A database of some 2000 geothermal gradient and heat flow calculations has been assembled across the African plate. This shows many basins in many different settings to have higher than expected geothermal gradients than would have been expected from a McKenzie thermal decay model. Geothermal gradients in excess of 40 deg C/km are surprisingly common, not only in active regimes such as the East African Rift and Red Sea but also in passive margin settings such as the eastern and western margins of the continent. The anomalies on the west Africa margin also seem to be reflected in similar anomalies on the Brazilian conjugate. These anomalies help to explain recent discovery trends, including several basins which turned out to be oil bearing where immaturity had once been predicted and others where more gas has been found than oil. Nearly all major African frontier discoveries of the 21st century are associated with thermal hotspots. The most obvious explanation for anomalies in long lived passive margin settings is intrusion of hot igneous rocks. Many anomalies, such as that associated with the Ahaggar in Algeria, are clearly associated with a shallow young igneous expression, though others are not. The width of many of the anomalies, e.g. 300km in the case of the Dakar hotspot in Senegal, appears suggestive of deep igneous effects, a model that has also been proposed for the Brazilian anomalies, which form a belt 100-300km wide close to the continental-ocean transition zone (COTZ). Although the public domain database on the west and east African margins is much poorer, available data suggest that the COTZ is also generally hot,highlighting a contrast with colder COTZs in the likes of the Gulf of Mexico and Black Sea. The data assembled here indicates that wider and generally higher ranges of heat flow should be applied in basin models across most African basins and petroleum systems.