EUDR soy compliance starts with a scope question that confuses many companies. Soy is the highest-volume agricultural commodity in the EUDR basket, and feed mills, food manufacturers and ingredient traders all touch it, but not all of them file due diligence statements. This guide sorts out who is actually an operator, what the Brazilian data landscape makes possible, and how to be ready for 30 December 2026, or 30 June 2027 for micro and small enterprises.
EUDR soy product scope
1201- soya beans, whether or not broken1208.10- soya bean flour and meal1507- soya bean oil and its fractions2304- oilcake and solid residues from soya oil extraction (the feed meal that moves in bulk)
The most consequential boundary: animals fed on soy are not in scope. A chicken raised on soy meal, the eggs it lays, and the pork chop from a soy-fed pig carry no EUDR obligation. The regulation attaches to the listed soy products themselves. So a feed compounder importing meal from Brazil files a DDS; the poultry farm buying that compound feed does not. Similarly, highly processed derivatives not listed in Annex I - lecithin under 2923, protein isolates - fall outside current scope, though the Commission can extend the annex over time.
Operator analysis for a typical chain
- Importing beans, meal or oil from outside the EU: operator; full due diligence, DDS per consignment or an annual statement for recurring flows.
- Crushing EU-grown soy: the farmer or first buyer placing EU soy on the market files; domestic production is covered too, with the same geolocation duty.
- Buying meal from an EU trader: reference the upstream DDS; large downstream companies must also confirm due diligence happened upstream.
- Food manufacturers using soy oil bought in the EU: downstream; keep the reference chain intact in your records.
Geolocation at farm scale
Soy is the anti-coffee: instead of thousands of sub-hectare plots per consignment, you get hundreds of large fields, nearly all above 4 hectares - which means polygons are mandatory, not points. The good news is that industrial soy farming is already digitised: Brazilian farms carry CAR (rural environmental registry) boundaries, and major traders operate farm-polygon databases built for their own zero-deforestation commitments. Ask for polygon files per origination area; the formats are standard (GeoJSON, shapefile-derived exports) and the geolocation rules on winding, self-intersection and coordinate precision apply as usual.
Deforestation screening in soy landscapes
Soy's deforestation exposure is concentrated: the Cerrado and the Amazon's arc of deforestation in Brazil, the Chaco in Argentina and Paraguay. Two screening notes matter. First, the EUDR cutoff is 31 December 2020 and its forest definition covers forest as defined by FAO - plots converted from Cerrado savanna woodland raise genuinely harder classification questions than rainforest clearing, and convergence-of-evidence screening (per the open WHISP methodology) is how you avoid both false comfort and false alarms. Second, the Amazon Soy Moratorium's private cutoff (2008) means many Amazon-facing suppliers already maintain clean polygon data - reuse it.
Country benchmarking and volumes
Brazil, Argentina and Paraguay sit in the standard tier of the country benchmark; the US, a major EU soy supplier, is low-risk, qualifying for simplified due diligence - information collection including full geolocation, without the formal risk-assessment and mitigation steps unless something in the file contradicts the low-risk assumption. Splitting your book between low-risk and standard-risk origination is a legitimate compliance-cost lever, and several EU feed buyers have already shifted volumes for exactly this reason.
Segregation, mass balance and what actually files
The soy trade runs on mass-balance certification - RTRS and ProTerra credits, mixed physical flows. EUDR does not accept credits: the plots in the DDS must be the plausible physical origin of the product in the consignment. That has pushed traders toward area mass balance and segregated EUDR-ready flows, where a defined set of farms feeds a defined elevator-to-vessel chain and the polygon file travels with the cargo. When negotiating supply, the question is not "is it certified" but "which polygon set will accompany this parcel, and does it cover the full volume". Where a trader offers certified-but-unmapped volumes, the certificate still has value as legality and practice evidence in your risk file - it just cannot substitute for the geometry.
The DDS for bulk cargo
A Panamax of soy meal is still one placing on the market: one DDS can cover the consignment, carrying the polygon set for the contributing farms and the quantity in net mass. Where cargoes blend origination areas, the polygon file covers all of them. The mechanics - activity type, HS codes, quantities, reference numbers returned by TRACES - are field-by-field in the DDS template guide.
Priorities for a soy importer
- Classify products:
1201/1208/1507/2304in scope; fed livestock and unlisted derivatives out. - Request farm polygons per origination area from traders; they exist - the negotiation is about passing them downstream.
- Screen polygons against the 2020 cutoff before contracting the season, not after loading.
- Decide consignment-level versus annual DDS filing based on your flow regularity, and set up TRACES access early.
The two dominant origins are covered in depth: soy from Brazil (the Cerrado gap the moratorium never closed) and soy from Argentina (Gran Chaco exposure and the sector-wide VISEC platform), with the 2020 cutoff guide explaining the forest-definition questions both raise. Compare with palm oil, where mills rather than farms anchor the data chain, and cattle, where soy's feed exemption contrasts sharply with cattle's birth-to-slaughter coverage.