Background: Bipolar disorder (BD) is highly heritable and shares genetic overlap with schizophrenia and major depressive disorder (MDD), yet its core pathophysiological mechanisms remain unresolved. While glutamatergic dysregulation has been emphasized, excessive synaptic pruning—particularly immune-mediated—has emerged as a candidate from schizophrenia research. We competitively tested pruning versus glutamatergic/plasticity pathways using multi-method genomic analysis of recent large-scale GWAS data, with parallel MDD replication to delineate shared and disorder-specific signals. Methods: European-ancestry BD GWAS summary statistics (effective N ≈ 137,097) and trans-ancestry MDD data (European subsample effective N ≈ 829,250) were analyzed via: (1) MAGMA gene-based and competitive gene-set testing with overlap-controlled pruning (core, expanded, specific) and glutamatergic sets plus controls; (2) stratified LD score regression for partitioned heritability; (3) S-PrediXcan transcriptome-wide association in six brain tissues; (4) two-sample Mendelian randomization for neuroplasticity proxies (e.g., educational attainment); and (5) supplementary gene set enrichment analysis (GSEA) with differential contrasts. Results: Pruning pathways dominated BD across methods: expanded/specific sets achieved Bonferroni significance in MAGMA and strong LDSC enrichment, persisting independently of glutamatergic overlap. TWAS showed activation-skewed pruning (upregulated microglial/autophagy activators) with modest glutamatergic signals; GSEA reinforced immune subpath skew (e.g., MHC, schizophrenia-associated). Mendelian randomization revealed genetically proxied higher educational attainment causally increases BD risk (IVW OR = 1.403, p = 6.69 × 10⁻⁵) but protects against MDD (OR ≈ 0.72). MDD showed robust heritability-level pruning but mediated expression and protective reserve effects. Conclusions: Immune-skewed synaptic pruning emerges as the primary pathway in BD, yielding inflammation-vulnerable circuits amplified into episodic instability by cognitive reserve—distinguishing it from MDD's downstream pruning deficits. These findings reframe BD toward neuroimmune-pruning models with implications for targeted therapeutics and cross-disorder classification.