More than 500 controlled experiments have collectively suggested that biodiversity loss reduces ecosystem productivity and stability. Yet the importance of biodiversity in sustaining the world's ecosystems remains controversial, largely because of the lack of validation in nature, where strong abiotic forcing and complex interactions are assumed to swamp biodiversity effects. Here we test this assumption by analysing 133 estimates reported in 67 field studies that statistically separated the effects of biodiversity on biomass production from those of abiotic forcing. Contrary to the prevailing opinion of the previous two decades that biodiversity would have rare or weak effects in nature, we show that biomass production increases with species richness in a wide range of wild taxa and ecosystems. In fact, after controlling for environmental covariates, increases in biomass with biodiversity are stronger in nature than has previously been documented in experiments and comparable to or stronger than the effects of other well-known drivers of productivity, including climate and nutrient availability. These results are consistent with the collective experimental evidence that species richness increases community biomass production, and suggest that the role of biodiversity in maintaining productive ecosystems should figure prominently in global change science and policy.