Using Satellite Data for Geopolitical Trades

Satellite imagery has been commercially available for over a decade, but its application as a trading signal has matured significantly with the rise of computer vision. The ability to monitor physical infrastructure at scale — ports, pipelines, military installations, agricultural land — gives analysts an information edge that predates the official data releases markets move on.

The Signal Layer

Commercial satellite providers including Maxar, Planet Labs, and Airbus publish high-resolution imagery at daily refresh rates over key geographies. Analysts apply object detection models to count tankers at oil terminals, measure shadow lengths on crude storage tanks to estimate fill levels, track agricultural equipment density to forecast crop yields, and monitor construction activity at industrial facilities. The signal is pre-fundamental — it reflects physical reality before it appears in financial disclosures.

The most granular applications go beyond simple counting. Machine learning models trained on years of historical imagery can detect subtle changes in facility activity — increased truck movement at a logistics hub, the presence of specialised equipment at a mining site, or the absence of activity at a facility that was previously operational. Each of these observations carries potential informational content ahead of official reporting.

Translating Imagery into Positions

The most cited use case is crude oil inventory estimation. Storage tanks with floating roofs cast measurable shadows that correlate with volume. When satellite counts at Cushing, Oklahoma diverge from EIA inventory reports, the divergence creates a tradeable gap. In 2019, analysis of Iranian oil tanker movements ahead of sanctions enforcement gave positioning windows of several weeks before official commentary confirmed the thesis.

Similar edge exists in agricultural commodities, where satellite-derived crop condition indices run several weeks ahead of USDA reports. Analysts monitoring Brazilian soybean fields during the 2012 drought had quantitative estimates of yield deterioration before the USDA revised its figures. The same methodology applies to soft commodities across multiple growing regions simultaneously — a level of geographical coverage that no ground-based survey can match.

Risk and Latency Considerations

Satellite data has a latency problem at the edges: cloud cover, refresh frequency, and model accuracy all introduce noise. The edge is also narrowing as the signal becomes commoditised through vendors such as OrbitalInsight and SpaceKnow. The practical advantage today lies less in the raw imagery than in the quality of the processing pipeline and the speed at which the output reaches a portfolio manager's desk.

Geopolitical applications carry an additional category of risk: the event that generates the satellite signal may not resolve in a predictable direction or timeframe. A buildup of military equipment near a border is observable. Whether it results in conflict, a diplomatic resolution, or a sustained standoff is not. Position sizing and stop discipline matter significantly more in geopolitical trades than in purely data-driven quantitative applications.