S10.8 · Food, Agriculture & Beverage
A trillion-dollar physical commodity flow run through a handful of trading houses whose information-asymmetry edge AI is narrowing.
There is no single market size here — this is a merchandising and trading margin layered on top of physical commodity flows worth trillions of dollars. Individual player revenues give the better sense of scale: ADM generated roughly $85B in 2024 revenue, Cargill roughly $160B in fiscal 2024 (privately held), and Bunge roughly $47B in 2023 revenue. A small number of trading houses handle the large majority of global grain and oilseed flow, and the margin they earn has always rested partly on knowing things others do not. That informational edge is exactly what AI is narrowing.
Trading margins compressed through 2024-2025 as crop supply normalized after the price spikes of 2021-2023 (company earnings commentary, industry trade press, 2025). Origination spans the US, Brazil, Argentina, the Black Sea region and Australia; trading and finance functions concentrate in Geneva, Singapore and Chicago. The group the industry calls the "ABCD" — ADM, Bunge, Cargill and Louis Dreyfus — together with the further consolidation the group has undergone in recent years, handles a large majority of global grain and oilseed trade, commonly cited in the 70-90% range depending on the specific commodity and region measured (Oxfam Cereal Secrets; PitchBook, 2024).
The segment sits downstream of farmland and row-crop production and upstream of processing and food and feed manufacturing. Margins run low-single-digit on enormous volumes, and the moat is cumulative rather than singular — scale, storage and logistics infrastructure, and hedging sophistication, no one of which decides the outcome alone. The balance sheet defines the business: inventory and receivables financed against volatile commodity prices make this extremely working-capital intensive, which is why participants are underwritten as much on cash-conversion-cycle discipline as on reported earnings.
The major houses have largely completed their forward integration into primary processing and crushing, which moved them into higher-margin ingredient economics further down the value chain — genuinely additive to what a pure trading margin could earn. What continues among the largest handlers is consolidation, and that is defense: a response to margin compression in the core trading business, not a source of new dollars.
AI's structural effect targets the information asymmetry that has underpinned trading-desk margins for decades. Systems that ingest weather, logistics, freight and crop-condition data can now generate the kind of origination and logistics-optimization insight that used to require an experienced trading desk and a network of relationships to assemble. Scale and infrastructure keep their value — a model does not own an elevator or a freight position — but the informational edge that justified part of the margin captured by the largest players narrows. AI-driven logistics and trading optimization is already running inside all four major trading houses' operations. Inside this segment, the effect is a present reality, not a forecast.