Schizophrenia, depression, and bipolar disorder are highly heritable psychiatric illnesses that share overlapping symptoms but also exhibit disorder-specific features. To dissect the cellular and developmental mechanisms of genetic risk, we integrated large-scale genome-wide association study (GWAS) data with human cortical single-nucleus RNA sequencing (snRNA-seq) data spanning gestation to adulthood (>700,000 nuclei from 106 donors). Gene-based analyses revealed 104 shared genes across disorders and convergent enrichment in synaptic pathways, alongside disorder-specific signals such as metal ion transport in schizophrenia. Using the single-cell disease relevance score (scDRS), we mapped polygenic risk across cortical cell types and developmental windows. Excitatory neurons were consistently implicated across all disorders from postnatal stages through adulthood, while inhibitory neurons showed broader vulnerability in depression and bipolar disorder, extending into the fetal period. Glial cells demonstrated disorder specificity: astrocytes were implicated across disorders during early postnatal synaptogenesis, oligodendrocyte precursor cells (OPCs) showed prolonged associations in depression, and mature oligodendrocytes were uniquely implicated in schizophrenia during childhood. These findings highlighted excitatory-inhibitory imbalance as a shared mechanism, alongside distinct glial and developmental trajectories contributing to disorder-specific pathophysiology. Our findings help to highlight the cortical cell types and developmental windows through which psychiatric genetic risk may act, offering insights into potential critical periods for intervention.