the number worth writing down. Setting the two net-of-nitrogen costs equal and solving gives a nitrogen value of $0.4098/lb N. Above that, DAP is the cheaper phosphate; below it, MAP is. The cheapest nitrogen on the retail sheet this month is anhydrous at $0.59 — 44% above the break-even. The dearest, UAN28 at $0.80, is 95% above it.

$0.41/lb N

The nitrogen price at which DAP and MAP cost exactly the same per pound of phosphate. Every nitrogen product on this month's retail sheet is priced above it. (Crop Root Zone calculation from DTN/Progressive Farmer retail data, week of Aug 10-14, 2026)

That cushion is what makes the conclusion durable rather than a curiosity of one week's print. Nitrogen would have to fall roughly 31% from anhydrous's current level — or phosphate would have to move a long way — before the naive calculation started giving the right answer again.

It is worth being clear that the nitrogen in these products is not a marketing rounding of the analysis. It is there because of how they are made. Both start as phosphoric acid and are neutralised with ammonia; MAP takes roughly one mole of ammonia per mole of acid and DAP takes about two. DAP is, quite literally, the more ammoniated product, and the 140 lb/ton nitrogen difference is the physical record of that extra ammonia. Crediting it is therefore not an accounting device applied after the fact — it is recovering a cost the manufacturer already paid and passed on in the ton price.

That has a consequence for which way the premium moves. Because DAP consumes about twice the ammonia per ton, its cost base is the more ammonia-sensitive of the two. A sustained rise in ammonia raises DAP's production cost faster than MAP's, which works to compress the MAP premium — and compressing that premium is the one mechanism that moves this market toward the crossover described in section 4. A grower who wants a single leading indicator for whether this conclusion is about to weaken should watch ammonia, not phosphate rock.

Ground Truth: The nitrogen credit is worth face value only if the nitrogen arrives when the crop can use it. A spring-applied or starter-placed DAP program collects the full credit, because the 27 lb of N per acre it carries displaces 27 lb the grower was going to buy anyway six weeks later. A fall-applied DAP program on warm soil may not: nitrate that moves out of the profile over winter is nitrogen the grower paid for in October and cannot credit in June. The flip in this table is a spring and starter conclusion with more confidence than it is a fall one, and any grower running the arithmetic on a fall P program should discount the credit by whatever they honestly believe their overwinter loss to be. The break-even is $0.41/lb N — if half the nitrogen is gone by spring, the effective credit on anhydrous-equivalent N drops to about $0.30 and MAP wins after all.

3. What it is worth on an acre

What's new: The invoice and the all-in cost point in opposite directions, and the gap between them is about $9/acre on a typical program.

Evidence: Take a 70 lb P2O5/acre rate — roughly crop removal for a 180 bu corn crop at 0.35-0.40 lb P2O5/bu. Delivering it takes 152.2 lb of DAP or 134.6 lb of MAP.

Source: Crop Root Zone calculation from DTN/Progressive Farmer retail prices, week of Aug 10-14, 2026. Removal rate assumption stated, not sourced — run your own rate.

The MAP invoice is $5.16/acre cheaper and the MAP program is $4.16/acre more expensive. Substituting the make-up nitrogen at anhydrous rather than urea narrows the all-in gap to $2.27/acre; at UAN28 it widens to $4.91. On 2,000 acres the urea case is about $8,300 — not a farm-saving number, but a real one, and it is being decided by a line item that does not appear in the phosphate comparison at all.

4. How far prices would have to move to change the answer

What's new: The conclusion is not sitting on a knife edge. Holding the nitrogen credit at urea's $0.74/lb N, either product would have to move far enough to cross the other in price before the ranking reverted.

Evidence: Solve for the MAP price at which its net-of-nitrogen phosphate cost equals DAP's $0.7072/lb. It is $898.26/ton — $61.74 below where MAP sits today, and $18.74 below DAP's current $917. Run it the other way and DAP would have to rise to $971.62/ton, $54.62 above today and $11.62 above MAP's current price, before MAP became the cheaper phosphate.

Source: Crop Root Zone calculations from DTN/Progressive Farmer retail prices, week of Aug 10-14, 2026.

That is the real robustness test, and it is a stronger result than the break-even nitrogen price. MAP carries more P2O5 per ton and is the preferred product for close-placed and blended applications, which is why it normally trades at a premium to DAP — as it does today, by $43. For the naive per-pound calculation to give the right answer, that premium would not merely have to compress; it would have to invert, and stay inverted. Nothing in this month's data points that way: both products moved higher, and the wider phosphate market is being supported by China's continued export restriction through August and by ammonia and sulphur input costs that squeeze producer margins rather than buyer prices (Argus Media phosphate market commentary, August 2026).

So there are two independent ways for a grower to satisfy themselves that this is not a one-week artifact. The nitrogen would have to fall 31% from anhydrous's current level to reach the $0.41 break-even, or the two phosphate products would have to swap places on the price sheet. Neither is a small move, and they are not correlated — which means the conclusion survives a fairly wide range of plausible autumn markets.

5. Where the calculation stops working

What's new: Three things break this, and none of them are arithmetic.

Evidence and caveats, stated plainly:

• The two products are not agronomically interchangeable. MAP's lower ammonia release makes it the safer choice in seed-furrow and close-band placement, where DAP can injure germinating seed at higher rates. If placement is the constraint, the cost ranking is not the decision, and a grower who moves to DAP for $4/acre and takes a stand loss has lost the argument by an order of magnitude.
• The credit assumes displacement. It is only real if the grower reduces another nitrogen purchase pound-for-pound. A grower already at their agronomic or regulatory nitrogen ceiling is not saving $0.74/lb on the DAP nitrogen — they are applying nitrogen they cannot count, and for them the naive calculation is the right one.
• These are national retail averages. DTN surveys retailers across the country and reports a mean. Local availability, freight and fill programs routinely move individual quotes further than the $43 spread this whole piece turns on, and a single-source terminal can reverse the ranking on delivered basis alone.

What to watch next: the ratio, not the levels. This conclusion is a statement about the nitrogen-to-phosphate price relationship, and it flips only if nitrogen falls under about $0.41/lb N while phosphate holds. Anhydrous is down month-over-month but still 27% above a year ago; DAP and MAP are both higher than a month ago. Both moves push the same way — deeper into the range where crediting the nitrogen is what decides which phosphate is cheap.

References

1. Russ Quinn, "5 Fertilizer Prices Slightly Lower Than Last Month; 3 Higher," DTN/Progressive Farmer, Aug 19, 2026 (retail survey week of Aug 10-14, 2026). https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/19/5-fertilizer-prices-slightly-lower-3
2. DTN/Progressive Farmer, "Fertilizers Start August With Mixed Prices as 5 Nutrients Decline, Led by UAN32," Aug 12, 2026. https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/12/fertilizers-start-august-mixed-5-led

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

There is a sentence in every farm-press wrap-up and every lender's credit memo this month that reads some version of "fertilizer prices are up around 7% on the year." It is arithmetically true and operationally useless. DTN's retail survey for the week of August 10-14 covers eight products; seven are higher than a year ago and one is lower, and the distance between the extremes is 33 percentage points (DTN/Progressive Farmer, Aug 19, 2026). The average of that set is a number that describes almost none of the growers paying it.

1. The fan, measured

What's new: Anhydrous ammonia is 27% more expensive than a year ago. UAN32 is 6% cheaper. Both are nitrogen. Both were bought last year by people who will buy them again this year.

Evidence: The full eight-product survey, with each product's year-over-year change:

Source: DTN/Progressive Farmer retail fertilizer survey, week of Aug 10-14, 2026 (published Aug 19, 2026). Bars scale positive YoY changes only; UAN32 is the single decliner.

Run the descriptive statistics on those eight changes and one number does the work of the whole article:

Source: Crop Root Zone calculation from the DTN survey figures above. Population standard deviation across the eight reported YoY changes.

8.7 points

The standard deviation of year-over-year change across the eight retail nutrients — larger than the +7.5% average move it is measured around. (Crop Root Zone calculation from DTN/Progressive Farmer retail data, week of Aug 10-14, 2026)

A coefficient of variation above 1.0 is the formal statement of the problem: the dispersion inside the basket exceeds the movement of the basket. When that is true, the average has less information in it than the spread does, and quoting the average to someone who has to write a cheque is close to quoting them nothing.

2. What it does to two real programs

What's new: The dispersion is not an abstraction about indices. It shows up as a two-figure difference in the fertilizer line of a corn budget, produced entirely by which products a grower runs.

Evidence: Take two growers on the same ground, both applying 180 lb N and 70 lb P2O5 per acre — the same agronomy, the same rates, the same year. Grower A runs anhydrous and DAP. Grower B runs UAN32 and MAP. Both programs credit the nitrogen carried in the phosphate product against the nitrogen bill.

Source: Crop Root Zone calculation. Current prices from DTN/Progressive Farmer, week of Aug 10-14, 2026. Year-ago prices are back-solved from the same survey's reported YoY percentages — see the caveat in section 4. Rates are stated assumptions, not a recommendation.

The gap in year-over-year experience is 21.3 percentage points. Neither grower did anything unusual; both ran mainstream programs that any agronomist would sign off on. One of them is going into 2027 budgeting having watched their fertilizer line rise by a fifth, and the other is going in having watched it fall slightly, and the difference is entirely product selection.

Ground Truth: The more useful finding here is not that A had a bad year — it is that the level gap between the two programs is closing, and closing from the direction that matters. A year ago Program B cost $53.16/acre more than Program A. Today it costs $23.74 more. The premium a grower pays for the convenience and application flexibility of a liquid UAN program has been cut by more than half in twelve months, not because UAN got cheap in absolute terms but because anhydrous stopped being cheap. That is a genuine change in the decision, and it is invisible in any average. Anyone who ruled out a liquid program on cost in 2025 was ruling on a $53 spread that no longer exists, and the case should be re-run on current quotes rather than on last year's conclusion.

3. Why nitrogen split from itself

What's new: The widest divergence in the survey is not between nutrient families. It is inside nitrogen, between two products that deliver the same element to the same crop.

Evidence: On a cost-per-pound-of-nitrogen basis the same survey reports urea at $0.74/lb N, anhydrous at $0.59, UAN28 at $0.80 and UAN32 at $0.72. Anhydrous is still the cheapest nitrogen a grower can buy — by 18% against UAN32 and 26% against UAN28 — and it is simultaneously the product that has risen most. Those two facts are usually assumed to be in tension, and they are not: a product can be the cheapest in the set and still be the one closing the gap.

That is what makes the year-over-year column misleading read on its own. A +27% move on the cheapest unit and a −6% move on a dearer one is convergence, not divergence, in the thing a grower actually buys — pounds of nitrogen. The basket's dispersion widened while the underlying nitrogen market narrowed. Both statements are true from the same table, and they support opposite conclusions about whether a grower should switch programs.

The phosphate and potash side did the opposite. DAP, MAP and 10-34-0 moved +11%, +7% and +7% — tightly clustered, all positive, all in the same direction, which is what a nutrient family behaves like when a common upstream cost is doing the work. Potash at +2% is the outlier of stillness in a survey full of movement, and it is the one line in the table that has essentially not participated in the last twelve months at all.

4. The line that didn't move

What's new: Potash at +2% is not just the smallest move in the survey. It is small enough that it needs explaining in a year when the average nutrient moved 7.5% and the widest moved 27%.

Evidence: Potash averaged $495/ton in the week of August 10-14, slightly higher than a month earlier and 2% above a year ago. Every other product in the survey moved at least three times as far from its year-ago level in one direction or the other. On a twelve-month view potash has effectively sat still through a period in which nitrogen ran up 27% at one end and down 6% at the other.

Source: Crop Root Zone calculation from DTN/Progressive Farmer retail survey, week of Aug 10-14, 2026. UAN32's 6-point move is negative; the rest are positive.

Low variance is itself information. Nitrogen prices move because nitrogen is manufactured continuously from natural gas by many producers in many countries, and both the input and the output trade actively — so the price is discovered constantly and reflects gas, freight, tenders and outages in something close to real time. Potash comes from a small number of very large mines with long lead times, high fixed costs and the ability to idle capacity, and much of it moves under contracts negotiated at intervals rather than continuously. A price series that barely moves for twelve months in a volatile market is the signature of a product whose price is administered against available capacity rather than cleared against daily supply and demand.

For a grower the practical consequence is about which input rewards timing. Nitrogen has offered a 33-point spread of outcomes over twelve months, which means when a grower prices nitrogen has mattered enormously. Potash has offered two points, which means it has barely mattered at all. Spending scarce attention equally across the nutrient budget is a mis-allocation when the dispersion is this uneven — the marketing effort belongs where the variance is, and this year all of the variance is in the nitrogen column.

5. What would falsify this, and what the numbers can't carry

What's new: Three limits, stated so the arithmetic above is not asked to do more than it can.

Evidence and caveats:

• The year-ago prices are reconstructed, not observed. DTN reports current dollar levels and year-over-year percentages, and several of those percentages are published as rounded whole numbers or, for the smallest moves, as qualitative descriptions. Back-solving a year-ago level from a rounded percentage carries error — on a $960 product a rounding of ±0.5 points is roughly ±$5/ton. That error is small against a 21-point program gap but it is not zero, and no conclusion here should be run to the dollar.
• These are national retail averages. DTN surveys retailers across the country and reports a mean. A single grower's quotes are set by their terminal, their freight, their fill-program timing and their prepay position, any of which can exceed the dispersion described here. The claim is about the shape of the national distribution, not about what any individual will be charged.
• The two programs are not equivalent in everything but price. Anhydrous requires different equipment, a different application window and a different safety regime than UAN. Some of the $23.74/acre remaining gap is buying real things. The point is that the size of what it buys has changed, not that the two programs were ever interchangeable.

What to watch next: the standard deviation, tracked as its own series. If the dispersion compresses back below the mean move, the sector's habit of quoting a single "fertilizer price" becomes defensible again. While it stays above — and it has been widening since anhydrous began its run — the only budget number worth anything is the one built from the grower's own product list. On the current print, that is not a stylistic preference. It is what a coefficient of variation of 1.16 means.

References

1. Russ Quinn, "5 Fertilizer Prices Slightly Lower Than Last Month; 3 Higher," DTN/Progressive Farmer, Aug 19, 2026 (retail survey week of Aug 10-14, 2026). https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/19/5-fertilizer-prices-slightly-lower-3
2. DTN/Progressive Farmer, "Fertilizer Prices Fall as UAN32 Leads Significant Drops Heading Into August," Aug 5, 2026. https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/05/fertilizer-prices-fall-uan32-leads
3. DTN/Progressive Farmer, "Fertilizers Start August With Mixed Prices as 5 Nutrients Decline, Led by UAN32," Aug 12, 2026. https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/12/fertilizers-start-august-mixed-5-led

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Third-quarter contract prices for molten sulfur delivered into Tampa settled at a record $705 per long ton, up $50/lt from the $655/lt agreed for the second quarter — an 8% rise on a number that had already broken the 2008 peak [1]. That is the headline, and it is the number that will be quoted all autumn.

It is also the smaller of the two numbers that matter. Assessed spot prices for sulfur exported from US Gulf refineries stood at $1,100–1,150 per tonne fob on 9 July [1]. Convert the contract to the same unit and the gap is not a rounding difference. It is the whole story of what the fourth quarter is going to cost.

1. The units have to be converted before the number means anything

A long ton is 2,240 pounds; a metric tonne is 2,204.6. The two differ by about 1.6%, which is small enough to ignore in conversation and large enough to matter when the underlying price is over $700. More importantly, the two quotes are on different bases entirely — one is a delivered quarterly contract into Tampa, the other is a spot export cargo fob the US Gulf.

Source: Argus Media, 3Q Tampa settlement and 9 July Gulf export assessment, 2026. Conversion at 1 long ton = 1.016 tonnes is Crop Root Zone's own.

On a common tonne basis the record contract is roughly $694, and spot is 58% to 66% above it. Both directions of that comparison deserve a caveat and both are stated here: a delivered Tampa contract and an fob Gulf export cargo are not the same product in the same place, and a quarterly contract is by design a smoothed number that trades some upside for security of supply. Neither caveat closes a gap of that size.

~$694/t

The record Q3 Tampa contract, converted out of long tons — against $1,100–1,150/t fob for spot Gulf export cargoes in early July. (Argus Media, 2026; conversion Crop Root Zone)

2. Where the tightness came from, and why it is not a US story

The supply cause is not domestic. Sulfur is overwhelmingly a by-product — of refining crude and of sweetening sour natural gas — so its supply is set by refinery and gas-processing runs, not by anyone's decision to produce sulfur. Nobody builds a sulfur mine to chase a price.

The tightening traces to the near-closure of the Strait of Hormuz, which sharply reduced shipments out of a region that supplies close to half of global sulfur production [1]. That is the specific mechanism worth holding onto: a by-product with essentially zero supply elasticity, concentrated in one exporting region, moving through one waterway.

What's new is not that sulfur is tight — it has been since the second-quarter settlement passed 2008 — but that the constraint is a logistics chokepoint rather than a production shortfall. Molecules that exist and cannot sail behave, in the short run, exactly like molecules that were never made.

