🍖 Food
29.77% of the Nitrogen, If Every Hectare Matched the Isotope. Half Needs 1.68 Croplands.
Blanketing FAO's implied cropland at 21.2 kg of nitrogen per hectare replaces 29.77% of 2024's 115 million tonnes. The 95% interval still includes zero.
29.77 percent.
FAO's 30 July 2026 brief puts 2024 agricultural use of inorganic fertilizer at 197 million tonnes of nutrients and 122 kilograms per hectare of cropland, with nitrogen at 115 million tonnes, so 197 million divided by 0.000122 tonnes per hectare is 1,614,754,098 hectares of implied cropland, and 21.2 kilograms on each of those hectares, the mean Martinez-Feria and colleagues traced from air into maize by flowering, is 34.233 million tonnes, and 34.233 divided by 115 is 0.2977.
Half is a different sentence
On 10 November 2026, Haskett, Schnabel and eight coauthors published a Nature Biotechnology comment on engineered nitrogen-fixing bacteria, and Dinneen and Lenharo, in the Nature news report corrected on 11 November, wrote that the researchers suggest those bacteria could eventually replace up to half of synthetic fertilizers, which is 57.5 million tonnes, or 2,712,264,151 hectares at 21.2 kilograms each, and 2,712,264,151 divided by 1,614,754,098 is 1.68.
The ratio is 1.68.
Switch Bioworks' release the same day says the authors estimate that replacing 10 to 20 percent could eliminate tens of millions of metric tons of carbon-dioxide-equivalent from manufacturing each year, which is 11.5 or 23.0 million tonnes of nitrogen, not 57.5, and the climate clause is about the factory, not about every tonne of nitrous oxide leaving a field.
21.2 kilograms, and a jug named 40
| Published figure | As printed | kg N per hectare | lb N per acre |
|---|---|---|---|
| Isotope mean, VT to R1, 16 sites | 21.2 kg, 11% of biomass N | 21.2 | 18.91 |
| 95% interval on that mean | supplemental Table S8 | −0.16 to 42.6 | −0.14 to 38.01 |
| On-farm fertilizer cut, 58 strips | 39 to 45 kg below the grower rate | 39 to 45 | 34.79 to 40.15 |
| Agronomic-optimum shift at Purdue | ACRE, 2021 to 2022 | 10.5 to 27.4 | 9.37 to 24.45 |
| Name on PROVEN 40 | 40 in the product name | 44.83 | 40 |
The interval includes zero.
It also reaches 42.6 kilograms per hectare, 38.01 pounds, near the name on the jug and far from the 18.91-pound mean, and 42.6 kilograms times the implied cropland is 68.789 million tonnes, 59.8 percent of 2024 use, so half lives only on that friendly edge. Martinez-Feria et al. print the range, and a dry Purdue 2022, 77 millimeters in 50 days, about 36 percent of normal, left little signal, as did two of three Wisconsin leaf-tip tests.
What 21.2 kilograms does to a corn country
| Area | Hectares | Nitrogen at 21.2 kg | Share of 115 Mt | Share of half |
|---|---|---|---|---|
| Paper's 2023 adoption estimate | 2 million | 0.0424 Mt | 0.0369% | 0.0737% |
| Company's "nearly 15 million acres" | 6.070 million | 0.1287 Mt | 0.112% | 0.224% |
| US corn planted, 30 June 2026 | 38.584 million | 0.8180 Mt | 0.711% | 1.42% |
| FAO-implied world cropland | 1,614.8 million | 34.233 Mt | 29.77% | 59.5% |
| Hectares required for half | 2,712.3 million | 57.5 Mt | 50% | 100% |
NASS, on 30 June 2026, counted 95,343 thousand planted acres, 95.3 million in the release, and 95,343,000 times 0.40468564224 hectares is 38,583,943 hectares, times 21.2 kilograms is 0.8180 million tonnes: 0.711 percent of world use, 1.42 percent of the half. Pivot's 16 June 2025 note puts a typical rate at 160 to 200 pounds, so 18.91 pounds is 11.8 and 9.46 percent of those poles, and its 37-pound cut on 172 fields is a company set, not the paper's 58 strips.
Frey, via the Nature news report, has the products on tens of millions of hectares over a decade. The 2025 note says nearly 15 million validated acres, 6,147 fields, 34 states. Fifteen million acres is 6.070 million hectares, and 20 million hectares is 3.29 times that. Since 2022 the note counts 129,000 tonnes less synthetic nitrogen, 0.112 percent of one year's 115. Those validated acres at 21.2 kilograms equal 0.129 million tonnes, within 310 tonnes of 129,000. That match is an accident of arithmetic, not a shared ledger.
Five percent, 2.1 percent, and the factory share
Gao and Cabrera Serrenho put nitrogen-fertilizer production and use, synthetic product plus manure, at about 5 percent of global emissions. Menegat, Ledo and Tirado count synthetic nitrogen only: 1,129.1 million tonnes of CO2-equivalent in 2018, plus or minus 171.1, or 2.1 percent of a roughly 54-gigatonne total, of which factories were 438.5 million tonnes (38.8 percent) and fields 58.6 percent. Divide 438.5 by their 107.7 million tonnes and the factory intensity is 4.071 tonnes of CO2-equivalent per tonne of nitrogen, so 10 and 20 percent of 2024's 115 are 46.8 and 93.6 million tonnes if 2018 intensity still holds. Martinez-Feria's lower cited factor, about 2.6 kilograms per kilogram of nitrogen, gives 29.9 and 59.8 instead. Either way the phrase "tens of millions" fits the factory and misses the larger field term, and the maize paper prints no nitrous-oxide flux.
