Global change has many facets, of which land use and the change of atmospheric chemistry are unquestioned primary agents, which induce a suite of secondary effects, including climatic changes. The largest single contribution to the compositional change of the atmosphere, CO2 enrichment, has (besides its influence on climate) immediate and direct effects on plants. Quantitatively, CO2 is the plant ‘food’ number one, and the rate of photosynthetic CO2 uptake by leaves is not yet CO2–saturated. This abrupt change of the biosphere's diet does and will affect all aspects of life, including our food. However, the plant and ecosystem responses are more subtle than had been assumed from the results of responses of isolated, well–fertilized and well–watered plants in greenhouses during the early days of CO2–enrichment research. In this article, I discuss potential responses of complex natural grassland and diverse forests, and address three key themes: CO2 and nutrients; CO2 and water; CO2 and plant–animal interactions. Examples from a suite of climatic regions emphasize that the most important ecosystem level responses to elevated CO2 will be introduced by differential responses of species. Atmospheric CO2 enrichment is a biodiversity issue. Classical physiological baseline responses of leaves to elevated CO2 can be overrun by biodiversity effects to such an extent that some of the traditional predictions may even become reversed. For instance, biodiversity effects may cause humid tropical forests (those which avoid destruction) to become more dynamic and store less, rather than more, carbon as CO2 enrichment continues. The abundance of certain life forms and species and their lifespans exert major controls over the half–life of carbon stored in forest biomass, and there is evidence that elevated CO2 can affect these controls and most likely does so already. Also, long–term hydrological consequences of atmospheric CO2 enrichment will be driven by biodiversity effects, given that some, but not all, species reduce their rate of water loss when exposed to a CO2–rich atmosphere. Such insights call for much more realistic experimental conditions and larger–scale test units, which permit biotic interactions across taxa and trophic levels to occur while simulating our CO2 future. The evidence currently available suggests that ecosystem processes reflect the composition of their biological inventory and this will be affected by a shift in carbon supply.