Evidence for the elasticity claim is in the price structure itself. A market that could respond to $1,100/t with additional supply would not hold a spread of that size against a contract settled six weeks earlier. The spread persists because the marginal tonne cannot be produced on purpose.

Ground Truth: The record contract is being read as the news. It is the opposite — it is the lagging indicator, and its lag is the forecast. A quarterly contract settled at ~$694/t while spot ran $1,100–1,150/t has not yet priced the disruption that caused the move; it has priced the quarter before it. Unless Gulf export spot falls by roughly a third before the fourth-quarter negotiation, the Q4 Tampa settlement has a step-up embedded in it that is already arithmetically visible, and the size of that step is not a forecast about sulfur — it is the distance between two numbers that are both already published.

3. Why a grower who never buys sulfur cares what Tampa settles

Molten sulfur delivered into Tampa is not a farm input. It is a feedstock price, and it sits upstream of two things that are farm inputs.

The first is sulfuric acid, which is burned from elemental sulfur and is the reagent that dissolves phosphate rock into phosphoric acid, and from there into MAP and DAP. A sustained rise in the sulfur contract is a rise in the variable cost of every tonne of processed phosphate, and it arrives with a lag set by how much acid a producer had already contracted.

The second is ammonium sulfate, where sulfur cost is a more direct input and where the pass-through is faster. Argus's own 2026 viewpoint on the sector is titled around exactly this transmission — sulfur costs supporting ammonium sulfate prices through the year [2].

Source: Product chemistry standard; transmission characterisation is Crop Root Zone's own read, informed by Argus Media's 2026 sulfur/amsul viewpoint.

The practical consequence for a fall program is that the sulfur line and the phosphate line are not independent, even though they arrive on separate invoices. A grower budgeting DAP and AMS separately is budgeting the same upstream molecule twice and may be treating one of them as a hedge against the other.

4. What a quarterly contract actually buys, and who is carrying the basis

The instinct on seeing a contract 60% below spot is that the contract buyer got a bargain. That is not what a quarterly settlement is, and reading it that way misprices the risk sitting on both sides of it.

A quarterly delivered contract is a supply commitment first and a price second. The buyer — typically a phosphate or sulfur-burning acid producer that cannot run without a continuous molten stream — is buying certainty of volume into a plant that has no economic ability to stop. The seller is buying certainty of offtake for a by-product it cannot warehouse cheaply and cannot stop producing.

That structure has two consequences visible in this quarter's numbers.

First, the contract cannot lag forever. The lag is a function of settlement frequency, not of anyone's view. A quarterly mechanism absorbs a shock in at most two settlements — the one being negotiated when the shock lands, and the one after. The second-quarter contract had already passed the 2008 peak and the third added $50/lt [1]. Two consecutive record settlements in a mechanism that resets four times a year is what catching up looks like from the inside.

Second, someone is short the difference. A buyer holding contract tonnes at ~$694/t in a market where incremental tonnes cost $1,100–1,150/t fob is not making a windfall; it is making the margin it contracted for, on the volume it contracted for. Any tonne it needs above contract volume is bought at spot. The exposure therefore sits precisely with whoever misjudged their own run rate — and in a quarter when phosphate producers have their own margin pressure, the incentive is to run to contract volume and no further.

Source: Contract structure is standard for quarterly delivered molten sulfur; the exposure characterisation is Crop Root Zone's own read.

For a fertilizer buyer several steps downstream, the practical translation is that the contract price is the one that predicts posted product prices, and the spot price is the one that predicts availability. They answer different questions, and this quarter they are answering them very differently.

5. What to watch, and the one thing that would break this read

The fourth-quarter Tampa settlement is the single number to watch, and the useful framing is not "will it rise" but "how much of the gap does it close." A settlement that closes half the distance to July spot would land near $900/t; one that closes none would say the contract mechanism has decoupled from the export market, which would itself be the story.

Three things would break the reading above:

1. Hormuz normalising. The constraint is a waterway. If transit normalises, the spot premium can unwind faster than any production response could deliver, because the molecules already exist. This is the largest single risk to the argument and it is a political variable, not an industrial one.
2. Refinery run rates rising into winter. More crude processed is more by-product sulfur, with no decision required by anyone.
3. Demand destruction in phosphate. If processed-phosphate producers cut runs rather than pay for acid, sulfur demand falls without any change in supply.

None of those is visible in the July assessment. All three are the reason the fourth-quarter number is a settlement to watch rather than one to assume.

Ground Truth: The most useful discipline in this market right now is unit hygiene, and it is not a pedantic point. The contract is published in dollars per long ton and the spot export market in dollars per tonne, and the two are quoted side by side in trade coverage without conversion. A buyer comparing $705 to $1,100 sees a 56% gap; the true gap on a common basis is 58–66%. The direction of the error is small here, but the habit is what matters — a sulfur price without a unit and a basis is not a usable number, and this is a market where both are currently moving.

References

1. Argus Media. "Tampa 3Q liquid sulphur price hits record $705/lt." 2026. Third-quarter Tampa molten sulfur contract settlement, prior-quarter comparison, and 9 July US Gulf export spot assessment. https://www.argusmedia.com/en/news-and-insights/latest-market-news/2851577-tampa-3q-liquid-sulphur-price-hits-record-705-lt
2. Argus Media. "Viewpoint: Sulfur costs to support amsul prices in 2026." 2026. https://www.argusmedia.com/en/news-and-insights/latest-market-news/2770651-viewpoint-sulfur-costs-to-support-amsul-prices-in-2026
3. US Geological Survey. "Sulfur — Mineral Commodity Summaries 2026." USGS, 2026. Cited for the by-product structure of sulfur supply. https://pubs.usgs.gov/periodicals/mcs2026/mcs2026-sulfur.pdf

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

December corn settled Friday at $5.08½, up 25¼ cents on the week and a 2.5-year high for the contract, while September corn settled at $4.83¾ (Pro Farmer, Aug 21, 2026; Brownfield Ag News, Aug 21, 2026). The 24¾-cent gap between them is the number circulating in marketing commentary this morning, generally with the observation that the market is paying growers to wait.

The gap is real. It is not a storage carry, and treating it as one is the most expensive arithmetic error available in an August marketing plan.

1. Why September-to-December is not a carry

A calendar spread is a storage signal only when both contracts can be satisfied by the same physical grain. September 2026 corn cannot be.

The September contract's first notice day is August 31, 2026, with expiration September 14 (CME contract specifications via Barchart, Aug 24, 2026). August 31 is also the last day of the 2025/26 corn marketing year. September 2026 is therefore the final old-crop contract: every bushel deliverable against it was grown in 2025 and has already been in storage for the better part of a year. December 2026 is the first new-crop contract, and the corn that satisfies it is standing in fields in Illinois today.

No bin turns one into the other. A grower cannot store 2025 corn through the 2026 harvest to capture that 24¾ cents in any economically meaningful way, because by December the old-crop bushel is competing directly against a freshly harvested one and carries an extra year of quality risk to do it. Nor, obviously, can 2026 corn be stored backwards into September.

Aug 21 settlements: Pro Farmer and Brownfield Ag News, Aug 21, 2026. Aug 24 figures are delayed intraday quotes from Barchart during the Aug 24 session, not settlements, and are labelled as such throughout. Contract dates from CME specifications via Barchart.

What the September–December spread actually prices is the handover between two crops: how tight the tail of the old one is against how large the new one looks. That is a supply question, and a valuable one — but it answers nothing about whether to fill a bin.

2. The first real carry: December to March

Both December 2026 and March 2027 are new-crop contracts. The same physical bushel, harvested in October, can be delivered against either. That spread is a storage price.

Aug 6 levels as reported in published settlement data; Aug 24 figures are delayed intraday quotes from Barchart. Spreads are Crop Root Zone calculations. Note that the flat price rose 58½ cents between these two readings while the spread moved less than a cent — the carry has been notably stable through a sharp rally.

Call it 14¾ cents over roughly three months — 4.92 cents per bushel per month, or about 2.83% of the December price over the interval, which annualises to roughly 11.3% on a simple basis (Crop Root Zone calculations).

The stability is the first thing worth noticing. Between August 6 and August 24 the December contract rallied 58½ cents on flooding in the eastern Corn Belt, a bullish Pro Farmer tour and a weaker dollar (Pro Farmer, Aug 21, 2026). The carry moved three-quarters of a cent. Flat price and storage price are answering different questions and the market is keeping them separate.

3. What storage actually costs

Now price the spread against the two ways a bushel can be stored.

Commercial. Published elevator rate structures cluster around an initial charge of about 13 cents per bushel plus roughly 4.5 cents per month thereafter, with the broader range across facilities running 9–16 cents for the first three months and 2–4 cents per month after that (Iowa State University Extension, Ag Decision Maker C2-24; published elevator service-rate schedules, 2026). Taking the first structure for a three-month hold: 13 + (3 × 4.5) = 26.5 cents.

On-farm. A grower who owns the bin pays no cash storage fee. The cost is the interest on the money not received, plus shrink, handling and quality risk. At a $5.20½ December price and a 7% operating rate over three months: $5.205 × 0.07 × 0.25 = 9.11 cents.

All figures Crop Root Zone calculations. Interest computed on the Aug 24 intraday December price. Shrink, drying, handling and quality risk are excluded from the on-farm line and would reduce the net; they are discussed in §5.

162% vs 56%

The December–March carry against the on-farm cost of storage, and against a typical commercial rate schedule. Same spread, same three months, same bushel. (Crop Root Zone calculations on Barchart quotes, Aug 24, 2026, and published elevator rate structures.)

Ground Truth: A spread that pays 162% of one storage cost and 56% of another is not ambiguous — it is sorted. The market is not undecided about whether corn should be stored; it has priced storage at a level that pays the grower who owns steel and refuses to pay the one who rents it. This is the mechanism behind a finding this desk has now reached from three different directions: TRZ-0109 established that US grain storage has been built at roughly a fifth of its two-decade rate since 2020, and TRZ-0119 established that the freight squeeze migrates onto basis in a way that rewards the grower with a bin and no forward sale. The December–March spread is the same conclusion expressed as a price. A bin bought at today's steel cost is being paid back, in this three-month window alone, at about 5.6 cents a bushel of capacity — and that is before any basis appreciation, which is where the larger money in on-farm storage has historically been.

4. The two spreads say different things, and both are true

Set the two side by side and the apparent contradiction dissolves.

Sep→Dec from Aug 21 settlements; Dec→Mar from Aug 24 intraday quotes. Crop Root Zone calculations.

A December contract at a 2.5-year high is a scarcity signal about this crop. A December–March carry near 11% annualised is an abundance signal about the quantity arriving in October. Those are not in conflict, because they are answers to different questions — one about level, one about timing.

Ground Truth: The practical consequence is that the two spreads point to opposite halves of a marketing plan, and a grower can act on both at once without inconsistency. The old-crop/new-crop premium argues for moving remaining 2025 bushels rather than carrying them into a harvest that will compete with them directly — the market is offering more for that corn now than the calendar will. The December–March carry argues for putting the 2026 crop in the bin rather than selling it off the combine. The single decision that is wrong under both readings is the common one: holding old crop and selling new crop at harvest. That combination pays the storage cost on the bushel the market wants moved and forfeits the carry on the bushel the market is willing to pay to have stored.

5. What this does not account for

Shrink, drying and quality. The on-farm line above is interest only. A crop that comes out of a flooded eastern Corn Belt at high moisture carries drying cost and shrink that the calculation ignores, and grain put away wet does not keep for three months. On a 1.5% shrink assumption alone, a $5.20 bushel loses about 7.8 cents — which would consume most of the 5.64 cent net. This is the single largest omission in the piece and it points the same direction in every case: it makes on-farm storage less attractive than the table shows, not more.

Basis is not in these numbers. Everything above is futures-to-futures. The bushel a grower actually sells is futures plus basis, and harvest basis widening is the reason many growers store in the first place. Excluding it understates the case for storage; including it would require a local basis forecast, which this desk does not publish.

Commercial rate structures vary enormously. The 26.5-cent figure is one common structure, not a national average, and the same table shows that a facility at the low end of the published range flips the conclusion entirely. A grower should price the actual schedule at the actual elevator before treating the commercial line as settled.

The Aug 24 quotes are intraday. December, September and March figures for August 24 are delayed quotes taken during a session in which corn was up more than 2%, not settlements. The spread arithmetic is robust to this — the Dec→Mar spread reads 15.50¢ on August 6 and 14.75¢ on August 24, so the conclusion does not depend on either single reading — but the flat prices should not be quoted as closes.

References

1. Pro Farmer — "After the Bell | Corn and soybeans see bullish weekly closes," Aug 21, 2026. https://www.profarmer.com/news/after-bell/after-bell-corn-and-soybeans-see-bullish-weekly-closes
2. Brownfield Ag News — "Closing Grain and Livestock Futures: August 21, 2026," Aug 21, 2026. https://www.brownfieldagnews.com/market-news/closing-grain-and-livestock-futures-august-21-2026/
3. ADM Investor Services — "Ag Market View for Aug 21, 2026," Aug 21, 2026. https://www.admis.com/ag-market-view-for-aug-21-2026/
4. Barchart — Corn Sep '26 (ZCU26), Dec '26 (ZCZ26) and Mar '27 (ZCH27) futures quotes and contract specifications, accessed Aug 24, 2026. https://www.barchart.com/futures/quotes/ZCZ26
5. Iowa State University Extension and Outreach — "Estimating a Grain Storage Rental Rate," Ag Decision Maker File C2-24, revised Aug 2024. https://www.extension.iastate.edu/agdm/wholefarm/pdf/c2-24.pdf
6. CME Group — Corn futures contract calendar and Variable Storage Rate documentation. https://www.cmegroup.com/markets/agriculture/grains/corn/calendar

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Pro Farmer put the 2026 US corn crop at 15.344 billion bushels on a 173.2 bu/acre national yield on Friday, August 21, closing a tour that sampled more than 3,000 fields across seven states. USDA's August 12 forecast was 180.7 bu/acre and 16.013 billion bushels. The gap is 7.5 bushels an acre and 669 million bushels — and on the relay carried by ADM Investor Services, a 173.2 national yield would be the lowest in six years. Soybeans went the other way: Pro Farmer's 53.3 bu/acre and 4.572 billion bushels sit above USDA's 52.7 and 4.519 billion, and would be a national record.

This desk has a stake in that first number. On Thursday, working from the tour's first three days, we published an implied national yield of roughly 174.1 bu/acre and an implied shortfall of about 585 million bushels — and we attached a caveat saying the estimate probably overstated the decline. The final number came in below ours. The caveat was reasonable, it was published in good faith, and it pointed the wrong way. Working out why is worth more than the nine-tenths of a bushel involved.

1. What Pro Farmer actually published

What's new: The tour's national estimates arrived with explicit error bands, which is more than most private surveys offer and more than USDA's August number carries on its face.

Evidence: Corn at 173.2 bu/acre and 15.344 billion bushels, with a stated range of ±1% — 171.5 to 175.0 bu/acre, or 15.191 to 15.498 billion bushels. Soybeans at 53.3 bu/acre and 4.572 billion bushels, with a wider ±2% band — 52.2 to 54.4 bu/acre, or 4.481 to 4.664 billion bushels.

Source: Pro Farmer, Aug 21, 2026; USDA WASDE and Crop Production, Aug 12, 2026.

Note what the two sets of numbers imply about acreage. Divide production by yield and Pro Farmer is working from about 88.59 million harvested corn acres against USDA's 88.62 million, and about 85.78 million soybean acres against USDA's 85.77 million. The two organisations are within a rounding error on area. This is a pure yield disagreement, which is worth stating plainly because a great deal of commentary treats tour-versus-USDA gaps as though acreage were also in play. Here it is not: essentially all 669 million bushels is the 7.5-bushel yield difference multiplied by an acreage base both sides accept.

669 million bushels

The corn production gap between Pro Farmer's final tour estimate and USDA's August forecast — equal to 40.5% of USDA's projected 1.653-billion-bushel carryout. (Pro Farmer, Aug 21, 2026; USDA WASDE, Aug 12, 2026)

2. The caveat that pointed the wrong way

What's new: On Thursday we applied the tour's first-three-days year-over-year sample change to last year's national yield and got about 174.1 bu/acre. We then wrote that the unweighted average gave South Dakota — the worst result in the sample, and a small share of national production — the same weight as Illinois, and that a production-weighted version would therefore land closer to USDA.

Evidence: The two reporting units added after we published are the two largest and best in the survey. Minnesota came in at 199.01 bu/acre and Iowa at 193.98 bu/acre. Both sit far above the national figure. On the tour's own comparisons, Minnesota was 8.7% above its three-year average and Iowa 1.4% above.

Source: Pro Farmer Crop Tour, Aug 21, 2026.

Adding two high-yielding states — one of them nearly nine points above its own three-year average — to a sample that had been anchored by South Dakota should have raised the mean. Instead the published national number landed 0.9 bushels below our partial-sample estimate.

There are two candidate explanations and only one of them survives.

The first is that the new data was worse than it looked. It was not, in absolute terms — 199 and 194 bushels are strong. But note the middle column: both states were still down year over year. Minnesota −1.9%, Iowa −2.2%. The pattern this desk flagged on Thursday, in which every single reporting unit came in below its 2025 tour result, did not break when the biggest states reported. It extended. A state can be far above its three-year average and still be below last year, because last year was a record; the two comparisons answer different questions, and only the year-over-year one feeds a national change calculation.