Yield held. The optimum may already have been passed.
Yield did not move.
Across 58 strips of about 11.26 hectares, the usual rate made 13.43 megagrams per hectare and the microbe plus a 39-to-45-kilogram cut made 13.42, contrast −0.0135, P equal to 0.871, below a minimum detectable effect of 0.166, while within-field variation fell, ratio 0.921, P equal to 0.0129. The authors, citing earlier surveys rather than these farms, say grower rates run 17 to 28 kilograms above the optimum, and at Purdue the fitted optimum moved only 10.5 to 27.4 kilograms. On 28 small plots the high dose and the restored fertilizer each added 0.21 megagrams.
Grove and Lee, on Woodward, Sible, Seebauer and Below, average 1.8 bushels an acre over four Illinois site-years, only one of them significant alone, and Michigan Farm News quotes Fred Below that the early 35-pound fertilizer equivalent "was down to about 10 by season's end," which is 11.2 kilograms a hectare. Their Kentucky rye trial gained 9 bushels in 2023, and in 2024 the gain cleared significance only at 215 pounds under heavier residue. Their summary of Franzen's NDSU circular SF2080 is 2 positive site-years out of 61 across crops and products, not this strain alone, a tally I am taking from Grove and Lee rather than from Franzen's circular or the Illinois supplement.
The company in the byline is not yet selling the microbe
Haskett, Bueno Batista, Samaddar and Schnabel work at Switch and hold equity, including US patent 12,649,699 and Schnabel's Stanford patent 12,460,213. Ané advises Pivot, Dixon advises Pivot and holds equity, and the note calls the firms independent. Ané's verdict on wild strains, in the Nature news piece, is "It doesn't work." Switch's 13 October release says a genetic timer made five times more nitrogen than always-on strains in a grow-room, with USDA and EPA trial clearance and a CTNBio approval that Mario Murakami, in Nature's report, calls the first transgenic bacterium cleared for Brazilian field tests. Temple's 1 September profile has Schnabel two to three years from a product: "It's too early to tell exactly how well this works in the field." Haskett, the same week as the comment: "Improved plant growth alone doesn't demonstrate that bacteria have transferred fixed nitrogen to crops."
What You Can Do
If a seller offers a 40-pound cut, ask which figure: 18.91 pounds from the isotope mean, an interval that still touches zero, or a Purdue optimum that moved 9.37 to 24.45 pounds. Flat yield after a cut is not replacement when the starting rate may already sit 17 to 28 kilograms above the optimum, so ask for season-end nitrogen. Funders and reporters should keep "up to half," the 10-to-20-percent factory case, and 21.2 kilograms in separate columns, and convert acres before comparing Frey's hectares with a 15-million-acre line. Nothing from Switch is for sale. What would change the arithmetic is a pre-registered multi-state interval, in kilograms per hectare, that excludes zero at season's end, on rates at or below the optimum, with a nitrous-oxide number beside it.
Limitations
This arithmetic did not prove the strains fix nothing, and the 10-to-20-percent line is the company release, because the comment and its tables are paywalled. Illinois figures are Grove and Lee plus a farm-press quotation, not a reanalysis, and the 61-site count is their summary of Franzen. At 121.5 or 122.5 kilograms per hectare, implied cropland is 1.621 or 1.608 billion hectares and 29.77 percent barely moves. FAO's July land brief puts irrigated area at 351 million hectares, about 22 percent of cropland, roughly 1.60 billion hectares. The 21.2-kilogram mean stops at flowering, a leaf tip may not stand for the plant, 2018 factory intensity may not describe 2024, and NASS still allows 93.0 to 97.7 million final acres.
The Strongest Case Against This Analysis
Grant the optimistic reading at full strength. Yield held after a 39-to-45-kilogram cut, variation fell, and the interval's top, 38.01 pounds, is close enough to the jug that a noisy tracer can hide a real 40-pound effect under a mean of 21.2. Illinois found air nitrogen in the crop, a Kentucky rye year added 9 bushels, and the release asked for 10 to 20 percent of manufacturing, not half the fertilizer supply this season. Ané's point, that wild microbes starve on cereals, is what a genetic switch is for, and the paper says extra fertilizer recovery, even if it inflates the isotope fraction, would still cut leaching and nitrous oxide. Twenty kilograms per hectare on millions of corn hectares is a real pile of ammonia, an interval through zero is what a genuine noisy effect looks like, and spreading a corn microbe across every orchard oversells a ceiling as a forecast.
That case can be right, and half can still be the wrong tense. The optimum moved 10.5 to 27.4 kilograms, the yield test is blind below 0.166 megagrams, grower rates already run hot by 17 to 28 kilograms, and Below's season-end figure is about 10 pounds. An estimated 2 million hectares in 2023, times 21.2 kilograms, is 0.0424 million tonnes, 0.0369 percent of the 115. Full cropland at the mean stops at 29.77 percent. A lab without the named equity has to print a season-end interval that excludes zero before 1.68 croplands is something to buy.
The Bottom Line
FAO counted 115 million tonnes of agricultural nitrogen in 2024. At 21.2 kilograms per hectare, copying that mean across the 1,614,754,098 hectares implied by 197 million tonnes divided by 122 kilograms replaces 34.233 million tonnes, 29.77 percent. Half of 115 is 57.5 million tonnes and 2,712,264,151 hectares, 1.68 times that cropland, while the 95 percent interval on the 21.2 still runs from −0.16 to 42.6 kilograms per hectare, and the product's own name converts to 44.83 kilograms, not 21.2.
Until the interval clears zero at the end of the season, half is a research sentence.