From the beginning of the 21th century, the issue of global food security has received increased scrutiny and a number of scenario studies concerned with the ability of the world agricultural system to feed the projected world population in 2030 and/or 2050 have been published. Among them, the FAO report (World Agriculture Towards 2030/250; Alexandratos and Bruinsma 2012) is the most well‐known projection exercise and very often held as a reference for baseline projections of world agricultural supply and demand. The most well‐known result of the Alexandratos and Bruinsma (2012)’s study, which is cited in nearly all studies dealing with world food security, is that world agricultural production has to increase by 60 % to satisfy global demand in 2050 (compared to production in 2005/2007). This 60 % increase in production would come mainly from an increase in crop yields (80 % of the production rise at the world level), some increase in cropping intensity (10 %) and limited land expansion (10 %). Because livestock production is less efficient than crop production in transforming solar energy into food calories, it is clear that reducing meat consumption, more precisely limiting meat consumption increases from current levels, would help solving the world quantitative food equation. Such a limitation could also have positive effects on both health and environment indicators. Nevertheless, livestock production and meat consumption deliver benefits that have to be taken into account. In that context, improving the energy and environmental footprint of livestock production and trade, including induced production and trade of crops for feed, is a priority for research, extension and actors (policy‐makers and livestock producers). In that perspective, the final section of the paper analyzes more specifically the question of land areas mobilized for feeding the world livestock. This question will be explored on the past (retrospective analysis) and for the next decades, at both the world and the regional level.