The second explanation, and the correct one, is that Pro Farmer's national estimate is not an average of its state samples at all. The tour publishes district and state sample results, then produces a national figure that incorporates its own adjustments — for grain fill still to come, for ear and pod weight assumptions the sample formula cannot capture, and for the variability its scouts observed. Pro Farmer's own framing of the corn crop, that it has "solid potential, but also more variability and fewer bushels than appearances might suggest," is a description of exactly that adjustment: fields that look good from the road and count well on ears can still disappoint on weight.

Ground Truth: Our 174.1 and Pro Farmer's 173.2 are not the same kind of number and should never have been expected to reconcile to a tenth. Ours was a mechanical extrapolation of a sample's year-over-year change; theirs is a judgment-adjusted forecast. The honest reading is not that we were 0.9 bushels off — it is that two methods sharing no adjustment logic landed within one bushel of each other, and 6.6 to 7.5 bushels below USDA. The agreement is the result. The residual is method. A reader who takes away "the tour and our estimate disagreed" has read this exactly backwards.

3. The ladder of estimates, and where USDA sits on it

What's new: There are now four published 2026 national corn yield figures built on materially different inputs, and they order themselves cleanly.

Evidence: DTN's Digital Yield Tour, published August 10, used satellite NDVI and models trained on Risk Management Agency data and produced 178.5. USDA's August 12 survey-based forecast is 180.7. Our sample extrapolation is 174.1. Pro Farmer's field-count estimate is 173.2.

Source: USDA, Aug 12, 2026; DTN, Aug 10, 2026; Pro Farmer, Aug 21, 2026. *Crop Root Zone estimate, method stated above — not a survey.

Scaled to USDA = ten blocks. Source: as above.

USDA sits at the top of the range, not in the middle of it. Every independent estimate published in August is below it, and the two built on physical field observation are the furthest below. That is not proof USDA is wrong — the August survey is the only one of the four with a legally compelled response base and objective yield plots behind it, and it is the only one that will be revised against actual harvested results. But when three methods with no shared inputs all land on the same side, the burden of the next revision is asymmetric.

There is a timing point worth carrying forward. DTN's own published comparison was against USDA's July 183.0, because it went out two days before the August WASDE. Measured against the August 180.7 that DTN gap is 2.2 bushels, not the 4.5 the original framing implied. USDA moved most of the way toward the satellite model on August 12. The set is converging, and it is converging downward.

4. What 669 million bushels does, and what it does not do

What's new: The temptation is to subtract the gap from carryout. Resist it.

Evidence: USDA's August balance sheet carries a 1.653-billion-bushel corn carryout. Take 669 million bushels straight off production with no other change and carryout falls to about 984 million — a 40.5% cut, and a genuinely tight number.

That arithmetic is correct and the conclusion it implies is not, for one reason: price rations use. December corn closed Friday at $5.08½, up 5 cents on the day, up 25¼ cents on the week, and at a two-and-a-half-year high for the contract — through its May high of $5.06½ though short of the $5.12½ contract high. A market that has already added 5.2% in a week is already discouraging feed demand and already making US corn less competitive in export tenders. Some meaningful fraction of any production cut USDA eventually makes will be absorbed by lower use, not by lower ending stocks.

Source: USDA WASDE, Aug 12, 2026; Pro Farmer, Aug 21, 2026. *Crop Root Zone scenarios. The demand-response split is illustrative, not modelled — we have not estimated corn's short-run elasticity and do not publish one here.

The middle row is the number the headlines will use. The bottom row is closer to how a balance sheet actually behaves, and we flag plainly that the 50% split is an illustration rather than an estimate. Anyone quoting a stocks-to-use ratio off the tour result this week is computing a third significant figure on a number with a 470-million-bushel standard error, which is roughly what USDA's own August forecast error implies.

5. Two crops, opposite news, one rally

What's new: The tour handed corn and soybeans directly opposing verdicts, and both markets went up.

Evidence: Pro Farmer put corn 7.5 bushels below USDA and soybeans 0.6 bushels above it — at a level that would be a national record. November soybeans nevertheless closed Friday at $12.39½, up 3 cents on the day and up 47 cents on the week.

Source: Closing grain futures, Aug 21, 2026, as relayed by Brownfield Ag News and ADM Investor Services.

A record yield estimate is not a reason for a 47-cent weekly rally. So the soybean move is not a supply story, and reading it as tour-driven is a category error. The market commentary carried with Friday's close attributes the week's strength across both pits to a weaker dollar, firmer fuel prices, continued export demand, and recent major flooding in the eastern Corn Belt. Those are macro and demand factors. They lift both boards regardless of what a pod count says.

Ground Truth: The two rallies have different half-lives and should not be marketed on one schedule. Corn's move rests on a production number that the September and October reports will either confirm or dismantle — it is a bet on a specific, scheduled, falsifiable event. Soybeans' move rests on demand and currency, which no crop report revises, and it arrived despite the best private yield estimate of the year. That makes the corn rally the one with event risk attached and the bean rally the one with none of it. This desk argued a fortnight of evidence ago that the two crops carry different amounts of unresolved forecast risk and should not come off one combine onto one marketing calendar; Friday's split verdict is the cleanest demonstration of it yet, and it points the same way.

6. What to watch

The September 11 WASDE is the first scheduled test. The relevant question is not whether USDA cuts — a cut is widely expected — but whether it cuts toward 178, which would leave the tour and the satellite models still arguing, or toward 175, which would concede most of the gap in one step. USDA has already moved 2.3 bushels once this season.

The second thing to watch is the harvested-acreage line, which nobody is watching. Both organisations are within 30,000 acres of each other today. August raised harvested area for both crops. If September revises area rather than yield, the production number can move without the yield debate resolving at all — and every ratio built on bushels per acre this month would be answering a question the balance sheet had stopped asking.

References

1. Pro Farmer, "Pro Farmer estimates 2026 U.S. corn crop at 15.344 billion bushels; soybeans at 4.572 billion bushels," Aug 21, 2026. https://www.profarmer.com/news/pro-farmer-estimates-2026-u-s-corn-crop-15-344-billion-bushels-soybeans-4-572-billion-bushels
2. Pro Farmer, "Recap: Pro Farmer Crop Tour concludes with sweep across Minnesota and Iowa," Aug 21, 2026. https://www.profarmer.com/news/live/live-coverage-pro-farmer-crop-tour-sweeps-across-iowa-and-minnesota
3. USDA, World Agricultural Supply and Demand Estimates and Crop Production, Aug 12, 2026.
4. DTN/Progressive Farmer, Digital Yield Tour, Aug 10, 2026.
5. ADM Investor Services, "Ag Market View for Aug 21, 2026." https://www.admis.com/ag-market-view-for-aug-21-2026/
6. Brownfield Ag News, "Closing Grain and Livestock Futures: August 21, 2026." https://www.brownfieldagnews.com/market-news/closing-grain-and-livestock-futures-august-21-2026/
7. Crop Root Zone, TRZ-0113, "Eight for Eight Below Last Year: the Crop Tour Sample and USDA Disagree by About 585 Million Bushels," Aug 20, 2026.
8. Crop Root Zone, TRZ-0114, "Same Fields, Same Weather, Different Odds: Corn 66% Below Trend, Soybeans 41%," Aug 20, 2026.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

The 2026 Pro Farmer Crop Tour finished its Iowa and Minnesota routes today, and the scouting data released so far has been consistent in a way that August rarely is. Through three days and eight reporting units, every single one came in below the same route's 2025 result — Ohio, South Dakota, Indiana, Nebraska, Illinois, and all three western Iowa crop districts (Pro Farmer, Aug 17-19, 2026). USDA's August WASDE, published August 12, carries a national corn yield of 180.7 bu/acre. Those two things are not yet reconciled, and the reconciliation is scheduled: Pro Farmer's national estimate posts Friday evening, August 21.

1. The sample, and how uniform it is

What's new: Uniformity is the finding. A tour that comes back with a wide spread of results tells you the crop is variable; a tour that comes back below last year everywhere tells you something narrower and more useful.

Evidence: Every unit reported through Wednesday, with both comparisons the tour publishes:

Source: Pro Farmer Crop Tour day-1 through day-3 recaps, Aug 17-19, 2026. Comparisons as published by Pro Farmer.

Eight of eight are below their 2025 tour result. Six of eight are also below the three-year tour average — the exceptions being western Iowa districts 1 and 7, which are marginally above it. The unweighted mean of the eight year-over-year changes is −6.65%, and the unweighted mean yield across them is 179.09 bu/acre.

2. The comparison USDA's number implies

What's new: USDA's August yield is 3.11% below last year. The tour sample is running roughly twice that far down.

Evidence: USDA's August 12 WASDE put the 2026 national corn yield at 180.7 bu/acre, cut from 183.0 in July, against a 2025 final of 186.5 — a decline of 3.11%. The tour sample's unweighted decline is 6.65%. Applying that to the 2025 national yield gives about 174.1 bu/acre.

Source: USDA WASDE, Aug 12, 2026; 2025 final per USDA as cited by DTN, Aug 10, 2026. Implied figure is a Crop Root Zone calculation and is expressly not a yield forecast — see section 4.

USDA's own production figure of 16.013 billion bushels against 180.7 bu/acre implies 88.62 million harvested acres. At that acreage, 6.6 bu/acre is 585 million bushels — against a projected carryout of 1.653 billion, roughly 35%.

585 million bu

The production difference between USDA's August corn yield and the figure implied by applying the tour sample's average year-over-year change to last year's national yield — about 35% of projected carryout. (Crop Root Zone calculation from Pro Farmer Crop Tour data Aug 17-19, 2026 and USDA WASDE, Aug 12, 2026)

3. What the market has already done about it

What's new: This is not an unpriced divergence. Corn has been moving toward the tour's view for a month.

Evidence: Corn settled at 476.92 cents/bu on August 20, up 0.83% on the day, 5.34% on the month and 23.16% on the year (Trading Economics, Aug 20, 2026). Soybeans were at 1,223.31 cents, up 0.09% on the day. The month-long move happened across the August WASDE and into the tour, which is exactly the window in which a smaller crop would get priced.

Ground Truth: The asymmetry has narrowed, and that is the part worth acting on rather than the headline gap. A month ago a bearish tour and a bullish tour were roughly equally unpriced. After a 5.34% move, a tour that confirms a materially smaller crop is partly in the market already, while a tour that lands closer to USDA is not. That does not make the tour's number less important — it makes the direction of surprise asymmetric, and any grower with unpriced new-crop bushels is holding a position whose payoff on Friday is no longer symmetric either. The decision that follows is about how much of the crop should be exposed to a single scheduled event, not about which way it will land. Note also that the soybean side of the tour has been far less uniform than the corn side — Illinois pod counts came in 2.9% above the three-year average and two of three western Iowa districts were up 8% and 14% — so a grower reading corn's uniformity across to their bean acres is reading across something the sample does not support.

4. A third estimate, from a method that shares nothing with the other two

What's new: There is a third independent read on the 2026 corn yield in the field this month, and it lands between USDA and the tour sample rather than outside them.

Evidence: DTN's Digital Yield Tour, published August 10, put the national corn yield at 178.5 bu/acre and soybeans at 52.1. Its method has almost nothing in common with either of the other two: it runs field-level models built on normalised difference vegetation index imagery, growing degree days and crop condition reports, trained on USDA Risk Management Agency insurance data rather than on NASS survey responses (DTN, Aug 10, 2026).

Source: USDA WASDE Aug 12, 2026; DTN Digital Yield Tour Aug 10, 2026; tour-implied figure is a Crop Root Zone calculation described in section 2.

Three methods with no shared inputs produce a 6.6 bu/acre spread, and they order themselves the way the sampling intensity would predict: the survey-based number is highest, the remote-sensed number is 2.2 bu lower — about 195 million bushels — and the boots-in-the-field number is lowest. Nothing here proves the field samplers are right. What it does establish is that USDA's figure is at the top of the current range rather than in the middle of it, and that the two non-government estimates both sit below it.

One timing detail matters and is easy to get wrong. The DTN estimate was published August 10, two days before the August WASDE, so its own stated comparison was against USDA's July yield of 183.0 — a gap of 4.5 bushels. Against USDA's revised August figure of 180.7 the gap is only 2.2. USDA moved most of the way toward the satellite model when it cut the yield on August 12. That is worth holding on to: one of the three estimates has already converged toward another since publication, which is a reminder that the spread in the table is a snapshot of a moving set, not a standing disagreement.

5. Why the implied number is a comparison and not a forecast

What's new: The 174.1 bu/acre figure above is a legitimate way to describe the sample and an illegitimate way to forecast the crop. The distinction matters enough to spend a section on.

Evidence and limits, stated plainly:

• The unweighted average is the biggest problem. It gives South Dakota the same weight as Illinois. South Dakota is the worst result in the sample and a far smaller share of national production, so an unweighted mean overstates the national decline by construction. A production-weighted version of this calculation would land closer to USDA than 174.1 does, and this piece does not compute one because the weights would have to be sourced separately and stated.
• The tour is a fixed-route sample, not a random one. Scouts pull ear counts and grain-length measurements from pre-set fields along set routes. That is a consistent methodology year to year, which is precisely what makes the year-over-year comparison the defensible part — and it is not a probability sample of the corn crop.
• Pro Farmer does not publish its state numbers as yield forecasts. The organisation applies its own adjustments before issuing a national estimate, and its national figure has historically differed from the raw arithmetic of its state results. Friday's number is the one Pro Farmer stands behind; the table in section 1 is not.
• The tour formula does not capture late grain fill. Corn yield in the tour is calculated from ear counts and measured ear length at the time of the stop. Kernel depth and test weight are still being set in late August, and a favourable September can add bushels the tour could not have seen.
• Two states are missing from this average. Minnesota and eastern Iowa were scouted Thursday and are not in the eight units above, and both are large enough to move a national figure. Their absence is not neutral either: the three western Iowa districts already in the sample are the best three results in it, so if eastern Iowa follows western Iowa the completed sample will read better than the eight-unit average above, not worse.
• Acreage moves independently of yield. USDA raised harvested area in the August report, so production and yield are not moving together this year, and a yield-only comparison of the kind above does not describe production.

There is one further asymmetry between the two numbers that is easy to miss. USDA's August yield sits inside a balance sheet in which harvested acreage was revised up, so the production figure of 16.013 billion bushels is being held up partly by area even as the yield came down. The tour measures only yield, on fields that were already planted. A grower comparing the two is therefore comparing a yield estimate against something that is not purely a yield estimate, and if the acreage revision itself is later trimmed, the production gap between the two views widens without either yield number changing. That is a second source of variance sitting underneath a comparison that already has plenty.

What to watch next: Friday evening's national corn and soybean figures, and specifically the distance between Pro Farmer's number and the raw arithmetic of its own state results. That distance is the size of the adjustment the organisation thinks its sampling method requires — and in a year when the sample is uniformly below last year, it is the single most informative number the tour will publish.

References

1. Pro Farmer, "Day 1 recap: Pro Farmer Crop Tour finds heat stress in South Dakota, excess moisture in Ohio," Aug 17, 2026. https://www.profarmer.com/news/live/live-coverage-2026-pro-farmer-crop-tour-scouts-set-sample-fields-across-7-states
2. Pro Farmer, "Day 2 recap: Crop Tour finds both Indiana and Nebraska lagging last year's corn yields, pod counts," Aug 18, 2026. https://www.profarmer.com/news/live/live-coverage-pro-farmer-crop-tour-rolls-indiana-nebraska
3. Pro Farmer, "Day 3 recap: Pro Farmer Crop Tour turns spotlight on Illinois and western Iowa," Aug 19, 2026. https://www.profarmer.com/news/live/live-coverage-pro-farmer-crop-tour-turns-spotlight-illinois-and-western-iowa
4. USDA, World Agricultural Supply and Demand Estimates, Aug 12, 2026.
5. DTN/Progressive Farmer, "DTN Digital Yield Tour Estimates US Average of 178.5 BPA Corn, 52.1 BPA Soybeans for 2026," Aug 10, 2026. https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/10/dtn-digital-yield-tour-estimates-us
6. Trading Economics, corn and soybean commodity quotes, Aug 20, 2026. https://tradingeconomics.com/commodity/corn

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Gregg Ibendahl at Kansas State published a state-level look at the 2026 soybean crop on August 19, and the number most worth pulling out of it is not the soybean number. It is the comparison the article makes in passing: the model puts the national probability of the soybean crop finishing below trend at 41%, and the corresponding figure for corn at 66% (Ibendahl, Kansas State University, Aug 19, 2026). Those two crops are grown in the same counties, in the same rotation, on the same rainfall. The model says one of them is a coin-flip-and-a-bit to disappoint and the other is closer to two-to-one.

1. The asymmetry

What's new: A 25-point gap in below-trend probability between corn and soybeans, out of one modelling framework applied to both crops in the same season.

Evidence: The Kansas State work runs state-level trend-and-condition regressions and aggregates. Its headline figures:

Source: Gregg Ibendahl, Kansas State University Dept. of Agricultural Economics, published via farmdoc daily, Aug 19, 2026. USDA figures from the August 12, 2026 WASDE.

