AbstractAerated concrete is a lightweight building material with excellent thermal and technical properties, relative to strength, excellent workability, and is economically advantageous. The basic composition of raw materials includes lime, cement, silica sand, and aluminum powder. Currently, the usability of secondary raw materials is being investigated in many construction sectors. The usability of secondary raw materials becomes a necessity from an ecological and economic point of view. The elimination of CO2 emissions is currently a major global issue. Secondary raw materials used in the construction industry are most often waste materials produced during the combustion of solid substances. These substances are often accumulated in landfills, which have a negative impact on the surrounding environment. It is necessary to constantly look for new ways of using these waste materials not only in the construction industry. The use of these raw materials reduces the price of the final product. It is advantageous if it also brings about a positive change in the properties of aerated concrete. The article deals with the possibility of replacing part or all of the sand content in the recipe of autoclaved aerated concrete with waste perlite. As part of the experimental verification, the influence of waste perlite filler on physical‐mechanical properties (e.g., bulk density, strength) and changes in microstructure is investigated. Primary sand is replaced by perlite in the amount of 10%, 30%, 50%, and 100% by volume in the experiments. The formation of porous structure, its characteristics, and mineralogical composition, in particular the formation of tobermorite, is also evaluated. The study has brought an interesting finding. The results show that the maximum beneficial amount of waste perlite is 30%. Up to this limit, waste perlite has a positive effect on the even distribution of pores in the matrix, with a zero‐phase content of more than 65%. The compressive strength of the samples with waste perlite increases by up to 22% of the values of the compressive strength of the reference sample.