Genetically Modified Seed Is Nearly All of the Corn in the Americas and None of the Rice Anywhere

Genetically modified seed covered 216.0 million hectares in 2025 across 30 countries — a record, 2.5% above the prior year, and 28.2% of the 765.9 million hectares of crop land the world plants. The market was worth USD 26.15 billion, or 33.4% of the USD 78.40 billion commercial seed market.

Then you break it down and it stops being one market at all.

Two of the four big crops are close to done. Cotton at 76.7% and soybean at 73.9% have almost no adoption left to win, and their remaining growth comes from trait refresh and price rather than from conversion. Maize, which is the largest value pool in the category at USD 12.35 billion, is only 35.8% converted — 75.3 million of 210.4 million hectares, roughly 135 million hectares still unconverted. And rice is zero.

So the premise most people carry about this market — that GM is what modern seed is — is true only inside the Americas, and only for three crops. Globally it is a partly-converted maize market attached to a saturated soybean and cotton complex, with one enormous crop that has never planted a single commercial hectare.

Maize is the whole headroom story, and it is geographic

The reason maize is only 35.8% converted is not that the trait is unavailable. It is that the maize that is GM is the maize grown in the United States, Brazil and Argentina, and the maize that is not GM is the maize grown in Asia and Africa.

Consider what the conversion is worth per hectare. GM maize generated about USD 163.9 a hectare in 2025, the highest per-hectare value of any crop implied by this data, and above the USD 121.1 average across all GM hectares. The value of every additional converted hectare is therefore the highest in the category — which is why the entire forward story of this market sits in germplasm adapted to tropical and sub-tropical conditions, and in regulatory access, and in nothing else.

That is an uncomfortable conclusion for anyone who models GM seed as a technology diffusion curve. Diffusion curves assume the constraint is time. Here the constraint is that the trait owner cannot sell into a country where the variety does not perform, and the varieties that perform in Malawi or Bihar are not the ones developed for Iowa.

China is the test, and it is running now

The Chinese programme moved from pilot to routine commercial planting in 2026, and the numbers are worth stating precisely because they define the only large-scale volume variable in this market.

The 2024 demonstration covered eight provinces and roughly 1.07 million mu. In 2025 it extended to thirteen provinces and passed 30 million mu. The 2026 season is the first conducted under ordinary commercial planting rules, with maize targeted at 57.85 million mu and soybean at 1.03 million mu across 23 provinces plus the Xinjiang Corps and Beidahuang — placing GM maize at close to 9% of a national maize seed area of about 650 million mu.

Three details matter more than the area target. Field efficacy against fall armyworm and stem borer is above 90%, with average yield gains of 8.5% and a cost-and-revenue benefit of RMB 132.5 per mu — the programme is holding because it pays, not because it is mandated. Cumulative variety approvals reached 239 by the fourth batch of 2026, 219 maize and 20 soybean, which is the supply-side evidence that the pipeline behind the target exists. And Beijing Dabeinong discloses that its events sit inside 133 of those 219 maize varieties — 60.73% of the national approval list, which makes the Chinese trait platform the genuinely new entrant in a market otherwise owned by three companies.

Asia-Pacific as a whole is the fastest-growing significant region at 9.05% a year, from USD 2.30 billion in 2025 to USD 4.60 billion in 2033. That growth rate is China, and it is a policy schedule rather than a market forecast.

Rice is not a slow conversion. It is a decision not taken.

Put the two crops side by side and the asymmetry is stark. Rice seed is a USD 6.35 billion market on roughly 172 million hectares — USD 36.9 a hectare, against USD 163.9 for GM maize. The gap is 4.44 times.

If rice converted at maize’s global penetration of 35.8%, that would be 61.6 million hectares. At maize’s per-hectare GM revenue it would be worth USD 10.1 billion — larger than the entire rice seed market as it stands, added on top of it.

That is the largest single option in world seed, and it is not being priced because it is not a technology question. China holds biosafety certificates for GM rice without commercial release, and has held them for many years, while moving corn and soybean to routine planting under the same regulator. The constraint is the export channel and the political weight of the staple, not the trait.

What the licensing shift changes

The mechanism by which the next 135 million maize hectares get converted is different from the one that converted the first 75 million.

In early 2026 Corteva settled all litigation with Bayer for a USD 610 million cash outflow, pulled its royalty-neutrality target forward two years to 2026, and set out to become a net out-licensor of seed technology. The settlement brought forward the availability of triple-stack maize traits for out-licensing by roughly five years, to as early as 2027. Branded seed is 67.3% of 2025 value and licensed technology 32.7%; the licensed share is forecast to reach 35.0% by 2033.

Read together, those two facts say something specific: access to elite traits is becoming purchasable rather than merely licensable. That widens the field of companies who can sell GM seed — which is how the unconverted maize area in Asia and Africa actually gets served, by regional breeders with local germplasm — while concentrating the ownership of the events in fewer hands. Value per hectare can rise at the same time as the seed margin available to a local licensee falls, and both are consistent with 5.90%.

The judgement

This is not one market growing at 5.90%. It is three things wearing the same label: a saturated American soybean and cotton complex where growth is price; a partly-converted maize pool where growth is geography and regulatory access; and a rice option that is a trade decision rather than a breeding one.

The forecast adds up to 5.90% because the first two roughly balance, and it contains nothing at all for the third. That is the right treatment. An option that requires a government to accept export risk does not belong in a base case — but it does belong in the watch list, and it is the only item on that list capable of moving this market by more than a percentage point.