Two things sit inside that table that pull in the same direction. On soybeans the model is above USDA — 53.4 against 52.7, implying production near 4,584 million bushels against USDA's 4,519 million. And on soybeans it assigns the lower below-trend probability. The corn side of the same framework does the opposite: the highest below-trend probability of the two, on the crop where the August field sampling has also been running below last year.

66% vs 41%

The modelled probability that the 2026 crop finishes below trend — corn against soybeans, from the same framework in the same week. (Gregg Ibendahl, Kansas State University, Aug 19, 2026)

2. The distinction the market keeps collapsing

What's new: "66% chance corn finishes below trend" is not "66% chance USDA cuts the yield again." Those are different sentences with different reference points, and reading one as the other is the most common way this class of number gets misused.

Evidence: Trend yield is a statistical construct — what the yield would be if this year performed like the fitted historical improvement path. The August WASDE is a forecast of the actual crop. The two can already disagree before anyone forecasts anything. If USDA's 180.7 is itself below the fitted trend, then a large share of that 66% probability is already satisfied by the number USDA has published, and expecting a further cut on the strength of the below-trend probability is double-counting.

This matters because the two statements imply opposite trades. "The crop is below trend" is a statement about the season. "USDA will revise lower" is a statement about a specific government estimate on a specific date. A grower can be right about the first and wrong about the second, and the second is the one that moves a futures contract on a report day.

Ground Truth: Use the asymmetry for sequencing, not for direction. The defensible read of 66% against 41% is not "corn goes up and beans don't" — it is that the corn bushel carries more unresolved forecast risk than the soybean bushel does this year, from the same model, in the same week. That is an argument about which crop a grower should be in less of a hurry to price, not about which way either one goes. It also cuts against the reflex of marketing both crops on one schedule because they came off the same farm. The model is saying the two crops are not carrying the same amount of unknown, and the marketing plan that treats them identically is the one plan the numbers actively contradict.

3. How wide the August number's error bar actually is

What's new: The Kansas State piece cites a companion corn analysis with a standard deviation of about 5.3 bu/acre on the August forecast. That is the single most useful number for anyone deciding how much weight to put on a WASDE.

Evidence: USDA's August production figure of 16.013 billion bushels against a yield of 180.7 implies 88.62 million harvested acres. One standard deviation of 5.3 bu/acre across that acreage is about 470 million bushels.

Source: standard deviation from the corn companion analysis cited in Ibendahl, Aug 19, 2026. Acreage, carryout and the bushel conversion are Crop Root Zone calculations from USDA WASDE, Aug 12, 2026.

A one-standard-deviation miss on the August corn yield moves production by more than a quarter of the entire projected carryout. That is not a criticism of USDA — a forecast made in the second week of August, before grain fill has finished, has no business being precise. It is a statement about how the number should be used. A stocks-to-use ratio computed off the August yield inherits an error bar wide enough that the ratio's third significant figure is decoration.

4. An independent method that points the same way on beans

What's new: The Kansas State model's more benign soybean read is corroborated this week by a completely different measurement — field pod counts from the Pro Farmer Crop Tour — and the corroboration is specific rather than general.

Evidence: The tour's corn results through Wednesday were uniformly below year-ago: eight of eight reporting units down against their 2025 route result. The soybean pod counts from the same stops were not. Illinois came in at 1,430.77 pods per 3'x3' square, down 3.3% from 2025 but up 2.9% against the three-year tour average. Two of the three western Iowa districts were up 8% and 14% against their averages, and the third was up 1.9%. Nebraska pods were essentially flat to average at −0.6% (Pro Farmer, Aug 17-19, 2026).

Source: Pro Farmer Crop Tour day-1 through day-3 recaps, Aug 17-19, 2026.

Four of the six units above show beans ahead of their three-year average while corn is at or below it. That is the same divergence the Kansas State model expresses as 41% against 66% — arrived at from ear counts and pod counts in fields rather than from condition-rating regressions, by people with no connection to the model. Two independent methods disagreeing would tell a grower to distrust both. Two independent methods splitting the two crops the same way, in the same week, is the more informative outcome, and it is the strongest single piece of support the asymmetry has.

The soybean model's quantitative gap to USDA points the same direction: 53.4 bu/acre against USDA's 52.7 is 0.7 bushels, or about 65 million bushels of production — a model that thinks USDA is slightly too pessimistic on beans, in a week when the field sampling was finding pod counts above average. Neither figure is large on its own. Together they describe a soybean crop that is doing better relative to expectations than the corn crop is, which is precisely what the below-trend probabilities said before either the pod counts or the model gap were examined.

5. What the numbers do not support

What's new: Four limits, and the first one is the author's own.

Evidence and caveats:

• There is no published soybean accuracy analysis. The Kansas State article states directly that a parallel historical accuracy study for soybeans has not yet been conducted and is planned for a future article. So the 41% has no published error bar of its own, and the 5.3 bu/acre standard deviation belongs to corn. The author is explicit that it should not transfer to soybeans — soybean yield is set later and by different weather, and its August forecast error has a different distribution.
• These are model outputs, not observations. Trend-and-condition regressions are fitted to crop condition ratings and historical yields. They are a disciplined way to summarise what is known in August; they are not a measurement of the crop and they do not know about anything the condition ratings have not picked up.
• The below-trend probabilities are national aggregates of state models. The article shows state-level trend deviations; a national probability assembled from state models carries assumptions about how state errors correlate, and correlated weather makes those errors far from independent.
• "Trend" is itself estimated. A different fitting window or functional form produces a different trend line and therefore a different probability of being below it. The 66/41 split is more robust than either number alone precisely because both were produced the same way.
• The pod-count corroboration is not independent of weather, only of method. Section 4 treats the tour and the model as two separate readings, and on methodology they are. But both are observing the same 2026 growing season, so a single widespread weather feature that affected corn and soybeans differently — a hot pollination window followed by an adequate August, for instance — would produce this exact pattern in both datasets without either one validating the other. Agreement between two methods looking at one season is weaker evidence than agreement between two methods looking at many, and this is one season.

It is also worth being clear about what a 66% below-trend probability is not worth in dollars. A probability without a magnitude does not size a decision: a 66% chance of finishing one bushel below trend and a 66% chance of finishing eight bushels below trend are the same probability and completely different outcomes. The Kansas State work publishes the odds, not the expected shortfall, and the 5.3 bu/acre standard deviation attaches to the August forecast error rather than to the deviation from trend. A grower who wants to convert the asymmetry into a position size needs the second number, and it is not in this article or in the companion one. Treating the 66% as a sizing input rather than a sequencing input is the most likely way to over-read it.

What to watch next: whether the promised soybean accuracy analysis lands before harvest, and what standard deviation it reports. If soybeans come back with a materially tighter August error bar than corn's 5.3 bu/acre, then the sequencing argument in section 2 strengthens considerably — the two crops would differ not only in their odds of disappointing but in how much is still genuinely unknown about each. If the soybean error bar comes back comparable to corn's, the asymmetry is about the season rather than about forecastability, and it should be expected to close as the crop finishes.

References

1. Gregg Ibendahl, "A Pre- and Post-Report Look at the 2026 U.S. Soybean Yield," farmdoc daily, Kansas State University Dept. of Agricultural Economics, Aug 19, 2026. https://farmdocdaily.illinois.edu/2026/08/a-pre-and-post-report-look-at-the-2026-u-s-soybean-yield.html
2. USDA, World Agricultural Supply and Demand Estimates, Aug 12, 2026.
3. DTN/Progressive Farmer, "USDA Raises Corn, Soybean Production With More Planted Acres for Both Crops," Aug 12, 2026. https://www.dtnpf.com/agriculture/web/ag/news/article/2026/08/12/usda-releases-august-crop-production-4

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

USDA's 2026 land values put the average value of US cropland at $6,020 an acre, up $190 from a year earlier — a 3.3% rise, and the first time the national cropland average has crossed $6,000 [1][2]. Farm real estate overall, cropland and pasture combined, averaged $4,500 an acre, up $150 or about 3.4%. Pastureland reached $2,000 [1][2].

In the same set of numbers, average cash rent for cropland came in at $160 an acre — one dollar lower than 2025 [1]. Irrigated cropland rent was $244, unchanged [1].

The two series are usually reported separately, and separately they read as unremarkable: land up a bit, rent flat. Divided into each other they say something the individual numbers do not.

1. The ratio nobody prints

Cash rent divided by land value is the cash yield a landowner earns on the asset before appreciation — the closest thing farmland has to a capitalization rate. Neither release publishes it. It is one division.

Source: USDA NASS Land Values 2026 Summary, released August 2026, as reported [1][2]. The 2025 value of $5,830 is derived by subtracting the stated $190 increase from the 2026 figure; the ratio and the change in it are Crop Root Zone's own calculation.

The implied yield on average US cropland fell roughly ten basis points in one year, and it fell for the least ambiguous possible reason: the numerator went down while the denominator went up.

Source: USDA NASS Land Values 2026 Summary [1][2]. Bar column scaled to the largest absolute move; the rent bar is shown as a negative.

2.66%

The implied cash yield on average US cropland in 2026 — $160 of rent against $6,020 of value. (Crop Root Zone calculation from USDA NASS Land Values 2026)

2. What a falling cap rate on farmland actually says

A capitalization rate falls when buyers accept less current income per dollar of asset. That happens for one of three reasons, and they have different implications for a grower.

Expected future income is rising. Buyers pay up today because they expect rents to be higher later. This is the benign explanation, and it is hard to square with a rent that just declined and with a crop-price environment that has not improved.

The discount rate is falling. Farmland competes with other yield assets; if the alternatives pay less, land can be worth more at the same rent. This is a capital-markets explanation and has nothing to do with agronomy.

The asset is being bought for something other than its operating income. Appreciation, portfolio diversification, estate considerations, development option value, or simply the desire of a neighbouring operator to control an adjacent parcel at nearly any price.

The 2026 numbers do not adjudicate between these. What they do establish is that whatever is setting the price of cropland this year, it is not the rent — because the rent moved the other way.

Ground Truth: The most consequential thing in this release is not that land hit a record. It is that the record was set while the asset's operating income fell, which severs the link a tenant intuitively relies on. Growers commonly reason that if land values rise, rent will follow, and budget defensively for it. This year says the reverse can happen, and it matters most for the operator negotiating 2027 ground: a landlord pointing to a 3.3% rise in land value is quoting a number that has just been shown not to drive rent. The counter is arithmetic and it is on the tenant's side — the 2026 data is the first recent instance where the two series decoupled, and it decoupled in the tenant's favour.

3. The averages hide the ground most readers farm

The national cropland average is not the number a Corn Belt operator transacts at. USDA's regional figure for the Corn Belt — Illinois, Indiana, Iowa, Missouri and Ohio — puts farm real estate at $8,590 an acre, well above the $4,500 national all-land average [2].

Source: USDA NASS Land Values 2026 Summary, as reported [1][2].

Two cautions belong with that table, and they are the reason this piece does not compute a Corn Belt cap rate. First, the $8,590 Corn Belt figure is farm real estate — cropland and pasture together — while the $6,020 is cropland only; they are not the same measurement. Second, the $160 rent is a national cropland average, and pairing it with a regional land value would produce a ratio that is not measuring anything. The high-value coastal states in the table are dominated by development option value and specialty production and should not be read as row-crop economics at all.

The honest version of the regional point is qualitative: cap rates compress fastest where land values are highest relative to what the ground can produce, and the national 2.66% is an average that includes a great deal of cheaper ground.

4. Where the return is now coming from

A landowner's total return on cropland is the cash yield plus the change in value. Both numbers are in this release, so the split can be computed rather than assumed.

Source: Crop Root Zone calculation from USDA NASS Land Values 2026 [1][2]. Cash yield is $160 ÷ $6,020; appreciation is the stated 3.3% rise in cropland value. Excludes property taxes, management cost and any leverage — this is a gross return on the unlevered asset, not a net return to an owner.

Fifty-five percent of the return came from the asset getting more expensive, and forty-five percent from what it earned. The direction of travel matters more than the level: appreciation is the larger share, and it is the share that is not contractually owed to anyone.

That distinction is the whole risk. Cash rent is a contract; a tenant owes it and generally pays it. Appreciation is a market price, owed by nobody, realised only on sale, and capable of going to zero or negative in a year without any change in the farming. An asset whose majority return component is the uncontracted one is a different holding from an asset where rent does most of the work — and on these numbers, 2026 is the year the split tipped.

The caveats belong in the open. This is a gross, unlevered figure: it is before property taxes, which in the higher-value Corn Belt states are a material drag on the cash yield, before any management cost, and before financing. An owner carrying debt sees both components amplified. A single year is also not a trend, and 2025's split would need the same computation on that year's own numbers to say whether the tipping is new.

5. The tenant's side of the same number

For an operator paying cash rent, this release is better news than the headline suggests, and the reason is the same arithmetic read from the other end.

Rent per acre is a cost line, and it declined by a dollar. That is not material on its own — $1 against $160 is 0.6%, which is inside the noise of any single farm's negotiation. What is material is the direction against the backdrop. Cash rent held roughly flat in a year when the asset it is charged on appreciated 3.3%, which suggests the rental market is being set by what the crop can pay rather than by what the land is worth.

That is the correct way for it to work, and it has not always held. It gives the tenant a specific, defensible position going into 2027 negotiations: rent is a function of crop margin, not of the landlord's balance sheet, and the 2026 national data is the evidence.

For the owner, the mirror image is less comfortable. An asset yielding 2.66% in cash, with the yield falling, is an asset whose return is increasingly dependent on continued appreciation. That is a different risk profile from the one farmland has historically been bought for, and it is more sensitive to the discount-rate channel than to anything happening in the field.

Ground Truth: Watch the gap between the cropland and irrigated series next year, because it is the cleanest available signal on which of the three explanations is operating. Irrigated cropland rent was unchanged at $244 while dryland cropland rent fell a dollar [1]. Irrigated ground carries a water right and a yield floor — it is the version of the asset whose income is least weather-dependent. If the irrigated rent holds while dryland rent softens further, the market is repricing production risk, which is an agronomic story a grower can act on. If both soften while values keep rising, the story is entirely in the discount rate and the buyer pool, and no amount of good farming changes it.

References

1. USDA National Agricultural Statistics Service. "Land Values 2026 Summary." Released August 2026. Average cropland value $6,020/acre (+$190, +3.3%); pastureland $2,000/acre; farm real estate $4,500/acre (+$150); cropland cash rent $160/acre (−$1); irrigated cropland cash rent $244/acre (unchanged). As reported by DTN/Progressive Farmer, "USDA Land Values Report: US Cropland Tops $6,000 an Acre in 2026," July 31, 2026. https://www.dtnpf.com/agriculture/web/ag/news/business-inputs/article/2026/07/31/usda-land-values-report-us-cropland
2. Farm Policy News, University of Illinois. "US Cropland Tops $6,000 per Acre for First Time in 2026." August 2026. Corn Belt farm real estate $8,590/acre and state-level values. https://farmpolicynews.illinois.edu/2026/08/us-cropland-tops-6000-per-acre-for-first-time-in-2026/
3. American Farm Bureau Federation. "Farmland Values Climb to New Highs as Cash Rents Hold Near Records." Farm Bureau Intel Markets, 2026. https://www.fb.org/intel/markets/farmland-values-climb-to-new-highs-as-cash-rents-hold-near-records

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Parts of Indiana took 15 to 18 inches of rain in the first weeks of August, and eastern Indiana into southeastern Ohio picked up more than 11 inches in a single week. Across the northern half of Illinois and Indiana and into eastern Ohio, it was among the top five wettest starts to any August dating back to 1893 (Pro Farmer, Aug 2026; AgWeb, Aug 2026). Corn and soybeans stood in water through a corridor running from central Illinois through central Indiana into eastern Ohio, and lodged and downed corn was reported through the same band after outflow winds accompanied the heaviest storms. In Madison County, Ohio, storms flattened cornfields outright.

That has been reported almost entirely as a yield story. It is more usefully read as a machine capacity story, because lodged corn is not lost in August — it is lost in October, at the header, and how much of it is lost is decided by equipment settings and ground speed rather than by anything that happened in the field this month.

1. The speed penalty, and what it does to a harvest calendar

Extension guidance for downed corn is consistent and unwelcome: reduce ground speed, flatten the corn head angle, and synchronise gathering chain speed to the new, slower ground speed. For picking corn up off the ground, University of Nebraska–Lincoln CropWatch puts the workable speed at as little as 1.5 mph (UNL CropWatch, "Assessing and Harvesting Downed Corn"). The gathering-chain synchronisation is specific: at 2 mph the chain should run about 27 revolutions per minute (UNL CropWatch, "Combine Adjustments for Downed Corn").

Field capacity is straightforward arithmetic. Acres per hour equals ground speed in mph times header width in feet, divided by 8.25, times field efficiency.

30-foot (12-row, 30-inch) corn head assumed throughout. All figures are Crop Root Zone calculations from the standard field-capacity formula at 80% field efficiency; speeds for downed corn are from UNL CropWatch guidance. Standing-corn speed of 4.5 mph is an illustrative Corn Belt working speed, not a source figure.

Bars scaled so the slowest case is ten blocks. Source: Crop Root Zone calculation as above.

