Purpose <p>This research develops a sustainable self-compacting geopolymer concrete that eliminates the need for mechanical vibration while improving flowability, compressive strength, and durability. Triple waste materials, such as recycled granite waste, rice husk ash, electric arc furnace slag, are incorporated to enhance performance and support circular economy goals.</p> Methods <p>self-compacting geopolymer concrete mixes with recycled granite waste, rice husk ash, electric arc furnace slag are prepared and tested. Laboratory data are used to train an artificial geographically weighted neural network, optimized via the linear congruential arctic fox algorithm. Response surface methodology (RSM) refined critical parameters affecting mechanical and physical properties.</p> Results <p>Sensitivity analysis identified curing temperature, Na₂SiO₃/NaOH ratio, NaOH molarity, and superplasticizer content as key influencers. The optimized mix (Na₂SiO₃/NaOH 2.5, liquid-to-solid ratio 0.22, 0.2% superplasticizer, 60&#xa0;°C oven curing for 24&#xa0;h) achieved 30&#xa0;MPa compressive strength, a slump flow of 740&#xa0;mm, a water absorption of 5.07–7.81%, and porosity 9.72–14.41%, demonstrating improved workability, microstructural density, and durability.</p> Conclusion <p>The triple-waste self-compacting geopolymer concrete efficiently optimizes performance without mechanical vibration. The proposed framework provides a reproducible and scalable approach for eco-efficient self-compacting geopolymer concrete design using industrial and agricultural by-products.</p> Graphical abstract <p></p>