The multipliers are 2.25× at 2.0 mph and 3.00× at 1.5 mph. In ten-hour days, a thousand-acre corn operation goes from roughly 7.6 harvest days to 17.2 or 22.9. A typical October–November window offers on the order of 30 to 40 suitable field days in the eastern Belt, so an operation that normally consumes a fifth of its window consumes half to three quarters of it — and every day of that expansion is a day the standing crop is exposed to the next weather event.

Fuel moves with the hours, not the acres. A combine at 2.0 mph burns close to the same per-hour rate as one at 4.5 and covers 44% of the ground, so fuel per acre rises by the same 2.25×. At the No. 2 farm diesel price of $4.65 a gallon reported for the week ending August 7 (USDA AMS, Illinois Production Cost Report) and an illustrative 10 gallons per hour, harvest fuel goes from about $5.84 to $13.14 an acre — a Crop Root Zone calculation on a stated consumption assumption, not a measured figure.

2. What is on the ground is worth more than it looks

The other half of the cost is what the header never picks up, and extension work has put a usable exchange rate on it.

Following windstorms in Nebraska, an assessment counted 14 ears on the ground in a row length equal to 1/1,000 of an acre, and put the associated loss at about 70 bu/ac (UNL CropWatch). That is a clean conversion factor:

~5 bu/ac

The yield represented by one ear on the ground per thousandth of an acre. (Crop Root Zone calculation from UNL CropWatch's 14-ear / 70 bu-per-acre assessment.)

Individual outcomes reported from the same events are far worse than the average suggests: one grower harvesting 230 bu/ac before a storm recovered 130 bu/ac after, and another went from 180 bu/ac to picking up just 40 bu/ac with a corn head (UNL CropWatch). Some operations resorted to windrowing stalks with a V-rake and picking them up with a bean head — an approach that recovers ears but adds an entire second field pass to an already extended calendar.

Source: UNL CropWatch, "Assessing and Harvesting Downed Corn" and "Harvesting Downed Corn." Percentages are Crop Root Zone calculations. These are individual reported cases from severe windstorm damage, not averages, and should not be applied to any region.

Ground Truth: Put the conversion factor against the national number everyone spent last week arguing about. Pro Farmer's final tour estimate came in 7.5 bu/ac under USDA — 173.2 against 180.7 — a gap this desk sized at 669 million bushels in TRZ-0116. At roughly 5 bu/ac per ear on the ground per thousandth of an acre, that entire national dispute is the size of one and a half ears. Not one and a half ears per plant; one and a half ears lying in the row over a thousandth of an acre. The point is not that the tour was wrong — it is that a yield survey and a combine are measuring two different quantities. A scout walking a field counts what grew. The header determines what arrives. In a year with a lodging corridor running from central Illinois to eastern Ohio, the wedge between those two numbers is wider than the entire gap between the two published national estimates, and no August survey of any design could have measured it, because the loss had not happened yet.

3. Why the capacity cannot simply be bought

The obvious response to a 2.25× hours multiplier is more machines — a second combine, a custom harvester, a longer day. Each of those is available to an individual operation and none of them is available to the region.

The constraint is that lodging damage is spatially correlated by construction. The same storm line that flattened corn in Madison County, Ohio moved through Indiana, Michigan, Illinois, Wisconsin and Iowa (AgWeb, Aug 2026). Every operation inside that corridor needs 2.25 times its normal machine hours in the same three weeks, and custom harvest capacity is drawn from the same corridor. A risk that hits every buyer of a service simultaneously cannot be transferred by buying that service; the price rises and the queue lengthens, but the aggregate hours do not appear.

This is the same structural point TRZ-0115 made from the manufacturer's side — that the equipment cycle is driven by replacement demand rather than by any single season's need — read from the field end. The fleet that will harvest the eastern Belt this October is the fleet that already exists, and a weather event in August cannot change its size.

Ground Truth: The throughput hit lands on a freight system this desk already found bid over tariff on every mode seven weeks before harvest peaks (TRZ-0119). The counterintuitive consequence is that lodging is mildly helpful to the logistics network and unambiguously bad for the grower. A harvest stretched from eight days to eighteen delivers the same bushels across more than twice as many days, which flattens exactly the delivery peak that overwhelms terminal and barge capacity — the elevator queue in the affected corridor may be shorter this year, not longer. The grower absorbs the entire cost of that smoothing, twice: once in field loss and once in extended weather exposure. And because a slower harvest arrives in smaller daily increments, more of it goes to on-farm storage rather than straight to a delivery point — which is the same direction the December–March carry is pushing (TRZ-0122), reached from a completely different mechanism. Two independent forces are moving the eastern Belt's 2026 crop into farm bins, and neither of them is a marketing decision.

4. What is settled here and what is not

The speed guidance is real and the arithmetic follows from it. The 1.5–2.0 mph figures are published extension recommendations, and the field-capacity formula is standard. The hours multipliers are therefore solid conditional on the speeds actually being needed.

The extent is not established. Nothing in the public reporting reviewed here quantifies how many acres in the affected corridor are lodged badly enough to require ground-speed reduction. "Lodged and downed corn reported through the same band" is a qualitative statement. Every per-acre and per-operation number in this piece is a conditional — what it costs if a field is in that condition — not a regional estimate, and it must not be scaled up to a state or national loss figure.

The severe cases are severe cases. The 43% and 78% recovery losses come from individual grower reports following windstorm damage, which is a different and generally more destructive mechanism than saturation lodging. They bound the top of the range; they do not describe it.

Some of the crop will stand back up. Corn lodged early enough and from the right growth stage can partially recover geotropically before harvest, and green-snap and root-lodged corn behave differently from stalk-lodged corn. The August condition is not the October condition, and the direction of that error runs against this piece's framing rather than for it.

References

1. Pro Farmer — "Eastern Corn Belt flooding will test 'rain makes grain' adage as Crop Tour weighs impact," Aug 2026. https://www.profarmer.com/news/eastern-corn-belt-flooding-will-test-rain-makes-grain-adage-crop-tour-weighs-impact
2. Pro Farmer / AgWeb — "Relentless Rain Muddies the Crop Production Picture Across the Corn Belt," Aug 2026. https://www.agweb.com/news/crops/crop-production/relentless-rain-muddies-crop-production-picture-across-corn-belt-ahead
3. University of Nebraska–Lincoln CropWatch — "Assessing and Harvesting Downed Corn." https://cropwatch.unl.edu/assessing-and-harvesting-downed-corn-unl-cropwatch-oct-26-2012
4. University of Nebraska–Lincoln CropWatch — "Combine Adjustments for Downed Corn." https://cropwatch.unl.edu/2017/combine-adjustments-downed-corn/
5. University of Nebraska–Lincoln CropWatch — "Harvesting Downed Corn." https://cropwatch.unl.edu/harvesting-downed-corn-unl-cropwatch-november-2-2012
6. USDA Agricultural Marketing Service — Illinois Production Cost Report, week ending Aug 7,
7. https://www.ams.usda.gov/mnreports/
8. Pro Farmer — "After the Bell | Corn and soybeans see bullish weekly closes," Aug 21, 2026, for the final tour yield estimate referenced in §2. https://www.profarmer.com/news/after-bell/after-bell-corn-and-soybeans-see-bullish-weekly-closes

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

USDA's Grain Transportation Report of August 20 describes a grain logistics system running at extraordinary volume. Corn inspections at Mississippi River terminals hit 1.1 million metric tons for the week ending August 13 — 176% above the prior three-year average and the third-highest weekly total of the year. Pacific Northwest inspections were 0.4 million tonnes, nearly 200% above average. Interior movements by rail to Mexico were 0.3 million tonnes, up 26%. In the Gulf, 29 oceangoing grain vessels loaded in the week ending August 13, 164% more than the same week last year.

In the same report, the number of barges moving downriver fell by 34 to 426, and New Orleans region unloads dropped 6% on the week to 684 grain barges.

Record throughput on a shrinking barge count is the kind of contradiction that usually resolves into a reporting-period artefact. This one does not, and working out why leads to a line most readers skip.

1. The volumes, and how big they actually are

What's new: Putting the inspection figures into bushels makes the scale legible in a way tonnes do not.

Evidence: At 39.368 bushels per metric ton of corn, the week's inspections convert as follows:

Source: USDA Grain Transportation Report, Aug 20, 2026, as relayed. Bushel conversion is a Crop Root Zone calculation at 39.368 bu/MT.

The four-week picture on the Mississippi system is the same story with the noise removed: an average of 494,000 short tons, up 19% from last year and 53% from the three-year average. This is not one strong week. It is a sustained run.

43.3 million bushels

Corn inspected for export at Mississippi River terminals in a single week, 176% above the prior three-year average — in August, before the new crop moves. (USDA Grain Transportation Report, Aug 20, 2026; bushel conversion Crop Root Zone)

2. The line nobody quotes

What's new: Buried in the rail section is a two-part figure that says more about system capacity than any tonnage number in the report.

Evidence: For the week ending August 13, secondary railcar bids had shuttle trains at $150 above tariff and non-shuttle bids at $50 above tariff. Class I railroads originated 26,705 grain carloads in the week ending August 8.

The secondary railcar market needs a sentence of explanation, because its sign is the whole point. Railroads sell grain car capacity at a posted tariff. Shippers who hold that capacity and do not need it can resell it, and shippers who need capacity they did not book can buy it. The resulting secondary price floats above or below tariff depending on whether the posted price is above or below what the market will bear.

When capacity is loose, the secondary market trades at a discount — shippers dump unwanted allocations at whatever they can get, and the figure prints as a negative number against tariff. That is the normal summer condition. A positive $150 premium on shuttles means the opposite: the railroad's own posted price is below the clearing price, and shippers are paying up in a secondary market to get cars the tariff market cannot supply.

Source: USDA Grain Transportation Report, Aug 20, 2026, and USDA forward barge rate data for August 2026 as relayed July 14, 2026.

Ground Truth: Four modes are bid to a premium in the same report week, and that is a different condition from any one of them being expensive. A single tight mode is a bottleneck, and bottlenecks are arbitraged — grain moves to rail when the river is short, and to truck when rail is short. When rail, barge, ocean and truck are all bid up simultaneously, there is nowhere to arbitrage to, and the system's response to the next increment of volume is not substitution but queueing. The secondary railcar premium is the cleanest evidence of it because, unlike a barge rate, it is a price shippers pay each other rather than one a carrier posts — it cannot be explained away as a tariff that has not been updated.

3. What the falling barge count probably means

What's new: A shrinking downriver fleet during a volume record is the report's genuine anomaly, and it has a physical candidate explanation.

Evidence: Barged grain movements totalled 638,650 tons for the week ending August 15 against 426 barges moving downriver. Divide the one by the other and the implied average load is about 1,499 tons per barge.

That figure needs a caution before it is used, and this desk would rather state the caution than bury it. The tonnage figure is a movements total for the barge grain system, while the barge count is a downriver count at reporting points; the two may not share the same scope, in which case the ratio is not a true average load. We publish it as an order-of-magnitude indicator only, and no conclusion below depends on it alone.

With that said, a standard covered hopper barge carries on the order of 1,500 short tons of grain at working draft and can carry meaningfully more — figures around 1,750 tons are commonly cited — when the river permits a full 12-foot draft. An implied load sitting at the lower end of that range, in a week of record demand when every operator has maximum incentive to load heavy, is consistent with draft restriction rather than with slack demand. Operators load light when the channel is shallow, and the third consecutive year of low autumn water on the Mississippi has been widely flagged as a risk into this harvest.

The alternative explanation is simpler and should be stated: barge counts move week to week for reasons including lock outages, weather and scheduling, and a 34-barge decline on a 460-barge base is a 7.4% move that could be noise. What makes the draft explanation more attractive is the company it keeps — a light average load alongside a 6% drop in New Orleans unloads, in a week when Gulf vessel loadings ran 164% above last year, describes a system whose demand is not the binding constraint.

4. Diesel, and what it costs per bushel

What's new: The diesel line in the same report is the largest single-week move in the set, and it converts directly into a number a grower recognises.

Evidence: Diesel rose 19.7 cents to $5.454 per gallon, and sits 174.1 cents above the same week last year — a 46.9% year-over-year increase from an implied $3.713.

Convert that to the farm-to-elevator haul:

Crop Root Zone calculation. Assumes a 900-bushel semi-hopper load, 100-mile round trip, 6 mpg — stated assumptions, not survey data. Fuel only; excludes labour, equipment and wait time.

Three and a bit cents a bushel is not a crisis on its own. It matters because it is additive to everything in section 2 and because it lands on the mode of last resort. Truck is what a country elevator uses when it cannot get a barge or a car — and its cost has risen 47% year over year in exactly the season when the other two are bid over tariff.

5. Why a smaller crop helps less than it looks

What's new: Pro Farmer's final estimate landed the day after this report, and the instinct is to read it as relief for a strained system. It is not, or at least not where it appears to be.

Evidence: Pro Farmer put the 2026 corn crop at 15.344 billion bushels against USDA's 16.013 billion — 669 million bushels less, or 4.2% fewer bushels needing to move. In isolation, 4.2% less freight demand takes real pressure off a system running over tariff.

The complication is that the export commitments are already sold. Vessel slots booked in the Gulf, rail shuttles bid at $150 over tariff, and sales already on the books do not shrink because a private survey lowered the yield forecast. The physical grain still has to reach the terminal on the contracted date. What changes is how hard the merchandiser has to work to source it.

Ground Truth: A smaller crop into an already-tight logistics system does not relieve the freight market. It migrates the squeeze from freight rates onto basis. The elevator that has sold bushels forward and now finds fewer of them in its draw area has to bid up to get them, and that bid is basis, not freight. So the practical read of Friday's tour number for anyone who stores or merchandises grain is not "cheaper freight this fall" — it is a firmer country bid competing against a fixed and expensive set of shipping commitments. Growers with on-farm storage and no forward sale are the party this arrangement favours; the elevator short the bushels is the party it does not.

6. What to watch

Three things, in order of how quickly they will tell you something.

The sign of the secondary railcar market. If shuttle bids fall back through tariff to a discount in September, the August premium was a pre-harvest positioning artefact and the system has slack after all. If the premium holds or widens into October, it was capacity.

The implied barge load. If the tons-per-barge ratio recovers toward the upper end of nominal capacity, the light August loads were scheduling. If it stays at or below 1,500 through September, the channel is the constraint and the fall programme gets rationed by draft.

Whether New Orleans unloads follow inspections up. Inspections at 176% above average with unloads falling 6% cannot both continue. One of them is describing the near future incorrectly, and the unload count is the one measuring what physically arrived.

References

1. USDA Agricultural Marketing Service, Grain Transportation Report, Aug 20, 2026, as relayed by IndexBox, "Grain Transportation Report August 20, 2026: Corn Exports Rise, Rail and Trucking Updates." https://www.indexbox.io/blog/usda-grain-report-corn-inspections-surge-rail-merger-review-resumes/
2. USDA AgTransport, grain barge rate data and forward curves for August 2026, as relayed July 14, 2026. https://www.indexbox.io/blog/usda-grain-barge-rate-report-mississippi-river-tariffs-and-forward-curves-july-14-2026/
3. USDA, Barge Dashboard, Open Ag Transport Data. https://agtransport.usda.gov/stories/s/Barge-Dashboard/965a-yzgy/
4. Pro Farmer, "Pro Farmer estimates 2026 U.S. corn crop at 15.344 billion bushels; soybeans at 4.572 billion bushels," Aug 21, 2026.
5. USDA, World Agricultural Supply and Demand Estimates, Aug 12, 2026.
6. Supply Chain Dive, "Low Mississippi River water levels disrupt barge shipping for third consecutive year." https://www.supplychaindive.com/news/mississippi-low-water-levels-third-year-barge-grain-shipping/727113/
7. American Farm Bureau Federation, Market Intel, "Low Mississippi River Levels Drive up Grain Transportation Costs." https://www.fb.org/market-intel/low-mississippi-river-levels-drive-up-grain-transportation-costs
8. Crop Root Zone, TRZ-0085, "The Memphis Gauge Hit Zero on August 3. Your Fall Fertilizer Rides the Same Barges as the Grain," Aug 11, 2026.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Deere reported fiscal third-quarter results on the morning of August 20: net income of $1.379 billion and $5.10 a share, against $1.289 billion and $4.75 a year earlier, on worldwide net sales and revenues of $12.608 billion, up 5% (Deere & Company, Aug 20, 2026). Both figures came in ahead of consensus. A grower reading the headline could reasonably conclude that the farm equipment market has turned. The segment table says something narrower and considerably more useful: the growth came from the parts of the company that do not sell to row-crop agriculture, and the row-crop segment shrank.

1. Where the quarter actually came from

What's new: Equipment operations sales rose $642 million year over year. Construction & Forestry alone accounted for $559 million of it — 87%. Production & Precision Agriculture subtracted $275 million.

Evidence: The three equipment segments, third quarter, against the year-ago quarter:

Source: Deere & Company third-quarter fiscal 2026 results, Aug 20, 2026.

Production & Precision Ag is the row-crop segment — large tractors, combines, planters, sprayers, the precision systems that go on them. It is the only one of the three that fell, on both lines. Of the $283 million of combined segment operating-profit growth, Construction & Forestry supplied $199 million and Small Ag & Turf supplied $137 million, while Production & Precision Ag gave back $53 million.

Source: Crop Root Zone calculation from Deere segment results, Aug 20, 2026.

2. The margin inversion inside agriculture

What's new: The premium row-crop segment now earns a lower operating margin than the lawn, compact-utility and small-tractor segment, and the gap has more than doubled in a year.

Evidence: Operating margin by segment, computed from the figures above:

Source: Crop Root Zone calculation from Deere segment results, Aug 20, 2026.

Small Ag & Turf was already ahead of Production & Precision Ag on margin a year ago, by 246 basis points. It is now ahead by 521. The high-horsepower, high-technology, high-ticket end of the business is the least profitable agricultural line Deere runs this quarter, on a percentage basis, and the direction of travel over the year is down while everything else moved up.

−6% / −9%

Production & Precision Agriculture sales and operating profit, year over year, in a quarter the company beat on both earnings and revenue. (Deere & Company third-quarter fiscal 2026 results, Aug 20, 2026)

3. What "the bottom" is a claim about

What's new: Chief executive John May said the company continues to believe 2026 will mark the bottom of the current ag equipment cycle, citing early order program trends, improving used-equipment inventories and adoption of advanced technology. The full-year guidance underneath that statement still has US and Canada large agriculture down 15-20%.

Evidence: The company guided fiscal 2026 net income to $4.75-5.00 billion, with US/Canada large agriculture down 15-20% and US/Canada small ag and turf flat to up 5%. So the bottom being called is a bottom in the rate of decline — a year that ends 15-20% lower is still a year that ends lower.

The evidence cited for the call is worth reading precisely, because two of the three items are inventory statements rather than demand statements. Early order programs are orders already taken for the following model year, which is the best forward indicator available and also the most interested one. Improving used-equipment inventories describe destocking — the working off of trade-ins that have been sitting on dealer lots and setting the price of the alternative to buying new. Only the third item, technology adoption, is a claim about what growers want rather than about what is in the channel.

Ground Truth: Shipments bottom when the factory stops cutting production. Farm equipment demand bottoms when crop margins recover. Those are different dates and only the first one is being called here. Everything in this release is evidence about the first: order books, dealer inventory, production rates. The second is set by the corn and soybean balance sheets, and on the same day this release landed, the Pro Farmer Crop Tour was returning corn yields below year-ago across every unit it had sampled while new-crop corn traded at 476.92 cents. For a grower the practical consequence runs the opposite way to the headline: the buyer's market of the last two years was built on the dealer inventory the company is now telling you is normalising. If that is true, the discounting window on high-horsepower equipment is closing before the farm-margin recovery that would make the equipment affordable has arrived. The worst possible sequence for a buyer is inventory normalising first and margins recovering second, and that is precisely the sequence this release describes.

4. The row-crop segment is shrinking as a share of the company

What's new: Production & Precision Ag has gone from 41.3% of Deere's equipment sales to 36.3% in twelve months — a 491 basis point fall in its share of the business.

Evidence: Each segment's share of equipment operations net sales, this quarter against the year-ago quarter:

Source: Crop Root Zone calculation from Deere segment results, Aug 20, 2026.

Agriculture as a whole — both ag segments together — fell from 70.5% of equipment sales to 67.1%. Within that, the shift is sharper than the total suggests, because Small Ag & Turf grew while the row-crop line contracted. Deere is not becoming less of a farm machinery company by intention; it is that one quarter of a down cycle in row-crop capital spending moves the mix by five points when the non-farm segments are growing at 12% and 18%.

That mix shift is the mechanism behind the bottom call, and it is worth naming as such. A manufacturer whose other segments are carrying the quarter has considerably more room to hold production discipline in the weak segment than one whose whole book is falling. Deere can afford to ship fewer combines this year because construction equipment is up 18% and turf is up 12%. A grower waiting for the discounting that comes from a manufacturer needing volume at any price should note that this particular manufacturer does not currently need it — which is a further argument that the buyer's-market window is a function of dealer inventory clearing rather than of factory desperation, and it closes when the inventory does.

5. What a grower should take from a manufacturer's quarter

What's new: Three readings that follow from the numbers, and one that does not.

Evidence:

• The used market is the number to track, not the new one. Trade-in values set the real cost of replacement, and "improving used-equipment inventories" is the company's way of saying the overhang that has been suppressing them is clearing. A grower whose replacement decision has been waiting for a better new-iron price should be checking whether the used side moves first, because it is the side that determines the net cheque.
• Guidance down 15-20% means the supply side keeps tightening. Fewer new units shipped into US and Canada large ag through the rest of the fiscal year is what eventually stops discounting. That is the mechanism by which a shipment bottom becomes a price floor, and it operates with a lag measured in quarters.
• The technology line is a pricing statement. "Increasing customer adoption of our advanced technologies" describes revenue per machine rising independently of unit volume. In a segment with falling units, that is the lever holding the sales line up, and a grower comparing a 2027 quote to a 2024 purchase is partly comparing content, not inflation.
• What does not follow: that the ag equipment market has turned. Nothing in this release says that. The beat came from construction and from small ag and turf, and the row-crop segment fell on both sales and profit.

Limits, stated plainly. These are three months of one manufacturer, which is not the machinery market — competitors' order books and the short-line makers may differ, and Deere's segment definitions do not map cleanly onto what a given farm buys. Quarterly segment results are also affected by shipment timing and by price realisation as well as by volume, and the release attributes part of the Production & Precision Ag decline to lower shipment volumes partially offset by favourable price realisation, which means the underlying unit decline is steeper than the 6% sales figure. Nothing here is a view on the company's securities; the numbers are used only as the best available read on what is happening to the cost and availability of farm capital equipment.

The used-equipment channel is where these two clocks actually meet, which is why it is the line in the release worth the most attention. Used values are set by the stock of trade-ins on dealer lots, and that stock is fed by new sales — every new combine delivered puts a used one into the channel. Two years of falling new sales therefore does two things in sequence: it starves the used channel of supply, which supports used values, and it removes the dealer's incentive to discount new iron to move inventory. Both effects raise the net cost of replacement for a grower, and both are consequences of exactly the destocking the company is presenting as evidence of recovery. The same sentence that reads as good news for the manufacturer reads as a closing window for the buyer, and nothing in the release resolves that tension because the release is not written for the buyer.

What to watch next: the fiscal fourth-quarter release in November, and specifically whether US/Canada large ag guidance for fiscal 2027 comes in above or below flat. That is the first data point that will test the bottom call with a forward number rather than a backward one, and it will land after harvest, when the crop margins that actually drive the replacement decision are known rather than forecast.

References

1. Deere & Company, "Deere Reports Third Quarter Net Income of $1.379 Billion," Aug 20, 2026. https://www.prnewswire.com/news-releases/deere-reports-third-quarter-net-income-of-1-379-billion-302856316.html
2. Pro Farmer, "Day 3 recap: Pro Farmer Crop Tour turns spotlight on Illinois and western Iowa," Aug 19, 2026. https://www.profarmer.com/news/live/live-coverage-pro-farmer-crop-tour-turns-spotlight-illinois-and-western-iowa
3. Trading Economics, corn commodity quote, Aug 20, 2026. https://tradingeconomics.com/commodity/corn

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

US grain storage capacity grew almost exactly in step with production for two decades. From 2000 to 2019, capacity was added at an average of 349 million bushels a year, against production growth of roughly 340 million bushels a year [1]. The system stayed in balance because it was built to.

From 2020 through 2025, total capacity added was 337 million bushels — not per year. In total. Over six years [1]. That is an average near 56 million bushels a year, roughly 84% below the prior two decades' rate (Crop Root Zone calculation from [1]).

1. The build rate, and what it cost the system

Source: Joe Janzen, "US Grain Storage Capacity Growth Has Stopped," farmdoc daily, University of Illinois, February 9, 2026 [1]. The 2020–2025 annual average and the percentage decline are Crop Root Zone's own calculation from the stated totals; the 2000–2019 implied total is the stated annual average carried across the period.

Capacity stood at just over 25 billion bushels in 2019 and approximately 25.3 billion in 2025 [1]. Had the earlier build rate continued, capacity would be near 27.5 billion bushels [1].

Source: farmdoc daily, February 9, 2026 [1]. Bar column scaled to the trend figure.

A 2.2-billion-bushel shortfall against trend is not a rounding error in a system this size — it is close to a tenth of the whole national capacity, and it accumulated quietly, one year of not-building at a time.

5%

The 2025 surplus of storage capacity over production. The post-2000 average is 15%. (farmdoc daily, University of Illinois, February 9, 2026)

2. The cushion is the number that matters, not the capacity

Total capacity is the wrong headline because production also grows. The useful measure is the ratio of capacity to production — the cushion the system carries.

That cushion averaged 15% across the post-2000 period. In 2025 it was 5% [1].

The difference between a 15% cushion and a 5% cushion is not a difference in how much grain can be stored. It is a difference in how much the system can absorb before it has to improvise. A 15% buffer absorbs a big crop, a slow export program and a farmer's decision to hold for a better basis, all at once. A 5% buffer absorbs one of them.

What's new is that the constraint is now structural rather than seasonal. A tight harvest is normal and has always been managed with temporary storage. A tight harvest against a capacity base that has not grown for six years is a different condition, because the improvisation is no longer the exception — it is the design.

Evidence is in the utilization figures. At the December 1, 2025 count, on-farm storage was 80% utilized and off-farm (commercial) storage 65% [1]. Those look comfortable, and they are the reason the constraint is invisible to most of the supply chain most of the year. December 1 is after the harvest peak has cleared. The binding moment is six to eight weeks earlier, when the crop arrives faster than it can be shipped.

3. Why the constraint appears as price, not as a shortage

Grain does not go unstored. It goes into bunkers, ground piles, bags and outdoor stacks, and that is exactly what has been absorbing the gap.

This matters because it changes where a grower experiences the problem. A storage shortage that is solved by improvisation shows up as:

Source: Channel characterisation is Crop Root Zone's own read; the underlying capacity and utilization data are from farmdoc daily, February 9, 2026 [1].

The last row is the one that compounds. On-farm storage is not primarily a logistics asset — it is a marketing option. A grower with bins can decline the harvest basis and wait. A grower without them sells into the weakest basis of the year, every year. When national capacity stops growing while production does not, that option gets scarcer and more valuable at exactly the moment it is most needed.

Ground Truth: The temptation is to read the flat build rate as a warning that grain will have nowhere to go. It will have somewhere to go — it always does, into bunkers and piles — and that is precisely why this is worse than it looks rather than better. A capacity constraint that is absorbed by temporary storage does not announce itself as a shortage; it settles quietly into harvest basis and commercial storage rates, where it is easy to attribute to something else. The six years of not building have already been paid for, in basis, by growers who mostly experienced it as an ordinary bad harvest basis. The 5% cushion is the arithmetic behind that experience, and it is the reason a grower comparing this year's harvest basis to a five-year average is comparing against a period in which the same erosion was already underway.

4. Growers own most of the capacity and less of the slack

National capacity splits roughly 13.6 billion bushels on-farm against 11.9 billion off-farm [2] — so growers collectively own about 53% of the storage in the country. That sounds like control. Applied to the December 1, 2025 utilization figures, it is less than it sounds.

Source: Capacity split from National Corn Growers Association, August 2025 [2]; utilization from farmdoc daily, February 9, 2026 [1]. Occupied, free and share-of-free columns are Crop Root Zone's own calculation. The 25.5 bn bu total from the split runs slightly above the ~25.3 bn bu capacity figure in [1]; the two are different vintages and the difference does not change the shares materially.

Growers own 53% of the capacity and hold 39% of the free space, because their bins run fuller. The commercial system, at 65% utilization, is carrying the majority of the country's genuine slack.

That asymmetry decides who has options. A grower whose own bins are 80% full at the December count has already committed most of their marketing flexibility for the year; the remaining choice is to sell or to pay someone else to store. The elevator with 35% of its space open has the flexibility — and prices it.

Ground Truth: The storage debate is usually framed as growers versus elevators competing for the same scarce space. The utilization split says something more specific: the scarce space is disproportionately on the farm, and the party with the flexibility is the one charging for it. That is not a criticism of elevators, who carry the working capital and the throughput obligation. It is an argument about where a marginal dollar of storage investment does the most good for a grower's own marketing position — and on these numbers, it is on-farm, precisely because the on-farm segment is the tighter one. It is also the segment where capacity has been hardest to add, since a farm bin is financed against one operation's balance sheet rather than a cooperative's.

5. Why nobody is building, and what would change it

The build-rate collapse is not a mystery and it is not primarily about steel prices. Permanent storage is a long-lived capital asset financed against expected returns, and the returns to storage are the carry — the spread between nearby and deferred prices that pays a holder to wait.

Three conditions have worked against it simultaneously:

1. Soft commodity prices. A bin is easiest to justify when the crop being stored is worth more.
2. High old-crop stocks. Existing space that stays full does not free up to justify new space; it just stays full.
3. A narrowing construction window. Permanent storage has to be built between harvest and the next planting, by contractors who are themselves capacity-constrained. A decision made late in the year cannot be executed for that year's harvest.

The third is the underappreciated one, and it is why this does not self-correct quickly. Even a decisive turn in the economics of storage cannot add meaningful permanent capacity for the crop currently in the field, or realistically for the one after it. The lag between deciding to build and having a bin is longer than the price signal that would justify it.

What to watch, in order of how early it would tell you something:

• Commercial storage rate announcements ahead of harvest. These are posted, they are public, and they move before basis does.
• Free days on delayed price contracts. A quiet reduction here is a capacity signal that never appears as a price.
• The December 1 utilization split. If on-farm utilization climbs above the 80% recorded for 2025 while off-farm stays near 65%, the burden is shifting onto growers' own bins — which is the group with the least ability to add more.
• Permanent construction starts, reported through the winter. This is the only one of the four that would indicate the constraint is actually being resolved rather than absorbed.

References

1. Janzen, Joe. "US Grain Storage Capacity Growth Has Stopped." farmdoc daily, Department of Agricultural and Consumer Economics, University of Illinois, February 9, 2026. Total capacity just over 25 bn bu (2019) and approximately 25.3 bn bu (2025); 337 mn bu added 2020–2025 against a 2000–2019 average of 349 mn bu/yr; trend-continuation capacity of 27.5 bn bu; December 1, 2025 utilization of 80% on-farm and 65% off-farm; capacity-to-production surplus of 15% average post-2000 versus 5% in 2025. https://farmdocdaily.illinois.edu/2026/02/us-grain-storage-capacity-growth-has-stopped.html
2. National Corn Growers Association. "A Look at Current U.S. Grain Storage Capacity." The Corn Economy, August 2025. Cited for the on-farm and off-farm split of national capacity. https://ncga.com/stay-informed/media/the-corn-economy/article/2025/08/a-look-at-current-u-s-grain-storage-capacity

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Between June 30 and July 11, 2026, the US Environmental Protection Agency completed registration actions on six herbicides, and the Weed Science Society of America welcomed the move as helping weed scientists "develop and implement safe and effective weed management programs with greater diversity and sustainability across many crops" (WSSA, Jul 2026). Three of the six — diflufenican, epyrifenacil and trifludimoxazin — carry tolerances for corn and soybean uses, published as final rules in the Federal Register on June 30, 2026 (Texas Agriculture / Texas Farm Bureau, Aug 7, 2026; Georgia Farm Bureau, 2026).

For a grower building a 2027 resistance-management plan, the number that matters is not six. Resistance is selected against a site of action, not against a molecule, and a rotation plan buys diversity only to the extent that the products in it attack different enzymes. Sorted that way, the announcement is a quarter of its headline size.

1. The six, sorted by site of action and by age

Sources: herbicide list and Group 14 / Group 27 assignments, WSSA, Jul 2026; diflufenican Group 12 / HRAC F1 assignment, WSSA classification of herbicide site of action; isoxaflutole 1998 conditional registration for field corn and topramezone 2005 registration, EPA registration fact sheets; florpyrauxifen-benzyl 2018 commercial release in rice, published weed-science literature. Table construction is Crop Root Zone's.

Three arithmetic facts fall straight out of it.

Six molecules, four sites of action. Groups 12, 14, 4 and 27, with Group 14 and Group 27 each represented twice. The molecule-to-site ratio is 0.67.

Three of the six are not new chemistry at all. Isoxaflutole has been a registered US herbicide since 1998 and topramezone since 2005 — 28 and 21 years respectively. Florpyrauxifen-benzyl reached commercial use in rice in 2018, eight years ago. The average age of those three is 19 years. What EPA acted on for them is new uses and tolerances, which is a real and useful regulatory outcome, but it is not new chemistry and it adds no site of action that did not already exist in a Midwest weed-control program.

Among the genuinely novel active ingredients, there are two sites of action, not three. Epyrifenacil and trifludimoxazin are both Group 14 PPO inhibitors (WSSA, Jul 2026; peer-reviewed characterisation of epyrifenacil as a systemic PPO-inhibiting herbicide, Pest Management Science, 2025). Only diflufenican sits outside — a Group 12 PDS inhibitor of the pyridinecarboxamide family.

Bars scaled so six is ten blocks. Crop Root Zone analysis of the sources above.

2. Why Group 14 is a different proposition from Group 12

Both new PPO molecules arrive into a site of action that Midwest pigweed populations have been selected against for two decades, and where the resistance is not merely present but mapped at the amino-acid level.

Target-site resistance to PPO inhibitors has been described mainly in broadleaf weeds through mutations in protoporphyrinogen oxidase 2. In Palmer amaranth, the Gly210 deletion is the most common, with Arg128Gly and Gly399Ala substitutions also common (peer-reviewed PPO2 inhibition-profile work, 2022).

The manufacturers' answer is direct and, on the published evidence, substantive: trifludimoxazin is a triazinane-type PPO inhibitor described as controlling PPO-resistant weed populations, and epyrifenacil is reported to have controlled all three of the Palmer amaranth biotypes named above. That is a genuine control claim supported by target-site work, and it should not be waved away.

Diflufenican is a different kind of arrival. It is described as the first effective PDS inhibitor herbicide for control of waterhemp, Palmer amaranth and other pigweed species in corn and soybean — a site of action those populations have simply never encountered in these crops (WSSA site-of-action classification; published characterisations of diflufenican).

1 of 6

Sites of action in the June–July package that are new to pigweed control in corn and soybean. (Crop Root Zone analysis of WSSA and EPA source material, Jul–Aug 2026.)

Ground Truth: The ranking a grower would draw from the control claims is the reverse of the one a resistance plan should use. The two molecules with the most impressive data — the PPO inhibitors that control known resistant biotypes — are the two entering a population that has already demonstrated it can generate PPO2 mutations, has been under selection pressure since the early 2000s, and carries the standing genetic machinery to do it again. Controlling ΔGly210, Arg128Gly and Gly399Ala is a statement about three mutations that already exist. It is not a statement about the fourth. Diflufenican has the least dramatic claim attached to it and is the only one of the three arriving at an enzyme that Midwest pigweed has no selection history against — which makes it the most valuable of the three to a rotation and the one most worth protecting from being used every year on every acre. The molecule that needs the discipline is the one whose marketing will least obviously demand it.

3. What the sorting is worth in practice

The operational implication is a sequencing question, and it has a clean answer.

A resistance-management program is usually specified as a number of effective sites of action per season — commonly two or more on the same weed, applied so that a survivor of one is exposed to the other. Adding a molecule to a group already in the rotation does not raise that count. It raises the number of products available at that count, which is a supply and cost benefit, and occasionally a spectrum or timing benefit, but not a diversity benefit.

Constructed by Crop Root Zone from the source material cited above. Product positioning reflects published descriptions, not label recommendations; the label governs.

Epyrifenacil is described as a fast-acting systemic PPO inhibitor designed for preplant burndown and cover-crop termination, and its tolerance covers field corn (forage, grain, stover), soybean (forage, hay, seed), rapeseed and wheat. Trifludimoxazin covers both pre-emergent and burndown timings. Both are useful; both are Group 14.

Ground Truth: There is a second-order consequence that argues for adopting the two PPO molecules anyway, and it runs opposite to §2. Cover-crop termination and preplant burndown in the eastern Belt have leaned heavily on glyphosate and on a narrow set of PPO products, and a burndown failure is what pushes a grower into a rescue application later in the season — which is itself a major driver of selection pressure. A more reliable burndown that controls existing resistant biotypes reduces the number of rescue passes, and rescue passes are disproportionately where resistance gets selected because they are applied to larger weeds at sub-lethal effective doses. So the two Group 14 molecules can improve resistance outcomes without adding a site of action, by removing the failures that generate the selection events. That is a real benefit and it is invisible in a site-of-action count — which is a caution against reading §2's arithmetic as the whole of the decision.

4. What this piece does not settle

Registration is not availability. EPA action on tolerances and registrations is a step toward commercial launch, not a launch. None of the three new active ingredients is established here as being on a retailer's shelf for 2027, and the sources reviewed do not give launch dates. A grower should not build a 2027 program around a product with no confirmed commercial supply.

Site-of-action group is not the whole of resistance risk. Cross-resistance, metabolic (non-target-site) resistance, and enhanced metabolism can defeat a molecule at an enzyme the population has never seen. Group 12's novelty to pigweed is a strong prior, not a guarantee, and metabolic resistance in waterhemp in particular has repeatedly surprised programs built on target-site reasoning.

The PFAS question is open and is a genuine regulatory exposure. Several of these approvals have been characterised in legal and advocacy commentary as involving compounds that may or may not meet a PFAS definition depending on which definition is applied (Goldberg Segalla, 2026; Center for Biological Diversity, Jul 1, 2026). This desk takes no position on the chemistry question. It is noted because definitional disputes of this kind have preceded label restrictions before, and a product's regulatory durability is part of what a grower is buying when they build a rotation around it.

The 1998 and 2005 dates describe first registration, not stasis. Isoxaflutole and topramezone have had label expansions, formulations and trait-system pairings in the intervening decades. Calling them "not new" is a statement about site of action, which is the only sense that matters for the argument here — not a claim that nothing about them has changed.

References

1. Weed Science Society of America — "WSSA Applauds EPA's Advancement of Six New Herbicides," Jul 2026. https://wssa.net/2026/07/wssa-applauds-epas-advancement-of-six-new-herbicides/
2. Texas Agriculture / Texas Farm Bureau — "EPA proposes registration for new crop protection chemistries," Aug 7, 2026. https://texasagriculture.texasfarmbureau.org/publication/?i=868350&article_id=5190437
3. Georgia Farm Bureau — "EPA proposes registration for new crop protection products," 2026. https://www.gfb.org/news/ag-news/post/epa-proposes-registration-for-new-crop-protection-products
4. Weed Science Society of America — "Classification of Herbicide Site of Action." https://www.weedscience.org/Documents/ShowDocuments.aspx?DocumentID=1193
5. Sada et al. — "Epyrifenacil, a new systemic PPO-inhibiting herbicide for broad-spectrum weed control," Pest Management Science, 2025. https://scijournals.onlinelibrary.wiley.com/doi/full/10.1002/ps.8244
6. "Inhibition profile of trifludimoxazin towards PPO2 target site mutations," Pest Management Science / PubMed, 2022. https://pubmed.ncbi.nlm.nih.gov/36178376/
7. US EPA — Pesticide Fact Sheet for Isoxaflutole, Sep 15, 1998. https://www3.epa.gov/pesticides/chem_search/reg_actions/registration/fs_PC-123000_15-Sep-98.pdf
8. US EPA — Pesticide Fact Sheet for Topramezone, Aug 10, 2005. https://www3.epa.gov/pesticides/chem_search/reg_actions/registration/fs_PC-123009_10-Aug-05.pdf
9. Goldberg Segalla — "Definition Critical: EPA Approves Three New Pesticides that May — (or May Not) — be PFAS," 2026. https://www.goldbergsegalla.com/blog/environmental-law-monitor/pfas/definition-critical-epa-approves-three-new-pesticides-that-may-or-may-not-be-pfas/
10. Federal Register — "Pesticide Product Registration; Receipt of Applications for New Uses," Aug 17, 2026. https://www.federalregister.gov/documents/2026/08/17/2026-16721/pesticide-product-registration-receipt-of-applications-for-new-uses-may-2026

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Switch Bioworks has begun advanced field trials of an engineered microbial nitrogen fertilizer on corn across multiple US Midwest sites, under authorisations from both USDA and EPA. What makes the programme worth a column is not the trial — every entrant in this category runs trials — but the way the company frames its own technology: a genetically encoded switch addressing what it calls a fundamental energy problem that has long limited microbial fertilizer performance.

That framing is correct, and it is the first time in several years of covering this category that this desk has seen a developer name the actual constraint out loud. Biological nitrogen fixation is not free. It is one of the most expensive reactions a living cell performs, and something has to pay for it. In a corn field, that something is photosynthate. This piece prices the transaction.

We have covered the other half of this question before. In July we set the replicated yield evidence for gene-edited nitrogen-fixing microbes — 2 bu/acre on average across nitrogen rates in Agronomy Journal work from Illinois, 4 bu/acre at moderate rates, 11 bu/acre in a Kentucky strip trial after a cover crop — against the retail nitrogen deck, and showed the product returns very differently depending on which nitrogen you actually buy. That was the price break-even. This is the carbon break-even, and it is the one that explains why the yield numbers keep coming in where they do.

1. What nitrogenase costs, in the units biochemistry uses

What's new: Nothing about the stoichiometry is new. What is new is applying it to a purchasing decision.

Evidence: The canonical stoichiometry for nitrogenase is 16 ATP consumed per molecule of N₂ reduced to two molecules of ammonia — eight ATP per NH₃. That is the textbook minimum under ideal conditions. In living systems the realised cost is higher: published figures put in-vivo consumption at roughly 20 to 30 ATP per N₂, because nitrogenase also reduces protons to hydrogen gas as an obligate side reaction, wasting a substantial share of the electron flux, and because the enzyme is irreversibly damaged by oxygen and must be protected at further metabolic expense.

For comparison, essentially every other nutrient a plant takes up costs it a transporter and an ion gradient. Nitrogen fixation costs it a dedicated metalloenzyme, an oxygen-exclusion strategy, and an ATP bill roughly two orders of magnitude larger per atom.

Biochemists convert that ATP bill into carbon, because carbon is what the cell actually spends. The best-characterised measurements come from legume nodules, where the system can be isolated and respiration measured directly:

Source: Respiratory/Carbon Costs of Symbiotic Nitrogen Fixation in Legumes, in Nitrogen Fixation in Agriculture, Forestry, Ecology and the Environment (Springer).

The wider range is the honest one to use for a product decision, because a grower is buying the whole organism, not the enzyme.

2. Converting carbon into bushels

What's new: The conversion is arithmetic, and to this desk's knowledge nobody selling into this category publishes it.

Evidence: A bushel of corn is 56 lb at 15.5% moisture, so 47.32 lb of dry matter. Corn grain runs about 45% carbon by dry weight — starch alone is 44.4% carbon and the oil fraction is higher. That puts roughly 21.3 lb of carbon in a bushel of corn.

Now take a microbial nitrogen product delivering 25 lb N/acre, which is at the upper end of what the category typically claims:

Crop Root Zone calculation. Carbon costs from the Springer legume figures above; grain carbon at 45% of 47.32 lb dry matter per bushel. This is a carbon-equivalence calculation, not a yield forecast — see section 4.

3.3 – 5.7 bushels

The grain-carbon equivalent of fixing 25 lb N/acre biologically, against replicated field evidence of roughly 2 bu/acre of yield gain from commercial microbial nitrogen products. (Crop Root Zone calculation from Springer carbon-cost figures; yield evidence from Agronomy Journal Illinois work, as reported Jul 27, 2026)

Ground Truth: Put those two numbers beside each other and the category's persistent disappointment stops looking like a marketing problem and starts looking like an accounting identity. The carbon required to fix a commercially meaningful quantity of nitrogen is larger than the yield gain the products demonstrate. That does not make the products worthless — see the next section, because the carbon is not all incremental — but it does mean the upside is bounded by something no amount of genetic engineering negotiates with. A buyer who has been waiting for the breakthrough version that delivers 60 lb N/acre should understand what 60 lb N/acre would cost the plant.

3. Why the products keep landing at 20-25 pounds

What's new: The convergence of independent products on a similar delivered-nitrogen claim is usually read as vendors copying each other's marketing. It is better read as several companies independently discovering the same ceiling.

Evidence: Products in this category have consistently been positioned around delivering roughly 20-25 lb N/acre. BioConsortia has reported wheat yields preserved with nitrogen rates cut by 50% in field trials — a claim about substitution rather than an absolute pound figure, but one that implies a similar order of magnitude on typical wheat programmes.

Consider what the rhizosphere can actually afford. A corn crop directs a modest single-digit-to-low-double-digit percentage of its net fixed carbon belowground as rhizodeposition — exudates, sloughed cells, mucilage — and that pool has to support the entire soil microbial community, of which diazotrophs are a small minority. At 2.78-4.81 g C per g N, supporting 25 lb of fixed nitrogen requires 70-120 lb of carbon routed specifically to organisms doing the fixing. That is a large claim on a shared and finite budget.

Ground Truth: The engineering problem Switch describes — a switch that decides when the microbe fixes — is real and worth solving, because a diazotroph that fixes nitrogen in the presence of abundant soil nitrate is wasting carbon for nothing, and one that shuts down permanently is useless. But regulation and thermodynamics are different constraints, and only one of them is negotiable. A better switch improves when the carbon is spent and how little is wasted. It does not change the price per pound. Any developer promising a step change in delivered nitrogen, rather than a step change in efficiency and consistency, is promising something the chemistry does not obviously permit.

4. The limitation, stated properly, because it matters

What's new: The calculation above is a ceiling argument. Read as a yield deduction it would be wrong, and this desk would rather say so than let a clean number do work it cannot support.

Evidence: Three qualifications, and the first is the important one.

The carbon is not all incremental. Rhizodeposition happens whether or not a diazotroph is present. A corn plant releases carbon into the rhizosphere as a matter of course, and a large share of it is consumed by soil organisms that give nothing back. If a nitrogen-fixing microbe captures carbon that was already leaving the plant, the marginal cost to yield is far below the gross figure — the plant is not spending more, it is getting something for what it was already spending. That is precisely the case for the category, and it is a legitimate one. The honest way to state it: the gross carbon cost is 70-120 lb for 25 lb N, and the net cost to the crop is that figure multiplied by the fraction of fixation carbon that represents genuinely additional draw on photosynthate. Nobody has published that fraction. It is the single number that would settle whether these products can work at scale, and it is missing.

The legume numbers may not transfer, and the direction of the error favours our argument. The 2.78-4.81 g C/g N range is measured in legume nodules, which are generally regarded as the efficient end of biological fixation: the plant supplies carbon directly, and leghemoglobin manages oxygen so the enzyme survives. Associative and endophytic diazotrophs on cereals have no comparable infrastructure. If cereal-associated fixation is less carbon-efficient than nodulated symbiosis — which is the expectation, not a certainty — the true cost is higher than the range we used. We have therefore taken the conservative figures, and the conclusion would only strengthen with better cereal-specific data.

Nitrogen is not the only thing these organisms do. Several products in the category also claim phosphorus solubilisation, hormonal effects on root architecture, or stress tolerance. A yield response attributed to fixed nitrogen may partly be something else, in which case the nitrogen accounting understates the product and overstates the mechanism. This cuts against our framing and we note it.

5. What this changes about buying

What's new: The carbon ceiling gives a buyer a screening rule that does not require waiting for another trial season.

Evidence: Price the alternative. On DTN's week-of-August-10 retail sheet, relayed August 19, the nitrogen deck stands as follows:

Source: DTN/Progressive Farmer, "5 Fertilizer Prices Slightly Lower Than Last Month; 3 Higher," Aug 19, 2026. $/lb N is a Crop Root Zone calculation from the posted price and standard analysis.

A microbial product priced at $15-20/acre must therefore displace 25.5 to 34.0 lb N/acre to break even against anhydrous, and 18.8 to 25.1 lb against UAN28. The anhydrous figure sits at or above the top of what the category claims to deliver — before any question of whether the displaced pounds actually show up in a dry August.

Scaled to $20/acre = ten blocks. Crop Root Zone calculation from the DTN deck above.

The screening rule follows directly: treat any claim above roughly 30 lb N/acre as requiring extraordinary evidence, because it implies a carbon commitment of 83-144 lb — four to seven bushels' worth — routed to a minority population in a shared rhizosphere. Ask a vendor two questions. First, what fraction of the fixation carbon is incremental draw on photosynthate rather than exudate that was leaving anyway? Second, on what nitrogen price was the product's return calculated? A return computed against UAN28 at $0.796/lb is 35% more flattering than the same return against anhydrous, and this desk has repeatedly found the marketing arithmetic runs off the expensive end of the deck.

Neither question asks a company to disclose anything proprietary about its organism. Both are answerable from work a serious programme would already have done. Their absence from the public record, across a category now more than a decade old, is itself informative.

References

1. World Fertilizer, "Microbial fertilizer enters field trials," Aug 7, 2026. https://www.worldfertilizer.com/environment/07082026/microbial-fertilizer-enters-field-trials/
2. Fertilizer Daily, "Switch Bioworks launches U.S. field trials for engineered microbial nitrogen fertilizer," Jun 4, 2026. https://www.fertilizerdaily.com/20260604-switch-bioworks-launches-u-s-field-trials-for-engineered-microbial-nitrogen-fertilizer/
3. "Respiratory/Carbon Costs of Symbiotic Nitrogen Fixation in Legumes," in Nitrogen Fixation in Agriculture, Forestry, Ecology and the Environment, Springer. https://link.springer.com/chapter/10.1007/1-4020-3589-6_11
4. "Enabling Biological Nitrogen Fixation for Cereal Crops in Fertilized Fields," ACS Synthetic Biology. https://pubs.acs.org/doi/10.1021/acssynbio.1c00049
5. BioConsortia, field trial results releases (nitrogen-fixation programme). https://www.bioconsortia.com/news/bioconsortia-releases-latest-field-trial-results
6. DTN/Progressive Farmer, "5 Fertilizer Prices Slightly Lower Than Last Month; 3 Higher," Aug 19, 2026.
7. Crop Root Zone, TRZ-0041, "Microbial Nitrogen Finally Has Its Numbers. Cheap Anhydrous Just Moved the Goalposts," Jul 27, 2026.
8. Crop Root Zone, TRZ-0091, "The Biostimulant Evidence Gap Is a Labeling Rule, Not a Science Problem," Aug 12, 2026.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Boron is the fastest-growing segment of the micronutrient fertilizer market, projected to expand at a 7.5% compound annual rate through 2031 against zinc's larger but slower base, and roughly 31% of the world's agricultural soils are reported boron-deficient. In North America about 35% of soils show zinc or boron shortfalls. The demand case is straightforward and this desk does not dispute it.

The application case is not straightforward at all, and it is almost never discussed in the terms that decide it. Boron has, by a wide margin, the narrowest gap between the rate that corrects a deficiency and the rate that causes injury of any nutrient a grower handles. It is also applied at the lowest rate of any product a spreader is asked to distribute. Those two facts are in direct mechanical conflict, and this piece quantifies the conflict.

This desk has argued before, on zinc, that the material price decides almost nothing about a micronutrient decision — that the argument is entirely about which critical soil-test level you believe. Boron is a further step down that road. Here even the rate decides less than the placement.

1. The window, and why it is genuinely unusual

What's new: Nothing in the thresholds is new. What is underappreciated is how close together they sit.

Evidence: Published guidance is consistent across sources and unusually specific about upper bounds:

Source: extension guidance summarised from University of Delaware Weekly Crop Update; Michigan State University Extension; Wisconsin "Soil and Applied Boron" (A2522); Lakehead University extension article; and Boron Toxicity and Deficiency in Agricultural Plants (PMC7073067).

Set the typical corrective rate of about 1 lb B/acre against a soil-application ceiling of 2 lb and the margin is a factor of two. There is no other nutrient in the standard programme where that is true. A grower who over-applies potassium by a factor of two has wasted money. A grower who over-applies nitrogen by a factor of two has wasted money and possibly lodged the crop. A grower who over-applies boron by a factor of two is at the published injury threshold.

Three further pieces of guidance follow directly from the narrowness and are worth reading as a set, because together they describe a nutrient the agronomists do not entirely trust the equipment to handle. Band application is not recommended. Broadcast application is advised one to two weeks ahead of planting rather than at planting. And extension sources recommend blending boron with other fertilizers applied at seeding specifically to ensure uniform application — that is, the blending is prescribed as a distribution remedy, not as a convenience.

2. What one pound per acre physically is

What's new: The particle count. To this desk's knowledge nobody publishes it, and it is the number that makes every other piece of boron advice make sense.

Evidence: Take a common granular boron source at roughly 15% B, applied to deliver 1.0 lb of actual boron per acre. That requires about 6.7 lb of product per acre — 3,025 grams spread over 43,560 square feet.

Granular fertilizer products are typically manufactured in the 2-4 mm size range, at a bulk particle density on the order of 1.7 g/cm³. Work the arithmetic through for a spherical granule:

Crop Root Zone calculation. Assumes 15% B product at 1.0 lb B/acre, spherical granules at 1.7 g/cm³. *At a 34,000 plants/acre corn population. Granule geometry is idealised; real granules are irregular and size-distributed, so these are order-of-magnitude figures, not precision counts.

~3.7 granules

The number of boron granules landing in the ground area belonging to a single corn plant, at a 1 lb B/acre broadcast rate, 3 mm granules and a 34,000 population. (Crop Root Zone calculation — assumptions stated above)

Ground Truth: At single-digit granules per plant, boron is not being spread; it is being placed at points. Every intuition a grower has about broadcast application — that the spreader averages things out, that a 15% coefficient of variation on the spread pattern is acceptable because neighbouring passes overlap — depends on there being enough particles per unit area for the law of large numbers to operate. At three to four particles per plant it does not operate. The variance a plant experiences is not the spreader's pattern variance; it is the variance of a handful of discrete events. A plant that catches eight granules instead of four has received twice the intended rate, and twice the intended rate is the published injury threshold. This is the mechanical reason the "hot spot" language appears in boron literature and almost nowhere else.

3. Why the industry's answer is impregnation, not calibration

What's new: The standard remedy is usually presented as a product-quality preference. It is better understood as the only available fix for a particle-count problem.

Evidence: Published work and manufacturer guidance both point the same direction. Boronated NPK fertilizers — with the boron source incorporated at the factory — deliver more uniform application than most bulk blends. Research comparing potash carrying incorporated boron (MOP+B) against traditional bulk blends found more uniform boron spatial distribution from the incorporated product. The mechanism described in the literature is exactly the one above: because boron is required in small amounts, delivering it evenly is a challenge, and blending small quantities with P and K products produces uneven application and boron hot spots that are prone to leaching.

The reason impregnation works is arithmetic rather than chemistry. Coating or incorporating boron across every granule in the blend converts a few large point sources into hundreds of thousands of small ones. If a 200 lb/acre blend carries its boron distributed across all its granules at 3 mm, the particle count driving boron placement rises from about 126,000 per acre to roughly 3.8 million — a thirty-fold improvement in the statistics, achieved without touching the spreader.

Crop Root Zone calculation on the assumptions in section 2, 3 mm granules. The second row assumes uniform incorporation across the carrier, which is the manufacturing claim, not a measured result.

Calibrating the spreader more carefully cannot fix this. Spreader calibration controls the rate and the pattern; it has no effect at all on how many particles carry the nutrient. This is the practical distinction a buyer should take from the section: for boron, product form is a bigger lever than application accuracy, and it is the cheaper of the two to change.

4. The crop that pays for the error is next year's

What's new: The rotation is where a boron mistake becomes expensive, and it is a year removed from the decision that caused it.

Evidence: Soybean and dry beans are described in extension guidance as highly sensitive to boron toxicity, to the point that boron fertilizer application to soybeans is advised against outright. Meanwhile, published reports on corn and soybean response to pre-plant boron are inconsistent — the upside in a corn-soybean rotation is unreliable while the downside in the bean year is well characterised.

That asymmetry is the whole decision. Consider the sequence in a standard corn-soybean rotation:

Source: Michigan State University Extension; University of Tennessee UT Crops News; University of Delaware Weekly Crop Update.

Ground Truth: A boron application is not a corn decision that happens to precede soybeans. It is a soybean decision taken a year early, by someone thinking about corn. The corn crop is tolerant enough to absorb the placement error the equipment inevitably creates, which means the error is invisible in the year it is made and diagnosed — if at all — as a patchy stand in a crop nobody applied boron to. Any grower running a corn-soybean rotation should evaluate a boron programme against the bean year's toxicity risk, not the corn year's response probability, because that is where both the sensitivity and the evidence are strongest.

5. What a buyer should actually ask

The screening questions are short and none of them concern price.

What form is the boron in? If the answer is a straight granular boron product bulk-blended at the plant, the particle-count problem in section 2 is live and no amount of spreader attention addresses it. Incorporated or boronated carriers are the structural fix.

What is the field's rotation? Continuous corn tolerates a placement error that a corn-soybean rotation carries into a sensitive crop. The same 1 lb/acre is a materially different risk in the two systems.

What did the soil test actually say, and against whose critical level? This desk found on zinc that published critical soil-test levels span a fourteenfold range across state guides, so the same sample produces opposite recommendations depending on whose table is used. Boron's soil test is not obviously better behaved, and the consequences of a false positive are worse, because the corrective action has a toxicity threshold at twice the rate.

Is foliar the better route here? The published foliar ceiling is 0.5 lb B/acre — a quarter of the soil ceiling — but foliar application dissolves the particle-count problem entirely, because the nutrient arrives in solution rather than as granules. That is a genuine advantage that the lower rate partly offsets, and it deserves to be evaluated on the placement question rather than dismissed on the rate one.

What a buyer should not do is treat boron as a cheap add-on to an existing blend because the material cost is trivial. The material cost is trivial. That has never been the constraint, and on this nutrient the thing that is cheap to buy is also the thing that is hard to place.

References

1. University of Delaware, Weekly Crop Update, "Boron Deficiencies and Toxicities in Vegetable Crops." https://sites.udel.edu/weeklycropupdate/?p=8121
2. "Boron Toxicity and Deficiency in Agricultural Plants," PMC7073067. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7073067/
3. University of Wisconsin, "Soil and Applied Boron" (A2522). http://corn.agronomy.wisc.edu/Management/pdfs/a2522.pdf
4. Michigan State University Extension, "Risks and benefits of applying boron fertilizer to soybeans." https://www.canr.msu.edu/news/risks_and_benefits_of_applying_boron_fertilizer_to_soybeans
5. University of Tennessee, UT Crops News, "Does boron fertilizer use increase corn and soybean yields?" https://news.utcrops.com/2020/04/does-boron-fertilizer-use-increase-corn-and-soybean-yields/
6. Silva et al., "Slow and Fast-Release Boron Sources in Potash Fertilizers: Spatial Variability, Nutrient Dissolution and Plant Uptake," Soil Science Society of America Journal, 2018. https://acsess.onlinelibrary.wiley.com/doi/full/10.2136/sssaj2018.02.0065
7. Lakehead University, "Boron the Mighty Micronutrient," Dr. Tarlok Singh Sahota, CCA.
8. Mordor Intelligence, micronutrient fertilizer market reports, 2026 (segment growth rates and deficiency incidence).
9. Crop Root Zone, TRZ-0076, "The Zinc Costs $6 an Acre and Lasts Three Years. The Argument Is Entirely About Which Critical Level You Use," Aug 7, 2026.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

The economics of a nitrification inhibitor are usually argued qualitatively — it protects the nitrogen, it reduces loss, it is cheap relative to a lost crop. The arithmetic is more interesting than the argument, and at current nitrogen prices it lands almost exactly on a knife edge.

Nitrapyrin-type treatment is commonly costed around $15–20 an acre, a figure widely cited as an adoption barrier for operations with tight margins [3]. University work on enhanced-efficiency fertilizers has suggested that inhibitors "could potentially reduce nitrogen rates 15–20%" while holding yield [2].

Put those two ranges together against what nitrogen actually costs this month, and the product very nearly pays for itself — and no more.

1. What a pound of nitrogen costs right now

DTN's early-August retail averages give three carriers with very different unit economics once converted to a nitrogen basis.

Source: DTN/Progressive Farmer retail fertilizer averages, early August 2026. The $/lb N conversion is Crop Root Zone's own: price per ton ÷ 2,000 ÷ nutrient fraction.

Anhydrous remains the cheapest nitrogen by a wide margin — $0.5872 a pound against $0.7370 for urea, a 26% premium for the granular product. That spread is the reason the same inhibitor decision produces different answers depending on what is in the tank.

Source: Crop Root Zone calculation from DTN/Progressive Farmer retail averages, early August 2026.

2. The break-even, run three ways

Take a 180-pound-per-acre nitrogen program — a reasonable central case for Corn Belt corn, and stated here as an assumption rather than a recommendation, since the right rate is a soil-test and yield-goal question this publication is not positioned to answer for any individual field.

Source: Crop Root Zone calculation. Nitrogen prices from DTN/Progressive Farmer retail averages, early August 2026; the 15–20% rate-reduction range from NC State University research [2]; the $15–20/acre inhibitor cost as commonly reported [3]. The 180 lb N/acre program rate is an assumption, not a recommendation. Net range spans worst case (15% cut, $20 inhibitor) to best case (20% cut, $15 inhibitor).

Read the last column carefully, because it is the finding. On anhydrous, the worst case is a loss of $4.15 an acre and the best case a gain of $6.14 — a genuine coin flip. On urea and UAN32, the worst case is essentially zero and the best case is $10–11.50 an acre to the good.

$105.70/acre

The nitrogen cost of a 180-pound program on anhydrous at $963/ton — against $15–20 an acre for the inhibitor that might let you cut it by 15–20%. (Crop Root Zone calculation from DTN retail averages, early August 2026)

The result is not a coincidence of this month's prices, and that is what makes it useful. An inhibitor's price is set by what the market believes it saves. What the arithmetic shows is that the product is priced almost precisely at the value of the rate reduction it is claimed to enable — which is what a competitive market for an input should produce, and which means a grower buying it for the rate cut is paying full value for it and capturing none of the surplus.

Ground Truth: If an inhibitor breaks even on the rate reduction alone, then the rate reduction is not the reason to buy it — it is the reason it costs what it costs. The return has to come from the loss-avoidance case, and that case is not an average; it is a tail. NC State's framing is the correct one: EEFs behave as "nitrogen-availability insurance" during risky weather, and in ideal conditions growers "often don't see any yield increase" from them [2]. That is exactly the payoff shape of an insurance product — a small certain cost against a large uncertain loss — and it should be underwritten the way insurance is: by asking how likely the loss is on this field, not by asking whether the product works on average. Wet springs on tile-drained, coarse-textured or fall-applied ground are where the tail lives. Flat, heavy, spring-applied ground is where the $15–20 is most likely to be a straight cost.

3. Why the carrier changes the answer

The reason urea and UAN32 look better than anhydrous in the table has nothing to do with the inhibitor's chemistry. It is entirely the denominator: a percentage rate cut on more expensive nitrogen is worth more dollars.

That produces a mild paradox worth naming. Anhydrous is the cheapest nitrogen and also, on this arithmetic, the carrier where an inhibitor is hardest to justify on rate alone — even though anhydrous is the product most often paired with one, because fall-applied ammonia is the classic loss-risk scenario.

The two facts are not in conflict; they are measuring different things. The rate-reduction case is weakest on anhydrous and the loss-avoidance case is strongest on it, because fall application on ammonia carries the longest exposure between application and uptake. A grower reasoning only from the table would reach the wrong conclusion about the product's most common use.

Source: Crop Root Zone's own read. The rate-reduction column follows directly from the $/lb N table above; the loss-avoidance column reflects the length of the interval between application and crop uptake, which is the period over which nitrification and leaching losses accumulate.

4. The subsidy is not the answer either

Incentive programs have appeared alongside these products — one pilot pays growers $3 an acre to use enhanced-efficiency fertilizers [2][3]. Against a $15–20 cost, that covers 15% to 20% of the price (Crop Root Zone calculation).

That is not nothing, and on the margin it converts some of the negative cells in the table above into positive ones. But it does not change the structure of the decision. A $3 payment against a $15–20 product moves the break-even; it does not move the product from a rate-reduction case to a rate-reduction case that clears comfortably.

The more important observation about the incentive is what it reveals about who wants the outcome. The payment exists because the environmental benefit of reduced nitrogen loss accrues to someone other than the grower. Losses to leaching and to nitrous oxide — a gas with roughly 300 times the warming potential of carbon dioxide [3] — are costs borne downstream and off-farm. A grower capturing only the on-farm value will rationally underinvest relative to what the full social calculation would justify, which is the standard shape of an externality, and a $3 payment is a partial and explicit acknowledgement of it.

The number to watch is the size of that payment relative to the product cost. At $3 against $15–20 it is a nudge. If it rises toward the full cost, the market for these products stops being an agronomic market and becomes a policy market, with the adoption decision driven by the programme rather than by the field.

5. What would change this arithmetic

Three things, in order of how fast they could move:

1. Nitrogen prices. Every one of these break-evens scales directly with the $/lb N line, and all three carriers fell in early August — anhydrous 7%, urea 5%, UAN32 14% from its June peak [1]. Falling nitrogen prices make inhibitors harder to justify on rate, which is the counterintuitive direction and the one currently in force.
2. Better rate-response data. The 15–20% reduction figure is a research estimate framed as a potential, not a measured field average [2], and the underlying literature reports "mixed impact on crop yield" [2]. A narrower, better-replicated number would move the table more than any price change.
3. Inhibitor pricing. $15–20 an acre is a commonly reported range rather than a quoted market. If competition or off-patent supply moves it toward $10, the rate-reduction case clears on every carrier in the table.

A note on the evidence, stated plainly because it bears on how much weight the table can carry: the 15–20% rate-reduction range and the $3/acre incentive figure both come from a 2022 university publication [2], and the $15–20/acre inhibitor cost is a commonly cited range rather than a dated market quote [3]. Only the nitrogen prices are current. The arithmetic is therefore best read as a framework for a grower to run with their own quoted inhibitor price, not as a settled answer — and the framework is the point, because the quote a grower is actually offered is the one number in this piece that no publication can source for them.

References

1. DTN/Progressive Farmer. "Fertilizers Start August With Mixed Prices as 5 Nutrients Decline, Led by UAN32," August 12, 2026, and "Fertilizer Prices Fall as UAN32 Leads Significant Drops Heading Into August," August 5, 2026. Retail averages: anhydrous $963/ton (−7% MoM), urea $678/ton (−5% MoM), UAN32 $458/ton. https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/12/fertilizers-start-august-mixed-5-led
2. North Carolina State University, Crop and Soil Sciences. "When Crops Eat First: Enhanced Efficiency Fertilizers for Improved Nitrogen Management." July 25, 2022. Source for the 15–20% potential nitrogen rate reduction (Alex Woodley), the "up to 50% of applied nitrogen can be lost" figure, the mixed yield-response finding, and the $3/acre pilot incentive. https://cals.ncsu.edu/crop-and-soil-sciences/news/when-crops-eat-first/
3. AgWeb. "Is A Nitrogen Inhibitor Worth The Cost?" Cited for the commonly reported $15–20/acre nitrapyrin treatment cost and the nitrous oxide global-warming-potential figure. https://www.agweb.com/news/crops/crop-production/nitrogen-inhibitor-worth-cost

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.